The increasing growth in the volume of biomolecular data has introduced significant challenges for extracting meaningful molecular-level insights, particularly in predicting interactions between biological sequences such as DNA, RNA, and proteins. These interactions are fundamental to complex biological processes, including gene regulation and immune response. Artificial Intelligence (AI) has played a major role in advancing discoveries in this field, enabling the identification of novel interactions, as demonstrated by various predictive modeling studies. Despite the growing number of scientific publications in this domain, accessibility to practical computational tools has not progressed at the same pace. Existing studies differ substantially in availability: some provide only methodological descriptions, others release source code exclusively for experimental reproducibility, and only a limited number deliver fully automated solutions ready for broad use. Given this context, this paper investigates state-of-the-art studies in biological sequence interaction prediction, emphasizing the public accessibility and usability of available tools, especially for researchers who are not experts in AI or computational methods. We compile and discuss the input requirements of current tools, along with the types of outputs they generate, enabling users to better understand the scenarios in which each solution can be effectively applied. Furthermore, we analyze accessibility-related aspects to support informed selection of tools according to user expertise, ranging from web-based servers with pretrained models that require minimal computational skills to fully end-to-end frameworks capable of training new models on user-defined datasets, though often lacking user-friendly interfaces.

Machine learning algorithms have been increasingly used to solve problems related to financial markets. Predicting the price or trends of assets in different markets is the subject of several studies. As other emerging markets, the Brazilian energy market has grown in number of traded volume and players, providing a business environment that can be explored through price trend forecasting techniques. The formation of prices of energy products depends on several factors. As the source of energy production in Brazil is predominantly through hydroelectric plants, information on water storage and natural flow between hydroelectric plants can directly impact price fluctuations. We propose a new strategy based on machine learning and complex networks for predicting the price trend in energy markets. The new strategy uses storage and flow data from the most important water stations in a hydroelectric power network. For this, we developed a feature selection method based on complex networks to identify and filter the main storage stations of the Brazilian water network. Data from the selected stations are then used as input for machine learning algorithms. We use three methods to extract centrality metrics from storage stations. Experimental results indicate that better performance is obtained by the proposed strategy, when compared to two baseline methods using different machine learning algorithms. The better performance is confirmed by a financial analysis of the results when simulating the investment strategies using algotrading.

Data production has followed an increased growth in the last years, to the point that traditional or batch machine-learning (ML) algorithms cannot cope with the sheer volume of generated data. Stream or online ML presents itself as a viable solution to deal with the dynamic nature of streaming data. Besides coping with the inherent challenges of streaming data, online ML solutions must be accurate, fast, and bear a reduced memory footprint. We propose a new decision tree-based ensemble algorithm for online ML regression named online extra trees (OXT). Our proposal takes inspiration from the batch learning extra trees (XT) algorithm, a popular and faster alternative to random forest (RF). While speed and memory costs might not be a central concern in most batch applications, they become crucial in data stream data learning. Our proposal combines subbagging (sampling without replacement), random tree split points, and model trees to deliver competitive prediction errors and reduced computational costs. Throughout an extensive experimental evaluation comprising 22 real-world and synthetic datasets, we compare OXT against the state-of-the-art adaptive RF (ARF) and other incremental regressors. OXT is generally more accurate than its competitors while running significantly faster than ARF and expending significantly less memory.

Recent technological advances have led to an exponential expansion of biological sequence data and extraction of meaningful information through Machine Learning (ML) algorithms. This knowledge has improved the understanding of mechanisms related to several fatal diseases, e.g. Cancer and coronavirus disease 2019, helping to develop innovative solutions, such as CRISPR-based gene editing, coronavirus vaccine and precision medicine. These advances benefit our society and economy, directly impacting people's lives in various areas, such as health care, drug discovery, forensic analysis and food processing. Nevertheless, ML-based approaches to biological data require representative, quantitative and informative features. Many ML algorithms can handle only numerical data, and therefore sequences need to be translated into a numerical feature vector. This process, known as feature extraction, is a fundamental step for developing high-quality ML-based models in bioinformatics, by allowing the feature engineering stage, with design and selection of suitable features. Feature engineering, ML algorithm selection and hyperparameter tuning are often manual and time-consuming processes, requiring extensive domain knowledge. To deal with this problem, we present a new package: BioAutoML. BioAutoML automatically runs an end-to-end ML pipeline, extracting numerical and informative features from biological sequence databases, using the MathFeature package, and automating the feature selection, ML algorithm(s) recommendation and tuning of the selected algorithm(s) hyperparameters, using Automated ML (AutoML). BioAutoML has two components, divided into four modules: (1) automated feature engineering (feature extraction and selection modules) and (2) Metalearning (algorithm recommendation and hyper-parameter tuning modules). We experimentally evaluate BioAutoML in two different scenarios: (i) prediction of the three main classes of noncoding RNAs (ncRNAs) and (ii) prediction of the eight categories of ncRNAs in bacteria, including housekeeping and regulatory types. To assess BioAutoML predictive performance, it is experimentally compared with two other AutoML tools (RECIPE and TPOT). According to the experimental results, BioAutoML can accelerate new studies, reducing the cost of feature engineering processing and either keeping or improving predictive performance. BioAutoML is freely available at https://github.com/Bonidia/BioAutoML.

One of the main challenges in applying machine learning algorithms to biological sequence data is how to numerically represent a sequence in a numeric input vector. Feature extraction techniques capable of extracting numerical information from biological sequences have been reported in the literature. However, many of these techniques are not available in existing packages, such as mathematical descriptors. This paper presents a new package, MathFeature, which implements mathematical descriptors able to extract relevant numerical information from biological sequences, i.e. DNA, RNA and proteins (prediction of structural features along the primary sequence of amino acids). MathFeature makes available 20 numerical feature extraction descriptors based on approaches found in the literature, e.g. multiple numeric mappings, genomic signal processing, chaos game theory, entropy and complex networks. MathFeature also allows the extraction of alternative features, complementing the existing packages. To ensure that our descriptors are robust and to assess their relevance, experimental results are presented in nine case studies. According to these results, the features extracted by MathFeature showed high performance (0.6350-0.9897, accuracy), both applying only mathematical descriptors, but also hybridization with well-known descriptors in the literature. Finally, through MathFeature, we overcame several studies in eight benchmark datasets, exemplifying the robustness and viability of the proposed package. MathFeature has advanced in the area by bringing descriptors not available in other packages, as well as allowing non-experts to use feature extraction techniques.

Choosing the most suitable algorithm to perform a machine learning task for a new problem is a recurrent and complex task. In multi-target regression tasks, when problem transformation methods are applied, this choice is even harder. The reason is the need to simultaneously choose the problem transformation method and the base learning algorithm. This work investigates how to bridge the gap of method/base learner recommendation for problems with multiple outputs. In meta-learning experiments, we use a large number of multi-target regression datasets to investigate whether using meta-learning can provide good recommendations. To do this, we compared the meta-models induced by 3 different ML algorithms, including three variations for each of them, and selected 58 meta-features that we believe are relevant for extracting good dataset descriptions for the meta-learning process. In the experimental results, the meta-models outperformed the baselines (Majority and Random) by recommending the most suitable solution for multi-target regression (for the transformation method and base-learner) with high predictive performance, including real-world applications. The meta-features and the relation between the transformation method and base-learner provided important insights regarding the optimal problem transformation method. Furthermore, when comparing the application of algorithm adaptation and problem transformation methods, our meta-learning proposal was capable of statistically overcoming all competitors, which resulted in a predictive performance using the best choice per problem.

a b s t r a c tMeta-learning is increasingly used to support the recommendation of machine learning algorithms and their configurations.These recommendations are made based on meta-data, consisting of performance evaluations of algorithms and characterizations on prior datasets.These characterizations, also called meta-features, describe properties of the data which are predictive for the performance of machine learning algorithms trained on them.Unfortunately, despite being used in many studies, meta-features are not uniformly described, organized and computed, making many empirical studies irreproducible and hard to compare.This paper aims to deal with this by systematizing and standardizing data characterization measures for classification datasets used in meta-learning.Moreover, it presents an extensive list of meta-features and characterization tools, which can be used as a guide for new practitioners.By identifying particularities and subtle issues related to the characterization measures, this survey points out possible future directions that the development of meta-features for meta-learning can assume.

Recent studies have shown that hybrid forecasting models tend to be a powerful tool to forecast univariate time series. However, most of these models are applied to time series of specific domains and do not report general performance analysis for several time series application domains. In this work, we designed a procedure that uses the Seasonal-Trend decomposition based on Loess as a preprocessing step to model the time series components separately using a machine learning algorithm and a seasonal naive forecaster. Finally, we analyze under which conditions our proposed framework can improve a standard machine learning model's predictive performance. Results have shown that our hybrid forecasting framework achieves a significant advantage in comparison to standard machine learning.

This work investigates information retrieval methods to address the existing difficulties on the Preliminary Search, part of the law making process from the Brazilian Chamber of Deputies. For such, different preprocessing approaches, stemmers, language models, and BM25 variants were compared. Two legislative corpora from Chamber were used to build and validate the pipeline. All texts were converted to lowercase and had stopwords, accentuation, and punctuation removed. Words were represented by their stem combined with word unigram and bigram language models. Retrieving the bill that was originated from a specific job request, the BM25L with Savoy stemmer reached a R@20 of 0.7356. After removing queries with inconsistencies or which made reference exclusively to attachments, to other job requests, or to bills, the R@20 increased to 0.94.

