Abstract Accurate crop yield forecasting requires identifying informative predictors from high-dimensional meteorological time series with multiple time lags. Such settings often lead to overfitting, reduced interpretability, and increased computational cost. This paper proposes a two-phase feature-selection framework, termed CS $$_\text {MRMR}$$ MRMR -DWES $$_R$$ R , to address these challenges in agricultural time-series regression. In the first phase, a supervised filter ranks features using correlation statistics with a maximum relevance-minimum redundancy criterion, retaining a compact subset of informative variables. In the second phase, a wrapper-based discrete weighted evolutionary strategy automatically selects the optimal feature subset by directly optimizing the predictive performance of a SARIMAX regression model. The proposed framework is evaluated on 20 years of monthly avocado and mango yield data from the Axarquía region of Málaga in Spain under multiple lag configurations. This feature selection might slightly change over time and depends entirely on the crop type and geographical location. The proposed CS $$_\text {MRMR}$$ MRMR -DWES $$_\text {R}$$ R framework inherently requires high-performance computing (HPC) capabilities due to the scale and complexity of the agricultural time-series datasets under study. Experimental results show that CS $$_\text {MRMR}$$ MRMR -DWES $$_R$$ R improves forecasting accuracy while reducing feature dimensionality and computational cost compared with baseline and existing feature-selection methods. These findings highlight the effectiveness of the proposed approach for high-dimensional agricultural time-series analysis.

Late detection of plant diseases leads to irreparable losses for farmers, threatening global food security, economic stability, and environmental sustainability. This research introduces the Multi-Objective Genetic Algorithm Explainer (MOGAE), a novel model-agnostic local explainer for image data aimed at the early detection of citrus diseases. MOGAE enhances eXplainable Artificial Intelligence (XAI) by leveraging the Non-dominated Sorting Genetic Algorithm II (NSGA-II) with an adaptive Bit Flip Mutation (BFM) incorporating densify and sparsify operators to adjust superpixel granularity automatically. This innovative approach simplifies the explanation process by eliminating several critical hyperparameters required by traditional methods like Local Interpretable Model-Agnostic Explanations (LIME). To develop the citrus disease classification model, we preprocess the leaf dataset through stratified data splitting, oversampling, and augmentation techniques, then fine-tuning a pre-trained Residual Network 50 layers (ResNet50) model. MOGAE’s effectiveness is demonstrated through comparative analyses with the Ensemble-based Genetic Algorithm Explainer (EGAE) and LIME, showing superior accuracy and interpretability using criteria such as numeric accuracy of explanation and Number of Function Evaluations (NFE). We assess accuracy both intuitively and numerically by measuring the Euclidean distance between expert-provided explanations and those generated by the explainer. The appendix also includes an extensive evaluation of MOGAE on the melanoma dataset, highlighting its versatility and robustness in other domains. The related implementation code for the fine-tuned ResNet50 and MOGAE is available at https://github.com/KhaosResearch/Plant-disease-explanation . • MOGAE eliminates several critical hyperparameters LIME requires, simplifying the explanation process. • MOGAE delivers more accurate and interpretable explanations than LIME, enhancing the reliability of the model’s decisions. • The integration of adaptive Bit Flip Mutation with densify and sparsify operators in NSGA-II significantly improves the algorithm’s exploration capabilities within MOGAE. • Comprehensive evaluations of MOGAE are conducted using both the citrus leaf diseases dataset and the melanoma dataset, demonstrating its versatility and effectiveness across different domains.

Feature selection has a prominent role in high-dimensional datasets to increase classification accuracy, decrease the learning algorithm computational time, and present the most informative features to decision-makers. This paper proposes a two-stage hybrid feature selection for high dimensional medical datasets: Maximum Pattern Recognition - Multi-objective Discrete Evolution Strategy (MPR-MDES). MPR is a rapid filter ranker that significantly outperforms existing frequency-based rankers in recognizing non-linear patterns, effectively eliminating a majority of non-informative features. Then, the wrapper Multi-objective Discrete Evolution Strategy(MDES) uses the remaining features and obtains sets of solutions which are automatically presented to decision-makers.The experiments conducted on large medical datasets demonstrate that MPR-MDES achieves considerable improvements compared to state-of-the-art methods, in terms of both classification accuracy and dimensionality reduction. In this sense, the proposal successfully performs when presenting informative feature sets to decision-makers. The implementation is available on https://github.com/KhaosResearch/MPR-MDES.

Feature selection in small-sample high-dimensional datasets enhances classification accuracy and reduces computational time for model training. This paper introduces the filter Classification Error Impurity (CEI) as a frequency-based ranker that improves upon existing methods by better identifying non-linear patterns. According to this, the top features identified by the ensemble of CEI, along with Mutual Information (MI) and Fisher Ratio (FR), form the feature space utilized by the Adaptive Genetic Algorithm with External Repository (AGAwER) to identify the optimal feature combination in a wrapper approach. AGAwER leverages an external repository, incorporating the best solutions to enrich the Genetic Algorithm’s (GA) population, thus promoting diversity and enhancing exploration. As a result, a hybrid method called CMF-AGAwER is proposed, which surpasses existing modern feature selection methods. The implementation and data are accessible on GitHub at https://github.com/KhaosResearch/CMF-AGAwER.

In the last years, Learning Management systems (LMSs) are acquiring great importance in online education, since they offer flexible integration platforms for organising a vast amount of learning resources, as well as for establishing effective communication channels between teachers and learners, at any direction. These online platforms are then attracting an increasing number of users that continuously access, download/upload resources and interact each other during their teaching/learning processes, which is even accelerating by the breakout of COVID-19. In this context, academic institutions are generating large volumes of learning-related data that can be analysed for supporting teachers in lesson, course or faculty degree planning, as well as administrations in university strategic level. However, managing such amount of data, usually coming from multiple heterogeneous sources and with attributes sometimes reflecting semantic inconsistencies, constitutes an emerging challenge, so they require common definition and integration schemes to easily fuse them, with the aim of efficiently feeding machine learning models. In this regard, semantic web technologies arise as a useful framework for the semantic integration of multi-source e-learning data, allowing the consolidation, linkage and advanced querying in a systematic way. With this motivation, the e-LION (e-Learning Integration ONtology) semantic model is proposed for the first time in this work to operate as data consolidation approach of different e-learning knowledge-bases, hence leading to enrich on-top analysis. For demonstration purposes, the proposed ontological model is populated with real-world private and public data sources from different LMSs referring university courses of the Software Engineering degree of the University of Malaga (Spain) and the Open University Learning. In this regard, a set of four case studies are worked for validation, which comprise advance semantic querying of data for feeding predictive modelling and time-series forecasting of students’ interactions according to their final grades, as well as the generation of SWRL reasoning rules for student’s behaviour classification. The results are promising and lead to the possible use of e-LION as ontological mediator scheme for the integration of new future semantic models in the domain of e-learning.

Earth Observation (EO) based on Remote Sensing (RS) is gaining importance nowadays, since it offers a well-grounded technological framework for the development of advanced applications in multiple domains, such as climate change, precision agriculture, smart urbanism, safety, and many others. This promotes the continuous generation of data-driven software facilities oriented to advanced processing, analysis and visualization, which often offer enhanced computing capabilities. Nevertheless, the development of knowledge-driven approaches is still an open challenge in remote sensing, besides they provide human experts with domain knowledge representation, support for data standardization and semantic integration of sources, which indeed enhance the construction of advanced on-top applications. To this end, the use of ontologies and web semantic technologies have shown high success in knowledge representation in many fields, in which the Earth Observation is not an exception. However, as argued by the research community, there is large room for improvement in the specific case of remote sensing, where ontologies that consider the special nature and structure of different satellital and airborne data products are demanded. This article addresses, in first instance, part of this need by proposing a semantic model for the consolidation, integration, reasoning and linking of data (and meta-data), in the context of satellital remote sensing products for EO. With this objective, an OWL ontology has been developed and an RDF repository has been generated to allow advanced SPARQL querying. Although the proposal has been designed to consider remote sensing data products in general, the current study is mainly focused on the Sentinel 2 satellite mission from the Copernicus Programme of the European Space Agency (ESA). Four different use cases are showcased to check potentials of the proposed semantic model in terms of ontology integration, federated querying, data analysis and reasoning.