As consequence of the various genomic sequencing projects, an increasing volume of biological sequence data is being produced. Although machine learning algorithms have been successfully applied to a large number of genomic sequence-related problems, the results are largely affected by the type and number of features extracted. This effect has motivated new algorithms and pipeline proposals, mainly involving feature extraction problems, in which extracting significant discriminatory information from a biological set is challenging. Considering this, our work proposes a new study of feature extraction approaches based on mathematical features (numerical mapping with Fourier, entropy and complex networks). As a case study, we analyze long non-coding RNA sequences. Moreover, we separated this work into three studies. First, we assessed our proposal with the most addressed problem in our review, e.g. lncRNA and mRNA; second, we also validate the mathematical features in different classification problems, to predict the class of lncRNA, e.g. circular RNAs sequences; third, we analyze its robustness in scenarios with imbalanced data. The experimental results demonstrated three main contributions: first, an in-depth study of several mathematical features; second, a new feature extraction pipeline; and third, its high performance and robustness for distinct RNA sequence classification. Availability:https://github.com/Bonidia/FeatureExtraction_BiologicalSequences.

MOTIVATION: CRISPR-Cas are important systems found in most archaeal and many bacterial genomes, providing adaptive immunity against mobile genetic elements in prokaryotes. The CRISPR-Cas systems are encoded by a set of consecutive cas genes, here termed cassette. The identification of cassette boundaries is key for finding cassettes in CRISPR research field. This is often carried out by using Hidden Markov Models and manual annotation. In this article, we propose the first method able to automatically define the cassette boundaries. In addition, we present a Cas-type predictive model used by the method to assign each gene located in the region defined by a cassette's boundaries a Cas label from a set of pre-defined Cas types. Furthermore, the proposed method can detect potentially new cas genes and decompose a cassette into its modules. RESULTS: We evaluate the predictive performance of our proposed method on data collected from the two most recent CRISPR classification studies. In our experiments, we obtain an average similarity of 0.86 between the predicted and expected cassettes. Besides, we achieve F-scores above 0.9 for the classification of cas genes of known types and 0.73 for the unknown ones. Finally, we conduct two additional study cases, where we investigate the occurrence of potentially new cas genes and the occurrence of module exchange between different genomes. AVAILABILITY AND IMPLEMENTATION: https://github.com/BackofenLab/Casboundary. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Human Activity Recognition is focused on the use of sensing technology to classify human activities and to infer human behavior. While traditional machine learning approaches use hand-crafted features to train their models, recent advancements in neural networks allow for automatic feature extraction. Auto-encoders are a type of neural network that can learn complex representations of the data and are commonly used for anomaly detection. In this work we propose a novel multi-class algorithm which consists of an ensemble of auto-encoders where each auto-encoder is associated with a unique class. We compared the proposed approach with other state-of-the-art approaches in the context of human activity recognition. Experimental results show that ensembles of auto-encoders can be efficient, robust and competitive. Moreover, this modular classifier structure allows for more flexible models. For example, the extension of the number of classes, by the inclusion of new auto-encoders, without the necessity to retrain the whole model.

A central aspect of online decision tree solutions is evaluating the incoming data and enabling model growth. For such, trees much deal with different kinds of input features and partition them to learn from the data. Numerical features are no exception, and they pose additional challenges compared to other kinds of features, as there is no trivial strategy to choose the best point to make a split decision. The problem is even more challenging in regression tasks because both the features and the target are continuous. Typical online solutions evaluate and store all the points monitored between split attempts, which goes against the constraints posed in real-time applications. In this paper, we introduce the Quantization Observer (QO), a simple yet effective hashing-based algorithm to monitor and evaluate split point candidates in numerical features for online tree regressors. QO can be easily integrated into incremental decision trees, such as Hoeffding Trees, and it has a monitoring cost of $O(1)$ per instance and sub-linear cost to evaluate split candidates. Previous solutions had a $O(\log n)$ cost per insertion (in the best case) and a linear cost to evaluate split points. Our extensive experimental setup highlights QO's effectiveness in providing accurate split point suggestions while spending much less memory and processing time than its competitors.

Meta-Learning has been largely used over the last years to support the recommendation of the most suitable machine learning algorithm(s) and hyperparameters for new datasets. Traditionally, a meta-base is created containing meta-features extracted from several datasets along with the performance of a pool of machine learning algorithms when applied to these datasets. The meta-features must describe essential aspects of the dataset and distinguish different problems and solutions. However, if one wants the use of Meta-Learning to be computationally efficient, the extraction of the meta-feature values should also show a low computational cost, considering a trade-off between the time spent to run all the algorithms and the time required to extract the meta-features. One class of measures with successful results in the characterization of classification datasets is concerned with estimating the underlying complexity of the classification problem. These data complexity measures take into account the overlap between classes imposed by the feature values, the separability of the classes and distribution of the instances within the classes. However, the extraction of these measures from datasets usually presents a high computational cost. In this paper, we propose an empirical approach designed to decrease the computational cost of computing the data complexity measures, while still keeping their descriptive ability. The proposal consists of a novel Meta-Learning system able to predict the values of the data complexity measures for a dataset by using simpler meta-features as input. In an extensive set of experiments, we show that the predictive performance achieved by Meta-Learning systems which use the predicted data complexity measures is similar to the performance obtained using the original data complexity measures, but the computational cost involved in their computation is significantly reduced.

Automated recommendation of machine learning algorithms is receiving a large deal of attention, not only because they can recommend the most suitable algorithms for a new task, but also because they can support efficient hyper-parameter tuning, leading to better machine learning solutions. The automated recommendation can be implemented using meta-learning, learning from previous learning experiences, to create a meta-model able to associate a data set to the predictive performance of machine learning algorithms. Although a large number of publications report the use of meta-learning, reproduction and comparison of meta-learning experiments is a difficult task. The literature lacks extensive and comprehensive public tools that enable the reproducible investigation of the different meta-learning approaches. An alternative to deal with this difficulty is to develop a meta-feature extractor package with the main characterization measures, following uniform guidelines that facilitate the use and inclusion of new meta-features. In this paper, we propose two Meta-Feature Extractor (MFE) packages, written in both Python and R, to fill this lack. The packages follow recent frameworks for meta-feature extraction, aiming to facilitate the reproducibility of meta-learning experiments.

One of the main challenges in Clustering Analysis is choosing the optimal number of clusters. A typical methodology is to evaluate a validity index over the data and to optimize it as a function of the number of clusters. However, this process can have a high computational cost. In this work, we introduce a new approach for recommending the number of clusters for a particular dataset by using Meta-learning. As the predictive performance of the meta-models induced by Meta-learning is affected by how datasets are described by meta-features, we propose a new set of meta-features able to improve the predictive performance of meta-models used for recommending the number of clusters. Experimental results show that the proposed approach provides a good recommendation of the number of clusters. Additionally, the proposed meta-feature obtains better results than meta-features for clustering tasks found in the literature.

Imbalanced datasets may negatively impact the predictive performance of most classical classification algorithms. This problem, commonly found in real-world, is known in machine learning domain as imbalanced learning. Most techniques proposed to deal with imbalanced learning have been proposed and applied only to binary classification. When applied to multiclass tasks, their efficiency usually decreases and negative side effects may appear. This paper addresses these limitations by presenting a novel adaptive approach, E-MOSAIC (Ensemble of Classifiers based on MultiObjective Genetic Sampling for Imbalanced Classification). E-MOSAIC evolves a selection of samples extracted from training dataset, which are treated as individuals of a MOEA. The multiobjective process looks for the best combinations of instances capable of producing classifiers with high predictive accuracy in all classes. E-MOSAIC also incorporates two mechanisms to promote the diversity of these classifiers, which are combined into an ensemble specifically designed for imbalanced learning. Experiments using twenty imbalanced multi-class datasets were carried out. In these experiments, the predictive performance of E-MOSAIC is compared with state-of-the-art methods, including methods based on presampling, active-learning, cost-sensitive, and boosting. According to the experimental results, the proposed method obtained the best predictive performance for the multiclass accuracy measures mAUC and G-mean.

Community detection looks for groups of nodes in networks, mainly using network topological, link-based features, not taking into account features associated with each node. Clustering algorithms, on the other hand, look for groups of objects using features describing each object. Recently, link features and node attributes have been combined to improve community detection. Community detection methods can be designed to identify communities that are disjoint or overlapping, crisp or soft and static or dynamic. In this paper, we propose a dynamic community detection method for finding soft overlapping groups in temporal networks with node attributes. Our approach is based on a non-negative matrix factorization model that uses automatic relevance determination to detect the number of communities. Preliminary results on toy and artificial networks, are promising. To the extent of our knowledge, a dynamic approach that includes link and node information, for soft overlapping community detection, has not been proposed before.

Qualitative data clustering is a fundamental data analysis task, with applications in many areas, like medicine, sociology, and economics. An appealing way to deal with this task is via Integer Linear Programming, as it avoids inappropriate inferences by the final user. This approach has two main advantages: the data are directly used, without the need of being converted to quantitative values, and the optimal number of clusters is automatically obtained by solving the optimization problem. However, it might create large and redundant models, which can limit the size of the problems it can be applied. Recently, models that are more compact and able to avoid some redundancy have been proposed in the literature. These models consume less memory and are faster to obtain the optimal solution set. In this study, a new model is introduced and compared with the state-of-the-art alternatives using datasets from different application domains. Empirical results show that the new model outperforms its predecessors, achieving the optimal solution set with lower computational time and memory consumption.

In this work, the goal is to use clustering algorithms as recommender in a meta-learning system and, thus, to propose an unsupervised meta-learning approach. Meta-learning has been successfully used for recommendation of Machine Learning algorithms in several Data Mining tasks. Meta-learning can rank algorithms according to their adequacy for a new dataset and use this ranking to recommend algorithms. The recommendations are usually made by predictive meta-models induced by supervised Machine Learning techniques, therefore needing a target attribute. In many situations, the target attribute is not available or has a high computational cost. In these situations, the use of unsupervised meta-models (as clustering algorithms) could provide important insights from Machine Learning experiments, like the interpretation of the partitions found by these clustering algorithms. Here, clustering algorithms are used as unsupervised meta-models. Experimental results show that the proposed approach achieved better clustering quality.