Modern applications of Big Data are transcending from being scalable solutions of data processing and analysis, to now provide advanced functionalities with the ability to exploit and understand the underpinning knowledge. This change is promoting the development of tools in the intersection of data processing, data analysis, knowledge extraction and management. In this paper, we propose TITAN, a software platform for managing all the life cycle of science workflows from deployment to execution in the context of Big Data applications. This platform is characterised by a design and operation mode driven by semantics at different levels: data sources, problem domain and workflow components. The proposed platform is developed upon an ontological framework of meta-data consistently managing processes and models and taking advantage of domain knowledge. TITAN comprises a well-grounded stack of Big Data technologies including Apache Kafka for inter-component communication, Apache Avro for data serialisation and Apache Spark for data analytics. A series of use cases are conducted for validation, which comprises workflow composition and semantic meta-data management in academic and real-world fields of human activity recognition and land use monitoring from satellite images.

BACKGROUND: The analysis of gene expression levels is used in many clinical studies to know how patients evolve or to find new genetic biomarkers that could help in clinical decision making. However, the techniques and software available for these analyses are not intended for physicians, but for geneticists. However, enabling physicians to make initial discoveries on these data would benefit in the clinical assay development. RESULTS: Melanoma is a highly immunogenic tumor. Therefore, in recent years physicians have incorporated immune system altering drugs into their therapeutic arsenal against this disease, revolutionizing the treatment of patients with an advanced stage of the cancer. This has led us to explore and deepen our knowledge of the immunology surrounding melanoma, in order to optimize the approach. Within this project we have developed a database for collecting relevant clinical information for melanoma patients, including the storage of patient gene expression levels obtained from the NanoString platform (several samples are taken from each patient). The Immune Profiling Panel is used in this case. This database is being exploited through the analysis of the different expression profiles of the patients. This analysis is being done with Python, and a parallel version of the algorithms is available with Apache Spark to provide scalability as needed. CONCLUSIONS: VIGLA-M, the visual analysis tool for gene expression levels in melanoma patients is available at http://khaos.uma.es/melanoma/ . The platform with real clinical data can be accessed with a demo user account, physician, using password physician_test_7634 (if you encounter any problems, contact us at this email address: mailto: [email protected]). The initial results of the analysis of gene expression levels using these tools are providing first insights into the patients' evolution. These results are promising, but larger scale tests must be developed once new patients have been sequenced, to discover new genetic biomarkers.

MOTIVATION: Multiple sequence alignment (MSA) is an NP-complete optimization problem found in computational biology, where the time complexity of finding an optimal alignment raises exponentially along with the number of sequences and their lengths. Additionally, to assess the quality of a MSA, a number of objectives can be taken into account, such as maximizing the sum-of-pairs, maximizing the totally conserved columns, minimizing the number of gaps, or maximizing structural information based scores such as STRIKE. An approach to deal with MSA problems is to use multi-objective metaheuristics, which are non-exact stochastic optimization methods that can produce high quality solutions to complex problems having two or more objectives to be optimized at the same time. Our motivation is to provide a multi-objective metaheuristic for MSA that can run in parallel taking advantage of multi-core-based computers. RESULTS: The software tool we propose, called M2Align (Multi-objective Multiple Sequence Alignment), is a parallel and more efficient version of the three-objective optimizer for sequence alignments MO-SAStrE, able of reducing the algorithm computing time by exploiting the computing capabilities of common multi-core CPU clusters. Our performance evaluation over datasets of the benchmark BAliBASE (v3.0) shows that significant time reductions can be achieved by using up to 20 cores. Even in sequential executions, M2Align is faster than MO-SAStrE, thanks to the encoding method used for the alignments. AVAILABILITY AND IMPLEMENTATION: M2Align is an open source project hosted in GitHub, where the source code and sample datasets can be freely obtained: https://github.com/KhaosResearch/M2Align. CONTACT: [email protected]. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Multiple sequence alignment (MSA) plays a core role in most bioinformatics studies and provides a framework for the analysis of evolution in biological systems. The MSA problem consists in finding an optimal alignment of three or more sequences of nucleotides or amino acids. Different scores have been defined to assess the quality of MSA solutions, so the problem can be formulated as a multiobjective optimization problem. The number of proposals focused on this approach in the literature is scarce, and most of the works take as base algorithm the NSGA-II metaheuristic. So, there is a lack of a study involving a set of representative multiobjective metaheuristics to deal with this complex problem. Our main goal in this paper is to carry out such study. We propose a biobjective formulation for the MSA and perform an exhaustive comparative study of six multiobjective algorithms. We have considered a number of problems taken from the benchmark BAliBASE (v3.0). Our experiments reveal that the classic NSGA-II algorithm and MOCell, a cellular metaheuristic, provide the best overall performance.

In last years, enhancing the vehicular traffic flow becomes a mandatory task to minimize the impact of polluting emissions and unsustainable fuel consumption in our cities. Smart Mobility optimisation emerges then, with the goal of improving the traffic management in the city. With this aim, we propose in this paper an optimisation strategy based on swarm intelligence to find efficient cycle programs for traffic lights deployed in large urban areas. In concrete, in this work we focus on the improvement of the traffic flow with the global purpose of reducing contaminant emissions (CO 2 and NO x ) and fuel consumption in the analyzed areas. For the sake of standardization, we follow European Union reference framework for traffic emissions, called HandBook Emission FActors (HBEFA). As a case study, we have concentrated in two extensive urban areas in the cities of Malaga and Seville (in Spain). After several comparisons between different optimisation techniques (Differential Evolution and Random Search), as well as other solutions provided by experts, our proposal is shown to obtain significant reductions of fuel consumption and gas emissions.

Optimal staging of traffic lights, and in particular optimal light cycle programs, is a crucial task in present day cities with potential benefits in terms of energy consumption, traffic flow management, pedestrian safety, and environmental issues. Nevertheless, very few publications in the current literature tackle this problem by means of automatic intelligent systems, and, when they do, they focus on limited areas with elementary traffic light schedules. In this paper, we propose an optimization approach in which a particle swarm optimizer (PSO) is able to find successful traffic light cycle programs. The solutions obtained are simulated with simulator of urban mobility, a well-known microscopic traffic simulator. For this study, we have tested two large and heterogeneous metropolitan areas with hundreds of traffic lights located in the cities of Bahía Blanca in Argentina (American style) and Málaga in Spain (European style). Our algorithm is shown to obtain efficient traffic light cycle programs for both kinds of cities. In comparison with expertly predefined cycle programs (close to real ones), our PSO achieved quantitative improvements for the two main objectives: 1) the number of vehicles that reach their destination and 2) the overall journey time.

In a previous work, it was empirically shown that certain numbers of informants different from the standard "two" and the expensive "all" may provide the Particle Swarm Optimization (PSO) with new essential information about the search landscape, leading this algorithm to perform more accurately than other existing versions of it. Here, we extend this study by analyzing the internal behavior of PSO from the point of view of the evolvability. Our motivation is to find evidences of why such number of 6+/-2 informant particles, perform better than other neighborhood formulations of PSO. For this task, we have evaluated different combinations of informants for an extensive set of problem functions. Using fitness-distance correlation and fitness-fitness cloud analyses we have tested the accuracy of the resulting landscape characterizations. The results suggest that, in spite of certain deviation to the global optimum, a number of 6 informants in PSO can generate new improved particles for a longer time, even in complex problems with multi-funnel landscapes.