MOTIVATION: With the recent advances in DNA sequencing technologies, the study of the genetic composition of living organisms has become more accessible for researchers. Several advances have been achieved because of it, especially in the health sciences. However, many challenges which emerge from the complexity of sequencing projects remain unsolved. Among them is the task of assembling DNA fragments from previously unsequenced organisms, which is classified as an NP-hard (nondeterministic polynomial time hard) problem, for which no efficient computational solution with reasonable execution time exists. However, several tools that produce approximate solutions have been used with results that have facilitated scientific discoveries, although there is ample room for improvement. As with other NP-hard problems, machine learning algorithms have been one of the approaches used in recent years in an attempt to find better solutions to the DNA fragment assembly problem, although still at a low scale. RESULTS: This paper presents a broad review of pioneering literature comprising artificial intelligence-based DNA assemblers-particularly the ones that use machine learning-to provide an overview of state-of-the-art approaches and to serve as a starting point for further study in this field.

Meta-learning has been successfully used for algorithm recommendation tasks. It uses machine learning to induce meta-models able to predict the best algorithms for a new dataset. In this paper, meta-models are applied to a set of meta-features, describing a dataset, to predict the performance of clustering algorithms applied to this dataset. The paper also proposes a new set of meta-features, based on correlation and dissimilarity measures. Experimental results show that these meta-features improve the recommendation. Additionally, this paper evaluates the importance of each meta-feature for the recommendation.

Noise is often present in real datasets used for training Machine Learning classifiers. Their disruptive effects in the learning process may include: increasing the complexity of the induced models, a higher processing time and a reduced predictive power in the classification of new examples. Therefore, treating noisy data in a preprocessing step is crucial for improving data quality and to reduce their harmful effects in the learning process. There are various filters using different concepts for identifying noisy examples in a dataset. Their ability in noise preprocessing is usually assessed in the identification of artificial noise injected into one or more datasets. This is performed to overcome the limitation that only a domain expert can guarantee whether a real example is indeed noisy. The most frequently used label noise injection method is the noise at random method, in which a percentage of the training examples have their labels randomly exchanged. This is carried out regardless of the characteristics and example space positions of the selected examples. This paper proposes two novel methods to inject label noise in classification datasets. These methods, based on complexity measures, can produce more challenging and realistic noisy datasets by the disturbance of the labels of critical examples situated close to the decision borders and can improve the noise filtering evaluation. An extensive experimental evaluation of different noise filters is performed using public datasets with imputed label noise and the influence of the noise injection methods are compared in both data preprocessing and classification steps.

Image segmentation is a key issue in image processing. New image segmentation algorithms have been proposed in the last years. However, there is no optimal algorithm for every image processing task. The selection of the most suitable algorithm usually occurs by testing every possible algorithm or using knowledge from previous problems. These processes can have a high computational cost. Meta-learning has been successfully used in the machine learning research community for the recommendation of the most suitable machine learning algorithm for a new dataset. We believe that meta-learning can also be useful to select the most suitable image segmentation algorithm. This hypothesis is investigated in this paper. For such, we perform experiments with eight segmentation algorithms from two approaches using a segmentation benchmark of 300 images and 2100 augmented images. The experimental results showed that meta-learning can recommend the most suitable segmentation algorithm with more than 80% of accuracy for one group of algorithms and with 69% for the other group, overcoming the baselines used regarding recommendation and segmentation performance.

This Research Full Paper introduces a machine learning methodology to automatically identify the learning style of students interacting with a Learning Management System. Studies in Cognitive Psychology and Pedagogy have already reported that each individual has a specific Learning Style, which describes her/his best means of perceiving and acquiring knowledge. The detection of the personal Learning Style of each student has long been made by using questionnaires; an analysis that demands too much effort, mainly in courses with hundreds of students. Therefore, the automatic modeling of learning styles has gained attention in the computing and education areas. This study compares different Machine Learning algorithms for the detection of students' Learning Styles. As such, a dataset is extracted from a real course in the Moodle learning platform. This course had 105 students interacting with 252 learning objects during 12 months. The learning styles were described using the classic model of Felder-Silverman. According to the experimental results using these data, a single machine learning algorithm was not able to induce models with predictive accuracy comparable to those from existing alternatives. However, when models from different algorithms were combined, it was possible to obtain a predictive accuracy superior to those reported in the related literature.

In imbalanced classification tasks, the training datasets may show class overlapping and classes of low density. In these scenarios, the predictions for the minority class are impaired. Although assessing the imbalance level of a training set is straightforward, it is hard to measure other aspects that may affect the predictive performance of classification algorithms in imbalanced tasks. This paper presents a set of measures designed to understand the difficulty of imbalanced classification tasks by regarding on each class individually. They are adapted from popular data complexity measures for classification problems, which are shown to perform poorly in imbalanced scenarios. Experiments on synthetic datasets with different levels of imbalance, class overlapping and density of the classes show that the proposed adaptations can better explain the difficulty of imbalanced classification tasks.

Protein Function Prediction is a complex Hierarchical Multi-label Classification task where the functional classes involved are organized in a hierarchy. While many Machine Learning methods have been proposed for this task, very few studies were performed for feature selection in such hierarchical scenarios. In this paper, we investigate feature selection techniques for hierarchical multi-label classification of protein functions. As decision trees are natural feature selectors, we rely on a hierarchical multi-label decision tree induction algorithm to extract features represented by the internal nodes of the tree. We also investigated the performance of a ReliefF-based non-hierarchical multi-label feature selection technique on the hierarchical scenario. We tested the different techniques on two classifiers, based on neural networks and genetic algorithms. The experimental results show that, in very few cases, the existing feature selection techniques were able to improve the classifiers performances, showing the need for developing feature selectors specifically to consider hierarchical class relationships.

When a Machine Learning algorithm is applied to a dataset, the predictive performance of the algorithm depends on how suitable its bias is to the the data distribution in the dataset, which leads researchers to create a large number of algorithms. The recommendation of the most suitable algorithm for a new dataset can occur by trial and error, trying a large number of algorithms with distinct bias. However, this approach usually has a high computational cost. This cost could be reduced if the most suitable algorithm(s) could be recommended. Meta-learning has been successfully used for recommendation of the best Machine Learning algorithm in several Machine Learning tasks. Meta-learning can rank algorithms according to their adequacy for a new dataset and use this ranking to recommend the algorithms to be used. As the recommended ranking is based on dataset features, dataset characterization (using meta-features) is of crucial importance for the successful use of meta-learning. Clustering is one of the main application of Machine Learning algorithms, however few works investigate the use of meta-learning for the recommendation of clustering algorithms. Moreover, the existing works use a poor methodology for the evaluation of the algorithm recommendation method and a small number of datasets. This paper proposes a comparison between two types of meta-features for clustering algorithm recommendation using meta-learning. Experimental results show in which situations the use of each type of meta-features is more suitable.

As Collaborative Filtering becomes increasingly important in both academia and industry recommendation solutions, it also becomes imperative to study the algorithm selection task in this domain. This problem aims at finding automatic solutions which enable the selection of the best algorithms for a new problem, without performing full-fledged training and validation procedures. Existing work in this area includes several approaches using Metalearning, which relate the characteristics of the problem domain with the performance of the algorithms. This study explores an alternative approach to deal with this problem. Since, in essence, the algorithm selection problem is a recommendation problem, we investigate the use of Collaborative Filtering algorithms to select Collaborative Filtering algorithms. The proposed approach integrates subsampling landmarkers, a data characterization approach commonly used in Metalearning, with a Collaborative Filtering methodology, named CF4CF. The predictive performance obtained by CF4CF using benchmark recommendation datasets was similar or superior to that obtained with Metalearning.

Abstract Food trucks are a widely popular fast food restaurant alternative, whose differentiating factor is their proximity to customers. Their popularity has stimulated the expansion of available options, which now includes several different types of cuisines, consequently making the choice by customers a challenging issue. From data obtained via a market research, in which hundreds of participants provided their food truck preferences, this paper focuses on the problem of food truck recommendation using a multilabel approach. In particular, it investigates how to improve the recommendation task regarding a previous work, where some labels have never been predicted. In order to address this problem, different alternatives were investigated. One of these alternatives, the Ensemble of Single Label, proposed in this paper, was able to reduce it. Despite its simplicity, good predictive results were obtained when they were used in the investigated task. Among other benefits, all labels were correctly predicted at least for few instances.

Many real-world situations constantly generate concept-drifting data streams at high speed. These situations demand adaptive algorithms able to learn online in accordance with the most recent target function (concept). This paper presents Online Adap

A feedback solution for approximate optimal scheduling of switched systems with autonomous subsystems and continuous-time dynamics is presented. The proposed solution is based on policy iteration algorithm which provides the optimal switching schedule. Algorithms for offline, online, and concurrent implementation of the proposed solution are presented. For online and concurrent training, gradient descent training laws are used and the performance of the training laws is analyzed. The effectiveness of the presented algorithms is verified through numerical simulations.

The problem of information overload motivated the appearance of Recommender Systems.From the several open problems in this area, the decision of which is the best recommendation algorithm for a specific problem is one of the most important and less studied.The current trend to solve this problem is the experimental evaluation of several recommendation algorithms in a handful of datasets.However, these studies require an extensive amount of computational resources, particularly processing time.To avoid these drawbacks, researchers have investigated the use of Metalearning to select the best recommendation algorithms in different scopes.Such studies allow to understand the relationships between data characteristics and the relative performance of recommendation algorithms, which can be used to select the best algorithm(s) for a new problem.The contributions of this study are two-fold: 1) to identify and discuss the key concepts of algorithm selection for recommendation algorithms via a systematic literature review and 2) to perform an experimental study on the Metalearning approaches reviewed in order to identify the most promising concepts for automatic selection of recommendation algorithms.

A large amount of information is continuously generated in intensive health care. An analysis of these data streams can supply valuable insights to improve the monitoring of the patients. The volume, frequency and complexity of data, which come unlabeled, make their analysis a challenging task. Machine learning (ML) techniques have been successfully employed for mining data streams to extract useful knowledge for health care monitoring. It includes the detection of changes in the behavior of sensors, failures on machines or systems, and data anomalies. Anomaly (or outlier) detection is a ML task that aims to find exceptions or abnormalities in a dataset. These exceptions, in a medical context, can represent a new disease pattern, an event to be further investigated, behavior changes or potential health complications. Despite of its analysis in data streams is a challenging task, temporal abstractions techniques should help due to they deal with the management and abstraction of time based data, offering high level of visualization of each data object in its context. The aim of this paper is to review recent research in anomaly detection and temporal abstractions and discuss the application of their combination to intensive care data streams.

This work addresses the problem of selecting Tensor Factorization algorithms for the Context-aware Filtering recommendation task using a metalearning approach. The most important challenge of applying metalearning on new problems is the development of useful measures able to characterize the data, i.e. metafeatures. We propose an extensive and exhaustive set of metafeatures to characterize Context-aware Filtering recommendation task. These metafeatures take advantage of the tensor's hierarchical structure via slice operations. The algorithm selection task is addressed as a Label Ranking problem, which ranks the Tensor Factorization algorithms according to their expected performance, rather than simply selecting the algorithm that is expected to obtain the best performance. A comprehensive experimental work is conducted on both levels, baselevel and metalevel (Tensor Factorization and Label Ranking, respectively). The results show that the proposed metafeatures lead to metamodels that tend to rank Tensor Factorization algorithms accurately and that the selected algorithms present high recommendation performance.

Biometric authentication systems have certain limitations. Recent studies have shown that biometric features may change over time, which can entail a decrease in recognition performance of the biometric system. An adaptive biometric system addresses this problem by adapting the biometric reference/template over time, thereby tracking the changes automatically. However, the use of these systems usually requires the adoption of a high threshold value to avoid the inclusion of impostor patterns into the genuine biometric reference. In this study, we hypothesize that score normalization procedures, which have been used to improve the recognition performance of biometric systems through a better refinement of their decision, can also improve the overall performance of adaptive systems. With such a normalization, a better threshold choice could also be made, which would then increase the number of genuine samples used for adaptation. To the best of our knowledge, this is the first investigation towards the use of score normalization to enhance adaptive biometric systems dealing with the change of user features over time. Through a systematic experimental design tested on two behavioral biometric traits, the obtained results indeed support our conjecture. Moreover, the experimental results show that the performance gain brought by adaptation can have a higher overall impact than score normalization alone. (C) 2017 Elsevier Ltd. All rights reserved.

The use of smartphone devices has increased over the last years, as illustrated by the growth in smartphone sales. These devices are currently used for several services, such as bank account access, social networks and storage of personal information

The problem of selecting learning algorithms has been studied by the meta-learning community for more than two decades. One of the most important task for the success of a meta-learning system is gathering data about the learning process. This data is used to induce a (meta) model able to map characteristics extracted from different data sets to the performance of learning algorithms on these data sets. These systems are built under the assumption that the data are generated by a stationary distribution, i.e., a learning algorithm will perform similarly for new data from the same problem. However, many applications generate data whose characteristics can change over time. Therefore, a suitable bias at a given time may become inappropriate at another time. Although meta-learning has been used to continuously select a learning algorithm in data streams, data characterization has received less attention in this context. In this study, we provide a set of guidelines to support the proposal of characteristics able to describe non-stationary data over time. This guidance considers both the order of arrival of the examples and the type of variables involved in the base-level learning. In addition, we analyze the influence of characteristics regarding their dependence on data morphology. Experimental results using real data streams showed the effectiveness of the proposed data characterization general scheme to support algorithm selection by meta-learning systems. Moreover, the dependent meta-features provided crucial information for the success of some meta-models.

BACKGROUND: Hierarchical Multi-Label Classification is a classification task where the classes to be predicted are hierarchically organized. Each instance can be assigned to classes belonging to more than one path in the hierarchy. This scenario is typically found in protein function prediction, considering that each protein may perform many functions, which can be further specialized into sub-functions. We present a new hierarchical multi-label classification method based on multiple neural networks for the task of protein function prediction. A set of neural networks are incrementally training, each being responsible for the prediction of the classes belonging to a given level. RESULTS: The method proposed here is an extension of our previous work. Here we use the neural network output of a level to complement the feature vectors used as input to train the neural network in the next level. We experimentally compare this novel method with several other reduction strategies, showing that it obtains the best predictive performance. Empirical results also show that the proposed method achieves better or comparable predictive performance when compared with state-of-the-art methods for hierarchical multi-label classification in the context of protein function prediction. CONCLUSIONS: The experiments showed that using the output in one level as input to the next level contributed to better classification results. We believe the method was able to learn the relationships between the protein functions during training, and this information was useful for classification. We also identified in which functional classes our method performed better.

BACKGROUND: Discovery of microRNAs (miRNAs) relies on predictive models for characteristic features from miRNA precursors (pre-miRNAs). The short length of miRNA genes and the lack of pronounced sequence features complicate this task. To accommodate the peculiarities of plant and animal miRNAs systems, tools for both systems have evolved differently. However, these tools are biased towards the species for which they were primarily developed and, consequently, their predictive performance on data sets from other species of the same kingdom might be lower. While these biases are intrinsic to the species, their characterization can lead to computational approaches capable of diminishing their negative effect on the accuracy of pre-miRNAs predictive models. We investigate in this study how 45 predictive models induced for data sets from 45 species, distributed in eight subphyla/classes, perform when applied to a species different from the species used in its induction. RESULTS: Our computational experiments show that the separability of pre-miRNAs and pseudo pre-miRNAs instances is species-dependent and no feature set performs well for all species, even within the same subphylum/class. Mitigating this species dependency, we show that an ensemble of classifiers reduced the classification errors for all 45 species. As the ensemble members were obtained using meaningful, and yet computationally viable feature sets, the ensembles also have a lower computational cost than individual classifiers that rely on energy stability parameters, which are of prohibitive computational cost in large scale applications. CONCLUSION: In this study, the combination of multiple pre-miRNAs feature sets and multiple learning biases enhanced the predictive accuracy of pre-miRNAs classifiers of 45 species. This is certainly a promising approach to be incorporated in miRNA discovery tools towards more accurate and less species-dependent tools. The material to reproduce the results from this paper can be downloaded from http://dx.doi.org/10.5281/zenodo.49754 .

International audience

The Traveling Salesman Problem (TSP) is one of the most studied optimization problems.Various metaheuristics (MHs) have been proposed and investigated on many instances of this problem.It is widely accepted that the best MH varies for different instances.Ideally, one should be able to recommend the best MHs for a new TSP instance without having to execute them.However, this is a very difficult task.We address this task by using a meta-learning approach based on label ranking algorithms.These algorithms build a mapping that relates the characteristics of those instances (i.e., the meta-features) with the relative performance (i.e., the ranking) of MHs, based on (meta-)data extracted from TSP instances that have been already solved by those MHs.The success of this approach depends on the quality of the meta-features that describe the instances.In this work, we investigate four different sets of meta-features based on different measurements of the properties of TSP instances: edge and vertex measures, complex network measures, properties from the MHs, and subsampling landmarkers properties.The models are investigated in four different TSP scenarios presenting symmetry and connection strength variations.The experimental results indicate that meta-learning models can accurately predict rankings of MHs for different TSP scenarios.Good solutions for the investigated TSP instances can be obtained from the prediction of rankings of MHs, regardless of the learning algorithm used at the metalevel.The experimental results also show that the definition of the set of meta-features has an important impact on the quality of the solutions obtained.

Classification trees are a powerful tool for mining non-stationary data streams. In these situations, mas sive data are constantly generated at high speed and the underlying target function can change over time. The iadem family of algorithms is based on Hoeffding’s and Chernoff’s bounds and induces online deci sion trees from data streams, but is not able to handle concept drift. This study extends this family to deal with time-changing data streams. The new online algorithm, named iadem-3, performs two main actions in response to a concept drift. Firstly, it resets the variables affected by the change and main tains unbroken the structure of the tree, which allows for changes in which consecutive target functions are very similar. Secondly, it creates alternative models that replace parts of the main tree when they significantly improve the accuracy of the model, thereby rebuilding the main tree if needed. An online change detector and a non-parametric statistical test based on Hoeffding’s bounds are used to guaran tee this significance. A new pruning method is also incorporated in iadem-3, making sure that all split tests previously installed in decision nodes are useful. The learning model is also viewed as an ensem ble of classifiers, and predictions of the main and alternative models are combined to classify unlabeled examples. iadem-3 is empirically compared with various well-known decision tree induction algorithms for concept drift detection. We empirically show that our new algorithm often reaches higher levels of accuracy with smaller decision tree models, maintaining the processing time bounded, irrespective of the number of instances processed.

Germplasm collections are an important strategy for conservation of diversity, a challenge in ecoinformatics. It is common to select a core to represent the genetic diversity of a germplasm collection, aiming to minimize the costs of conservation, while ensuring the maximization of genetic variation. For the problem of finding a core for a germplasm collection, we proposed the use of a constrained multi-objective artificial immune algorithm (MAIS), based on principles of systematic conservation planning (SCP), and incorporating heterozygosity information. Therefore, optimization takes genotypic diversity and variability patterns into account. As a case study, we used Dipteryx alata molecular marker information. We were able to identify within several accessions, the exact entries that should be chosen to preserve species diversity. MAIS presented better performance measure results when compared to NSGA-II. The proposed approach can be used to help construct cores with maximal genetic richness, and also be extended to in situ conservation. As far as we know, this is the first time that an AIS algorithm is applied to the problem of finding a core for a germplasm collection using heterozygosity information as well.