In the standard particle swarm optimization (PSO), a new \nparticle’s position is generated using two main informant elements: \nthe best position the particle has found so far and the \nbest performer among its neighbors. In fully informed PSO, \neach particle is influenced by all the remaining ones in the \nswarm, or by a series of neighbors structured in static \ntopologies (ring, square, or clusters). In this paper, we generalize \nand analyze the number of informants that take part in \nthe calculation of new particles. Our aim is to discover if a \nquasi-optimal number of informants exists for a given \nproblem. The experimental results seem to suggest that 6 to 8 \ninformants could provide our PSO with higher chances of \nsuccess in continuous optimization for well-known benchmarks.

In current modern cities, the increasing number of traffic lights that control the vehicular traffic flow requires a highly complex scheduling. Thousands of red lights, that have to be optimally programmed, are nowadays operating in congested urban areas. Therefore, automatic intelligent systems are indispensable tools for optimally tackling this task. In this work, we propose a Swarm Intelligence approach that, coupled with the SUMO traffic simulator, is able to find successful cycle programs of traffic lights for large urban areas. In concrete, we have focused on a metropolitan area of the city downtown of Córdoba (in Spain). The experiments and comparisons with other techniques reveal that our proposed approach obtains significant profits in terms of traffic flow and global trip time.

To evaluate the search capabilities of a multiobjective algorithm, the usual approach is to choose a benchmark of known problems, to perform a fixed number of function evaluations, and to apply a set of quality indicators. However, while real problems could have hundreds or even thousands of decision variables, current benchmarks are normally adopted with relatively few decision variables (normally from 10 to 30). Furthermore, performing a constant number of evaluations does not provide information about the effort required by an algorithm to get a satisfactory set of solutions; this information would also be of interest in real scenarios, where evaluating the functions defining the problem can be computationally expensive. In this paper, we study the effect of parameter scalability in a number of state-of-the-art multiobjective metaheuristics. We adopt a benchmark of parameter-wise scalable problems (the Zitzler-Deb-Thiele test suite) and analyze the behavior of eight multiobjective metaheuristics on these test problems when using a number of decision variables that range from 8 up to 2048. By using the hypervolume indicator as a stopping condition, we also analyze the computational effort required by each algorithm in order to reach the Pareto front. We conclude that the two analyzed algorithms based on particle swarm optimization and differential evolution yield the best overall results.

In this work we compare the use of a particle swarm optimization (PSO) and a genetic algorithm (GA) (both augmented with support vector machines SVM) for the classification of high dimensional microarray data. Both algorithms are used for finding small samples of informative genes amongst thousands of them. A SVM classifier with 10-fold cross-validation is applied in order to validate and evaluate the provided solutions. A first contribution is to prove that PSOSVMis able to find interesting genes and to provide classification competitive performance. Specifically, a new version of PSO, called geometric PSO, is empirically evaluated for the first time in this work. In this sense, a comparison of this approach with a new GASVMand also with other existing methods of literature is provided. A second important contribution consists in the actual discovery of new and challenging results on six public datasets identifying significant in the development of a variety of cancers (leukemia, breast, colon, ovarian, prostate, and lung).

This paper presents a new distributed Differential Evolution (dDE) algorithm and evaluates it according to the standard procedure set in the special session of continuous optimization of CEC’05. We statistically validate our results in continuous optimization versus several other efficient techniques. Our distributed Differential Evolution is simple and accurate, at the same time amenable, to be applied to a wide variety of problems, especially for noisy and multimodal functions.

In this work we compare the use of a particle swarm optimization (PSO) and a genetic algorithm (GA) (both augmented with support vector machines SVM) for the classification of high dimensional microarray data. Both algorithms are used for finding small samples of informative genes amongst thousands of them. A SVM classifier with 10- fold cross-validation is applied in order to validate and evaluate the provided solutions. A first contribution is to prove that PSOsvm is able to find interesting genes and to provide classification competitive performance. Specifically, a new version of PSO, called Geometric PSO, is empirically evaluated for the first time in this work using a binary representation in Hamming space. In this sense, a comparison of this approach with a new GAsvm and also with other existing methods of literature is provided. A second important contribution consists in the actual discovery of new and challenging results on six public datasets identifying significant in the development of a variety of cancers (leukemia, breast, colon, ovarian, prostate, and lung).

No abstract available.

No abstract available.