Hierarchical Multi-label Classification (HMC) is a classification task where classes are organized in a hierarchical taxonomy, and instances can be simultaneously classified in more than one class. This paper investigates the HMC problem of classifying proteins in functions organized according to the Gene Ontology hierarchical taxonomy. This is a complex task, since the Gene Ontology hierarchy is organized as a Directed Acyclic Graph with thousands of classes hierarchically represented. We propose a neural network-based method to incorporate label-dependency during learning. The experimental results show that the proposed method achieves competitive results when compared to the state-of-the-art methods from the literature.

In some practical classification problems in which the number of instances of a particular class is much lower/higher than the instances of the other classes, one commonly adopted strategy is to train the classifier over a small, balanced portion of the training data set. Although straightforward, this procedure may discard instances that could be important for the better discrimination of the classes, affecting the performance of the resulting classifier. To address this problem more properly, in this paper we present MOGASamp (after Multiobjective Genetic Sampling) as an adaptive approach that evolves a set of samples of the training data set to induce classifiers with optimized predictive performance. More specifically, MOGASamp evolves balanced portions of the data set as individuals of a multiobjective genetic algorithm aiming at achieving a set of induced classifiers with high levels of diversity and accuracy. Through experiments involving eight binary classification problems with varying levels of class imbalancement, the performance of MOGASamp is compared against the performance of six traditional methods. The overall results show that the proposed method have achieved a noticeable performance in terms of accuracy measures.

Many classification algorithms, such as Neural Networks and Support Vector Machines, have a range of hyper-parameters that may strongly affect the predictive performance of the models induced by them. Hence, it is recommended to define the values of these hyper-parameters using optimization techniques. While these techniques usually converge to a good set of values, they typically have a high computational cost, because many candidate sets of values are evaluated during the optimization process. It is often not clear whether this will result in parameter settings that are significantly better than the default settings. When training time is limited, it may help to know when these parameters should definitely be tuned. In this study, we use meta-learning to predict when optimization techniques are expected to lead to models whose predictive performance is better than those obtained by using default parameter settings. Hence, we can choose to employ optimization techniques only when they are expected to improve performance, thus reducing the overall computational cost. We evaluate these meta-learning techniques on more than one hundred data sets. The experimental results show that it is possible to accurately predict when optimization techniques should be used instead of default values suggested by some machine learning libraries.

Classification is one of the most common machine learning tasks. SVMs have been frequently applied to this task. In general, the values chosen for the hyper-parameters of SVMs affect the performance of their induced predictive models. Several studies use optimization techniques to find a set of hyper-parameter values that induces classifiers with good predictive performance. This paper investigates the hypothesis that a simple Random Search method is sufficient to adjust the hyper-parameters of SVMs. A set of experiments compared the performance of five tuning techniques: three meta-heuristics commonly used, Random Search and Grid Search. The experimental results show that the predictive performance of models using Random Search is equivalent to those obtained using meta-heuristics and Grid Search, but with a lower computational cost.

Enhanced authentication mechanisms are currently needed in several situations. Mainly due to the widespread use of the Internet, data exposure became a source of growing concern. Commonly used login and password credentials may not provide enough security in this scenario, as they may be easily stolen or guessed in some cases. The use of biometrics is a prominent alternative for user authentication, such as by the use of keystroke dynamics. This biometric technology allows the recognition of users by their typing rhythm, which can be performed using data provided by a common keyboard. However, recent work has shown that typing rhythm changes over time. As a result, a static biometric model can become outdated, decreasing the predictive performance of the system. In light of this fact, there is a need for new techniques able to dynamically adapt user models over time. This paper evaluates, in a data stream context, algorithms proposed in the literature for user authentication based on keystroke dynamics. Modifications to these algorithms are also proposed and evaluated. A study of the behaviour of the algorithms over time under several aspects is also performed. According to our experiments, adaptive methods can improve predictive performance of user recognition by keystroke dynamics.

Hierarchical multilabel classification is a complex classification problem where an instance can be assigned to more than one class simultaneously, and these classes are hierarchically organized with superclasses and subclasses, that is, an instance can be classified as belonging to more than one path in the hierarchical structure. This article experimentally analyses the behavior of different decision tree–based hierarchical multilabel classification methods based on the local and global classification approaches. The approaches are compared using distinct hierarchy‐based and distance‐based evaluation measures, when they are applied to a variation of real multilabel and hierarchical datasets' characteristics. Also, the different evaluation measures investigated are compared according to their degrees of consistency, discriminancy, and indifferency. As a result of the experimental analysis, we recommend the use of the global classification approach and suggest the use of the Hierarchical Precision and Hierarchical Recall evaluation measures.

Dealing with distributed data is one of the challenges for clustering, as most clustering techniques require the data to be centralized. One of them, k-means, has been elected as one of the most influential data mining algorithms for being simple, scalable, and easily modifiable to a variety of contexts and application domains. However, exact distributed versions of k-means are still sensitive to the selection of the initial cluster prototypes and require the number of clusters to be specified in advance. Additionally, preserving data privacy among repositories may be a complicating factor. In order to overcome k-means limitations, two different approaches were adopted in this paper: the first obtains a final model identical to the centralized version of the clustering algorithm and the second generates and selects clusters for each distributed data subset and combines them afterwards. It is also described how to apply the algorithms compared while preserving data privacy. The algorithms are compared experimentally from two perspectives: the theoretical one, through asymptotic complexity analyses, and the experimental one, through a comparative evaluation of results obtained from a collection of experiments and statistical tests. The results obtained indicate which algorithm is more suitable for each application scenario.

Noise filters are preprocessing techniques designed to improve data quality in classification tasks by detecting and eliminating examples that contain errors or noise. However, filtering can also remove correct examples and examples containing valuable information, which could be useful for learning. This fact usually implies a margin of improvement on the noise detection accuracy for almost any noise filter. This paper proposes a scheme to improve the performance of noise filters in multi-class classification problems, based on decomposing the dataset into multiple binary subproblems. Decomposition strategies have proven to be successful in improving classification performance in multi-class problems by generating simpler binary subproblems. Similarly, we adapt the principles of the One-vs-One decomposition strategy to noise filtering, making the noise identification process simpler. In order to integrate the filtering results achieved in the binary subproblems, our proposal uses a soft voting approach considering a reliability level based on the aggregation of the noise degree prediction calculated for each binary classifier. The experimental results show that the One-vs-One decomposition strategy usually increases the performance of the noise filters studied, which can detect more accurately the noisy examples.

Data stream mining is an emergent research area that investigates knowledge extraction from large amounts of continuously generated data, produced by non-stationary distribution. Novelty detection, the ability to identify new or previously unknown situations, is a useful ability for learning systems, especially when dealing with data streams, where concepts may appear, disappear, or evolve overtime. There are several studies currently investigating the application of novelty detection techniques in data streams. However, there is no consensus regarding how to evaluate the performance of these techniques. In this study, we propose a new evaluation methodology for multiclass novelty detection in data streams able to deal with: i) unsupervised learning, which generates novelty patterns without an association with the true classes, where one class may be composed of a novelty set, ii) confusion matrix that increases overtime, iii) confusion matrix with a column representing unknown examples, i.e., those not explained by the model, and iv) representation of the evaluation measures overtime. We propose a new methodology to associate the novelty patterns detected by the algorithm, in an unsupervised fashion, with the true classes. Finally, we evaluate the performance of the proposed methodology through the use of known novelty detection algorithms with artificial and real data sets.

The use of pesticides in agriculture is essential to maintain the quality of large-scale production. The spraying of these products by using aircraft speeds up the process and prevents compacting of the soil. However, adverse weather conditions (e.g. The speed and direction of the wind) can impair the effectiveness of the spraying of pesticides in a target crop field. Thus, there is a risk that the pesticide can drift to neighboring crop fields. It is believed that a large amount of all the pesticide used in the world drifts outside of the target crop field and only a small amount is effective in controlling pests. However, with increased precision in the spraying, it is possible to reduce the amount of pesticide used and improve the quality of agricultural products as well as mitigate the risk of environmental damage. With this objective, this paper proposes a methodology based on Particle Swarm Optimization (PSO) for the fine-tuning of control rules during the spraying of pesticides in crop fields. This methodology can be employed with speed and efficiency and achieve good results by taking account of the weather conditions reported by a Wireless Sensor Network (WSN). In this scenario, the UAV becomes a mobile node of the WSN that is able to make personalized decisions for each crop field. The experiments that were carried out show that the optimization methodology proposed is able to reduce the drift of pesticides by fine-tuning of control rules.

Recent advances in cluster analysis highlight the importance of finding multiple meaningful partitions and point out to the need for approaches to evaluate them. They also suggest that the evaluation should consider knowledge of a domain expert. In this paper, we present a visualization method, called PVis1(Partition's Visualizer), that allows the integrated visualization of a collection of partitions. PVis allows to compare the content of a set of partitions. The comparison can be done with respect to priori knowledge provided by an expert. PVis can be useful in the discovery of relevant information to the domain experts performing cluster analysis. In order to illustrate our approach, we give an example of how to perform an exploratory analysis of collections of partitions. In order to do so, we use a well-known dataset from the Bioinformatics domain, regarding molecular classification of cancer.

We present a nature-inspired semi-supervised learning technique based on the flocking formation of certain living species like birds and fishes. Each data item is treated as an individual in the flock. Starting from random directions, each data item moves according to its surrounding items, by getting closer to them (but not too much close) and taking the same direction of motion. Labeled items play special roles, ensuring that data from different classes will belong to different, distant flocks. Experiments on both artificial and benchmark datasets were performed and show its classification accuracy. Despite the rich behavior, we argue that this technique has a sub-quadratic asymptotic time complexity, thus being feasible to be used on large datasets. In order to achieve such performance, a space-partitioning technique is introduced. We also argue that the richness behind this dynamic, self-organizing model is quite robust and may be used to do much more than simply propagating the labels from labeled to unlabeled data. It could be used to determine class overlapping, wrong labeling, etc.

BACKGROUND: Computational discovery of microRNAs (miRNA) is based on pre-determined sets of features from miRNA precursors (pre-miRNA). Some feature sets are composed of sequence-structure patterns commonly found in pre-miRNAs, while others are a combination of more sophisticated RNA features. In this work, we analyze the discriminant power of seven feature sets, which are used in six pre-miRNA prediction tools. The analysis is based on the classification performance achieved with these feature sets for the training algorithms used in these tools. We also evaluate feature discrimination through the F-score and feature importance in the induction of random forests. RESULTS: Small or non-significant differences were found among the estimated classification performances of classifiers induced using sets with diversification of features, despite the wide differences in their dimension. Inspired in these results, we obtained a lower-dimensional feature set, which achieved a sensitivity of 90% and a specificity of 95%. These estimates are within 0.1% of the maximal values obtained with any feature set (SELECT, Section "Results and discussion") while it is 34 times faster to compute. Even compared to another feature set (FS2, see Section "Results and discussion"), which is the computationally least expensive feature set of those from the literature which perform within 0.1% of the maximal values, it is 34 times faster to compute. The results obtained by the tools used as references in the experiments carried out showed that five out of these six tools have lower sensitivity or specificity. CONCLUSION: In miRNA discovery the number of putative miRNA loci is in the order of millions. Analysis of putative pre-miRNAs using a computationally expensive feature set would be wasteful or even unfeasible for large genomes. In this work, we propose a relatively inexpensive feature set and explore most of the learning aspects implemented in current ab-initio pre-miRNA prediction tools, which may lead to the development of efficient ab-initio pre-miRNA discovery tools.The material to reproduce the main results from this paper can be downloaded from http://bioinformatics.rutgers.edu/Static/Software/discriminant.tar.gz.

The ability to detect changes in the data distribution is an important issue in Data Stream mining. Detecting changes in data distribution allows the adaptation of a previously learned model to accommodate the most recent data and, therefore, improve

Decision-tree induction algorithms are widely used in machine learning applications in which the goal is to extract knowledge from data and present it in a graphically intuitive way. The most successful strategy for inducing decision trees is the greedy top-down recursive approach, which has been continuously improved by researchers over the past 40 years. In this paper, we propose a paradigm shift in the research of decision trees: instead of proposing a new manually designed method for inducing decision trees, we propose automatically designing decision-tree induction algorithms tailored to a specific type of classification data set (or application domain). Following recent breakthroughs in the automatic design of machine learning algorithms, we propose a hyper-heuristic evolutionary algorithm called hyper-heuristic evolutionary algorithm for designing decision-tree algorithms (HEAD-DT) that evolves design components of top-down decision-tree induction algorithms. By the end of the evolution, we expect HEAD-DT to generate a new and possibly better decision-tree algorithm for a given application domain. We perform extensive experiments in 35 real-world microarray gene expression data sets to assess the performance of HEAD-DT, and compare it with very well known decision-tree algorithms such as C4.5, CART, and REPTree. Results show that HEAD-DT is capable of generating algorithms that significantly outperform the baseline manually designed decision-tree algorithms regarding predictive accuracy and F-measure.

Hierarchical Multi-Label Classification (HMC) is a challenging task in data mining and machine learning. Each instance in HMC can be classified into two or more classes simultaneously. These classes are structured in a hierarchy, in the form of either a tree or a directed acyclic graph. Therefore, an instance can be assigned to two or more paths from the hierarchical structure, resulting in a complex classification problem with hundreds or thousands of classes. Several methods have been proposed to deal with such problems, including several algorithms based on well-known bio-inspired techniques, such as neural networks, ant colony optimization, and genetic algorithms. In this work, we propose a novel global method called GEHM, which makes use of grammatical evolution for generating HMC rules. In this approach, the grammatical evolution algorithm evolves the antecedents of classification rules, in order to assign instances from a HMC dataset to a probabilistic class vector. Our method is compared to bio-inspired HMC algorithms in protein function prediction datasets. The empirical analysis conducted in this work shows that GEHM outperforms the bio-inspired algorithms with statistical significance, which suggests that grammatical evolution is a promising alternative to deal with hierarchical multi-label classification of biological data.

Software effort estimation is an important task within software engineering. It is widely used for planning and monitoring software project development as a means to deliver the product on time and within budget. Several approaches for generating predictive models from collected metrics have been proposed throughout the years. Machine learning algorithms, in particular, have been widely-employed to this task, bearing in mind their capability of providing accurate predictive models for the analysis of project stakeholders. In this paper, we propose a grammatical evolution approach for software metrics estimation. Our novel algorithm, namely SEEGE, is empirically evaluated on public project data sets, and we compare its performance with state-of-the-art machine learning algorithms such as support vector machines for regression and artificial neural networks, and also to popular linear regression. Results show that SEEGE outperforms the other algorithms considering three different evaluation measures, clearly indicating its effectiveness for the effort estimation task.

This study reports the empirical analysis of a hyper-heuristic evolutionary algorithm that is capable of automatically designing top-down decision-tree induction algorithms. Top-down decision-tree algorithms are of great importance, considering their ability to provide an intuitive and accurate knowledge representation for classification problems. The automatic design of these algorithms seems timely, given the large literature accumulated over more than 40 years of research in the manual design of decision-tree induction algorithms. The proposed hyper-heuristic evolutionary algorithm, HEAD-DT, is extensively tested using 20 public UCI datasets and 10 microarray gene expression datasets. The algorithms automatically designed by HEAD-DT are compared with traditional decision-tree induction algorithms, such as C4.5 and CART. Experimental results show that HEAD-DT is capable of generating algorithms which are significantly more accurate than C4.5 and CART.

Decision tree induction is one of the most employed methods to extract knowledge from data, since the representation of knowledge is very intuitive and easily understandable by humans. The most successful strategy for inducing decision trees, the greedy top-down approach, has been continuously improved by researchers over the years. This work, following recent breakthroughs in the automatic design of machine learning algorithms, proposes a hyper-heuristic evolutionary algorithm for automatically generating decision-tree induction algorithms, named HEAD-DT. We perform extensive experiments in 20 public data sets to assess the performance of HEAD-DT, and we compare it to traditional decision-tree algorithms such as C4.5 and CART. Results show that HEAD-DT can generate algorithms that significantly outperform C4.5 and CART regarding predictive accuracy and F-Measure.

Clustering is an important tool for data analysis, since it allows the exploration of datasets with no or very little prior information. Its main goal is to group a set of data based on their similarity (dissimilarity). A well known mathematical formulation for clustering is the k-medoids problem. Current versions of k-medoids rely on heuristics, with good results reported in the literature. However, few methods that analyze the quality of the partitions found by the heuristics have been proposed. In this paper, we propose a hybrid Lagrangian heuristic for the k-medoids. We compare the performance of the proposed Lagrangian heuristic with other heuristics for the k-medoids problem found in literature. Experimental results presented that the proposed Lagrangian heuristic outperformed the other algorithms.

Support Vector Machines (SVMs) have become a well succeed technique due to the good performance it achieves on different learning problems. However, the performance depends on adjustments on its model. The automatic SVM parameter selection is a way to deal with this. This approach is considered an optimization problem whose goal is to find suitable configuration of parameters which attends some learning problem. This work proposes the use of Particle Swarm Optimization (PSO) to treat the SVM parameter selection problem. As the design of learning systems is inherently a multi-objective optimization problem, a multi-objective PSO (MOPSO) was used to maximize the success rate and minimize the number of support vectors of the model. Moreover, we propose the combination of Meta-Learning (ML) with MOPSO to the cited problem. ML is used to recommend SVM parameters, to a given input problem, based on well-succeeded parameters adopted in previous similar problems. In this combination, initial solutions provided by ML are possibly located in good regions in the search space. Hence, using a reduced number of candidate search points, the search process, to find an adequate solution, would be less expensive. We highlight that, the combination of search algorithms with ML was just studied in the single objective field and the use of MOPSO in this context has not been investigated. In our work, we implemented a prototype in which MOPSO was used to select the values of two SVM parameters for classification problems. In the performed experiments, the proposed solution (MOPSO using ML or Hybrid MOPSO) was compared to a MOPSO with random initialization, obtaining paretos with higher quality on a set of 40 classification problems.

Support Vector Machine (SVM) is a supervised technique, which achieves good performance on different learning problems. However, adjustments on its model are essentials to the SVM work well. Optimization techniques have been used to automatize this process finding suitable configurations of parameters which attends some learning problems. This work utilizes Particle Swarm Optimization (PSO) applied to the SVM parameter selection problem. As the learning systems are essentially a multi-objective problem, a multi-objective PSO (MOPSO) was used to maximize the success rate and minimize the number of support vectors of the model. Nevertheless, we propose the combination of Meta-Learning (ML) with a modified MOPSO which uses the crowding distance mechanism (MOPSO-CDR). In this combination, solutions provided by ML are possibly located in good regions in the search space. Hence, using a reduced number of successful candidates, the search process would converge faster and be less expensive. In our work, we implemented a prototype in which MOPSO-CDR was used to select the values of two SVM parameters for classification problems. In the performed experiments, the proposed solution (MOPSO-CDR using ML) was compared to the MOPSO-CDR with random initialization, obtaining pareto fronts with higher quality on a set of 40 classification problems.

BACKGROUND: This paper addresses the prediction of the free energy of binding of a drug candidate with enzyme InhA associated with Mycobacterium tuberculosis. This problem is found within rational drug design, where interactions between drug candidates and target proteins are verified through molecular docking simulations. In this application, it is important not only to correctly predict the free energy of binding, but also to provide a comprehensible model that could be validated by a domain specialist. Decision-tree induction algorithms have been successfully used in drug-design related applications, specially considering that decision trees are simple to understand, interpret, and validate. There are several decision-tree induction algorithms available for general-use, but each one has a bias that makes it more suitable for a particular data distribution. In this article, we propose and investigate the automatic design of decision-tree induction algorithms tailored to particular drug-enzyme binding data sets. We investigate the performance of our new method for evaluating binding conformations of different drug candidates to InhA, and we analyze our findings with respect to decision tree accuracy, comprehensibility, and biological relevance. RESULTS: The empirical analysis indicates that our method is capable of automatically generating decision-tree induction algorithms that significantly outperform the traditional C4.5 algorithm with respect to both accuracy and comprehensibility. In addition, we provide the biological interpretation of the rules generated by our approach, reinforcing the importance of comprehensible predictive models in this particular bioinformatics application. CONCLUSIONS: We conclude that automatically designing a decision-tree algorithm tailored to molecular docking data is a promising alternative for the prediction of the free energy from the binding of a drug candidate with a flexible-receptor.

This paper presents a survey of evolutionary algorithms that are designed for decision-tree induction. In this context, most of the paper focuses on approaches that evolve decision trees as an alternate heuristics to the traditional top-down divide-and-conquer approach. Additionally, we present some alternative methods that make use of evolutionary algorithms to improve particular components of decision-tree classifiers. The paper's original contributions are the following. First, it provides an up-to-date overview that is fully focused on evolutionary algorithms and decision trees and does not concentrate on any specific evolutionary approach. Second, it provides a taxonomy, which addresses works that evolve decision trees and works that design decision-tree components by the use of evolutionary algorithms. Finally, a number of references are provided that describe applications of evolutionary algorithms for decision-tree induction in different domains. At the end of this paper, we address some important issues and open questions that can be the subject of future research.

Real-world datasets commonly present high dimensional data, which means an increased amount of information. However, this does not always imply an improvement in learning technique performance. Furthermore, some features may be correlated or add unexpected noise, thereby reducing data clustering performance. This has motivated the development of feature selection methods to find the most relevant subset of features to describe data. In this work, we focus on the problem of unsupervised feature selection. The main goal is to define a method to identify the number of features to select after sorting them based on some criterion. This task is done by means of the False Nearest Neighbor technique, which is rooted in chaos theory. Results have shown that this technique gives a good approximate number of features to select. When compared to other techniques, in most of the analyzed cases, it maintains the quality of the generated partitions while selecting fewer features.

Decision tree induction algorithms are well known techniques for assigning objects to predefined categories in a transparent fashion. Most decision tree induction algorithms rely on a greedy top-down recursive strategy for growing the tree, and pruning techniques to avoid overfitting. Even though such a strategy has been quite successful in many problems, it falls short in several others. For instance, there are cases in which the hyper-rectangular surfaces generated by these algorithms can only map the problem description after several sub-sequential partitions, which results in a large and incomprehensible tree. Hence, we propose a new decision tree induction algorithm based on clustering which seeks to provide more accurate models and/or shorter descriptions more comprehensible for the end-user. We do not base our performance analysis solely on the straightforward comparison of our proposed algorithm to baseline methods. Instead, we propose a data-dependent analysis in order to look for evidences which may explain in which situations our algorithm outperforms a well-known decision tree induction algorithm.

Decision tree induction algorithms are widely used in knowledge discovery and data mining, specially in scenarios where model comprehensibility is desired. A variation of the traditional univariate approach is the so-called oblique decision tree, which allows multivariate tests in its non-terminal nodes. Oblique decision trees can model decision boundaries that are oblique to the attribute axes, whereas univariate trees can only perform axis-parallel splits. The majority of the oblique and univariate decision tree induction algorithms perform a top-down strategy for growing the tree, relying on an impurity-based measure for splitting nodes. In this paper, we propose a novel bottom-up algorithm for inducing oblique trees named BUTIA. It does not require an impurity-measure for dividing nodes, since we know a priori the data resulting from each split. For generating the splitting hyperplanes, our algorithm implements a support vector machine solution, and a clustering algorithm is used for generating the initial leaves. We compare BUTIA to traditional univariate and oblique decision tree algorithms, C4.5, CART, OC1 and FT, as well as to a standard SVM implementation, using real gene expression benchmark data. Experimental results show the effectiveness of the proposed approach in several cases.

In Hierarchical Multi-Label Classification problems, each instance can be classified into two or more classes simultaneously, differently from conventional classification. Additionally, the classes are structured in a hierarchy, in the form of either a tree or a directed acyclic graph. Hence, an instance can be assigned to two or more paths from the hierarchical structure, resulting in a complex classification problem with possibly hundreds of classes. Many methods have been proposed to deal with such problems, some of them employing a single classifier to deal with all classes simultaneously (global methods), and others employing many classifiers to decompose the original problem into a set of subproblems (local methods). In this work, we propose a novel local method named HMC-LMLP, which uses one Multi-Layer Perceptron per hierarchical level. The predictions in one level are used as inputs to the network responsible for the predictions in the next level. We make use of two distinct Multi-Layer Perceptron algorithms: Back-propagation and Resilient Back-propagation. In addition, we make use of an error measure specially tailored to multi-label problems for training the networks. Our method is compared to state-of-the-art hierarchical multi-label classification algorithms, in protein function prediction datasets. The experimental results show that our approach presents competitive predictive accuracy, suggesting that artificial neural networks constitute a promising alternative to deal with hierarchical multi-label classification of biological data.

In most classification problems, a classifier assigns a single class to each instance and the classes form a flat (non-hierarchical) structure, without superclasses or subclasses. In hierarchical multilabel classification problems, the classes are hi

Hierarchical Multilabel Classification is a problem where the classes involved are hierarchically structured, and examples can be assigned to more than one class simultaneously at a same hierarchical level. This paper describes and evaluates five different methods for this classification task, based on two approaches, named top-down and one-shot. In the top-down approach, the classification task is carried out by discriminating the classes, level by level, in the hierarchy. In the one-shot approach, the methods consider the whole set of classes at once in the classification. Based on the top-down approach, two new hierarchical methods (with label combination and with label decomposition) and the well-known binary hierarchical method are investigated using SVM classifiers. Other two methods from the literature, named HC4.5 and Clus-HMC, based on the one-shot approach, are also used. The methods are applied to ten biological datasets and evaluated using specific metrics for this kind of classification. The experimental results showed that the proposed methods can improve the classification accuracy.

Nowadays, microarray has become a fairly common tool for simultaneously inspecting the behavior of thousands of genes. Researchers have employed this technique to understand various biological phenomena. One straightforward use of such technology is identifying the class membership of the tissue samples based on their gene expression profiles. This task has been handled by a number of computational methods. In this paper, we provide a comprehensive evaluation of 7 commonly used algorithms over 65 publicly available gene expression datasets. The focus of the study was on comparing the performance of the algorithms in an efficient and sound manner, supporting the prospective users on how to proceed to choose the most adequate classification approach according to their investigation goals.

Learning Classifier Systems (LCSs) are rule-based systems that can be manipulated by a genetic algorithm. LCSs were first designed by Holland to solve classification problems and a lot of effort has been made since then, resulting in a broad number of different algorithms. One of these is called Organizational Classifier System (OCS), a LCSs that tries to organize its rule set favoring good rules to be together in the same organization. However, the proposal of OCS did not include the discovery mechanism. Recently, the OCS was applied to multi-label classification, a type of classification where one instance can have more than one associated label. The authors represented the multi-label classification problem as a default hierarchy and combined the organizational capabilities of OCS together with Smith's default hierarchy formation theory to solve a simple multi-label problem. The purpose of this paper is to extend this idea with the inclusion of a genetic algorithm for the discovery of new rules and present some initial results obtained using the new method. The preliminary results obtained show that the method is comparable to other multi-label techniques. Final discussions present the conclusions of the work and some directions for further research.

Clustering quality or validation indices allow the evaluation of the quality of clustering in order to support the selection of a specific partition or clustering structure in its natural unsupervised environment, where the real solution is unknown o

This paper presents and evaluates an approach to novelty detection that addresses it as the problem of identifying novel concepts in a continuous learning scenario, as an extension to a single-class classification problem. OLINDDA, an OnLIne Novelty

Support Vector Machines were originally proposed to solve two-class classification problems. When they are applied to multiclass classification problems, usually a decomposition approach is followed, in which the original multiclass problem is decomposed into multiple binary sub-problems, whose solutions are afterwards combined. There are several strategies to decompose the multiclass problem. Genetic Algorithms (GAs) can be used to optimize the decomposition according to the performance obtained in the overall multiclass problem solution. This paper presents a study on possible evaluation functions that can be used by the GA in order to evaluate a given decomposition.

This paper presents a survey of evolutionary algorithms designed for clustering tasks. It tries to reflect the profile of this area by focusing more on those subjects that have been given more importance in the literature. In this context, most of the paper is devoted to partitional algorithms that look for hard clusterings of data, though overlapping (i.e., soft and fuzzy) approaches are also covered in the paper. The paper is original in what concerns two main aspects. First, it provides an up-to-date overview that is fully devoted to evolutionary algorithms for clustering, is not limited to any particular kind of evolutionary approach, and comprises advanced topics like multiobjective and ensemble-based evolutionary clustering. Second, it provides a taxonomy that highlights some very important aspects in the context of evolutionary data clustering, namely, fixed or variable number of clusters, cluster-oriented or nonoriented operators, context-sensitive or context-insensitive operators, guided or unguided operators, binary, integer, or real encodings, centroid-based, medoid-based, label-based, tree-based, or graph-based representations, among others. A number of references are provided that describe applications of evolutionary algorithms for clustering in different domains, such as image processing, computer security, and bioinformatics. The paper ends by addressing some important issues and open questions that can be subject of future research.

Data clustering is an important technique to extract and understand relevant information in large data sets. In this paper, a clustering algorithm based on graph theoretic models and community detection in complex networks is proposed. Two steps are involved in this processing: The first step is to represent input data as a network and the second one is to partition the network into subnetworks producing data clusters. In the network partition stage, each node has a randomly assigned initial angle and it is gradually updated according to its neighbors angle agreement. Finally, a stable state is reached and nodes belonging to the same cluster have similar angles. This process is repeated, each time a cluster is chosen and results in an hierarchical divisive clustering. Simulation results show two main advantages of the algorithm: the ability to detect clusters in different shapes, densities and sizes and the ability to generate clusters with different refinement degrees. Besides of these, the proposed algorithm presents high robustness and efficiency in clustering.

The performance of artificial neural networks is largely influenced by the value of their parameters. Among these free parameters, one can mention those related with the network architecture, e.g., number of hidden neurons, number of hidden layers, activation function, and those associated with a learning algorithm, e.g., learning rate. Optimization techniques, often genetic algorithms, have been used to tune neural networks parameter values. Lately, other techniques inspired in Biology have been investigated. In this paper, we compare the influence of different bio-inspired optimization techniques on the accuracy obtained by the networks in the domain of gene expression analysis. The experimental results show the potential of use this techniques for parameter tuning of neural networks.

Machine Learning techniques have been largely applied to the problem of class prediction in microarray data. Nevertheless, current approaches to select appropriate methods for such task often result unsatisfactory in many ways, instigating the need for the development of tools to automate the process. In this context, the authors introduce the use of metalearning in the specific domain of gene expression classification. Experiments with the KNN-ranking method for algorithm recommendation applied for 49 datasets yielded successful results.

Support Vector Machines were originally designed to solve two-class classification problems. When they are applied to multiclass classification problems, the original problem is usually decomposed into multiple binary sub- problems. Afterwards, individual classifiers are induced to solve each of these binary problems. To obtain the final multiclass prediction, the outputs of these binary classifiers generated are combined. Genetic Algorithms can be used to optimize the combination of binary classifiers, defining the decomposition according to the performance obtained in the multiclass problem solution. This paper investigates several evaluation functions that can be used in order to evaluate the performance of the decompositions evolved by genetic algorithms.

There is an increasing concern with the quality of beverage products, like water, milk, wine and coffee. For wine, in particular, the evaluation is usually performed by human tasters, which may require a long time for their training. Moreover, the use of human tasters usually presents high financial and health costs and has a strong subjective component. The quality control of beverages may strongly benefit from the automatic monitoring of their properties by using taste sensors and intelligent systems. This work investigates how the application of a artificial intelligence, more specifically a hybrid case-based system, can lead to an efficient tool for wine quality monitoring. For such, a set of measures extracted by a set of taste sensors is analysed by an intelligent hybrid system. Experimental results show the ability of the proposed system to evaluate wine quality.

A clinical decision support system known as Leuko has been developed for leukemia diagnosis using a naive Bayes classifier. The system is able to recognize six types of white blood cells (WBC), including a malignancy. This paper investigates the use of support vector machines (SVMs) classifiers to recognize WBC for future leukemia diagnosis. Since SVMs are originally designed for the solution of two class problems, several strategies for their extension to this multiclass task are investigated and compared. The experimental results evidence the potential of SVMs to leukemia diagnosis and indicate that a hierarchical tree-based multiclass strategy can be better suited to a future update of the Leuko system.

The article presents the development of a Classifiers Combination based on machine learning techniques (Artificial Neural Networks, Support Vector Machines and C4.5 algorithm) which were able to increase the performance achieved by the Refractive Errors Measurement System (REMS) that analyzes Hartmann-Shack (HS) images from human eyes. The HS images are analyzed in order to extract relevant data for identification of refractive errors (myopia, hypermetropia and astigmatism). Those data are extracted using Gabor wavelets transform and afterwards, machine learning techniques are employed to carry out the image analysis.

Genetic algorithms (GAs) with gene dependent mutation probability applied to non-stationary optimization problems are investigated in this paper. In the problems studied here, the fitness function changes during the search carried out by the GA. In the GA investigated, each gene is associated with an independent mutation probability. The knowledge obtained during the evolution is utilized to update the mutation probabilities. If the modification of a set of genes is useful when the problem changes, the mutation probabilities of these genes are increased. In this way, the search in the solution space is concentrated into regions associated with the genes with higher mutation probabilities. The class of non-stationary problems where this GA can be interesting and its limitations are investigated.

In this paper, a network of coupled chaotic maps for pixel clustering is proposed. Time evolutions of chaotic maps in the network corresponding to a pixel cluster are synchronized with each other. Those synchronized trajectories are desynchronized with respect to the time evolutions of chaotic maps corresponding to other pixel clusters in the same image. A pixel motion mechanism is also introduced, which makes each group of pixels more compact and, consequently, makes the model robust enough to classify ambiguous pixels. Another feature of the proposed model is that the number of pixel clusters does not need to be previously known.

In this paper, a network of coupled chaotic maps for multi-scale image segmentation is proposed. Time evolutions of chaotic maps that correspond to a pixel cluster are synchronized with one another, while this synchronized evolution is desynchronized with respect to time evolution of chaotic maps corresponding to other pixel clusters in the same image. The number of pixel clusters is previously unknown and the adaptive pixel moving technique introduced in the model makes it robust enough to classify ambiguous pixels.

In this paper, a network of coupled chaotic maps for multi-scale image segmentation is proposed. Time evolutions of chaotic maps that correspond to a pixel cluster are synchronized with one another, while this synchronized evolution is desynchronized with respect to time evolution of chaotic maps corresponding to other pixel clusters in the same image. The number of pixel clusters is previously unknown and the adaptive pixel moving technique introduced in the model makes it robust enough to classify ambiguous pixels.

Although Artificial Neural Networks have been satisfactorily employed in several problems, such as clustering, pattern recognition, dynamic systems control and prediction, they still suffer from significant limitations. One of them is that the induced concept representation is not usually comprehensible to humans. Several techniques have been suggested to extract meaningful knowledge from trained networks. This paper proposes the use of symbolic learning algorithms, commonly used by the Machine Learning community, such as C4.5, C4.5rules and CN2, to extract symbolic representations from trained networks. The approach proposed is similar to that used by the Trepan algorithm, which extracts symbolic representations, expressed as decision trees, from trained networks. Experimental results are presented and discussed in order to compare the knowledge extracted from Artificial Neural Networks using the proposed approach and the Trepan approach. Results are compared regarding two aspects: fidelity and comprehensibility.

The article presents the main features of a new environment for case-based reasoning systems development, RaBeCa. The software developed to run under Windows, aims to have an architecture open enough to accept external contributions and comprehend tests allowing integration with other artificial intelligence approaches. The system already uses a neural network. A small introduction to the case-based reasoning paradigm and a brief description of existing analogous systems are shown.

One of the main obstacles to the widespread use of artificial neural networks is the difficulty of adequately defining values for their free parameters. This article discusses how Radial Basis Function (RBF) networks can have their parameters defined by genetic algorithms. For such, it presents an overall view of the problems involved and the different approaches used to genetically optimize RBF networks. A new strategy to optimize RBF networks using genetic algorithms is proposed, which includes new representation, crossover operator and the use of a multiobjective optimization criterion. Experiments using a benchmark problem are performed and the results achieved using this model are compared to those achieved by other approaches.

Vibration signals analysis is considered as an appropriate diagnosis method for detecting faults. Several techniques have been used for detecting vibration signals. In this work, artificial neural networks (ANN) were used to predict vibration signals using process parameters measured in a turbine. The ANN models can be viewed as "black-boxes". One way to improve their comprehensibility is to use a symbolic model. As a first step in this direction, a hybrid rule-based regression model was also tested.

This paper aims to analyze the advantages of using constructive algorithms to achieve a functional neural network topology automatically. A constructive neural network simulator, named Kipu, was implemented in order to accomplish the design and analysis of constructive algorithms. Four constructive algorithms were modified to allow the use of real-valued data. Experiments using three pattern recognition tasks were conducted in the Kipu simulator and the results suggest a better functioning of the constructive algorithms compared with the fixed topology MLP network trained with backpropagation algorithm.

The use of a linguistic representation for expressing knowledge acquired by learning systems is an important issue as regards to user understanding. Under this assumption, and to make sure that these systems will be welcome and used, several techniques have been developed by the artificial intelligence community, under both the symbolic and the connectionist approaches. This work discusses and investigates three knowledge extraction techniques based on these approaches. The first two techniques, the C4.5 and CN2 symbolic learning algorithms, extract knowledge directly from the data set. The last technique, the TREPAN algorithm extracts knowledge from a previously trained neural network. The CN2 algorithm induces if...then rules from a given data set. The C4.5 algorithm extracts decision trees, although it can also extract ordered rules, from the data set. Decision trees are also the knowledge representation used by the TREPAN algorithm.

This paper presents a completely integrated Boolean neural architecture, where a selforganizing Boolean neural network (SOFT) is used as a front-end processor to a feedforward Boolean network based on goal-seeking principles (GSN f). This paper will evaluate the advantages of the integrated SOFT-GSN f over GSN f by showing its increased effectiveness in an optical character recognition task.