Explainable Artificial Intelligence methods have gained significant traction for their ability to elucidate the decision-making processes of black-box models, particularly in high-stakes fields such as healthcare and finance. Among these, Local Interpretable Model-agnostic Explanations (LIME) stands out as a widely adopted post-hoc, model-agnostic approach that interprets black-box predictions by constructing an interpretable surrogate model on perturbed instances to approximate the local behavior of the original model around a given instance. However, the effectiveness of LIME can depend on the quality of the training data used by the black-box model. When trained on limited or low-quality data, the black-box model may yield inaccurate predictions for perturbed samples, resulting in poorly defined local decision boundaries and consequently unreliable explanations. This limitation is especially problematic in data-scarce settings. To overcome this challenge, we propose TransLIME, a novel end-to-end explainable transfer learning framework that improves the local fidelity and stability of LIME on limited tabular datasets by transferring relevant explainability knowledge from a related auxiliary source domain with a shifted distribution. Also, in TransLIME, only representative source prototype explanations obtained through clustering are transferred to the target domain, thereby reducing cross-domain exposure of both data and explanatory information during transfer. Experimental evaluations on real-world datasets demonstrate the effectiveness of the proposed framework in improving explanation quality in target domains with limited data.

Comprehending biological reproduction and cellular metabolism is facilitated by the Enzyme Commission, which matches protein sequences to the biochemical reactions they catalyse through EC numbers. In recent years, several methods have been proposed for predicting enzyme function. However, these methods still encounter challenges. Firstly, traditional methods for manually designing enzyme features are complex and cumbersome, lacking an effective generalized method for embedding enzyme sequences. Secondly, the distribution gap between different enzymes is significant, which resulting in existing methods struggling to predict multilevel enzyme functions. Thirdly, traditional enzyme function prediction models only extract single view feature of enzyme, so there is still room for further improving the ability of these models to extract enzyme data. To address these challenges, a new multilevel enzyme function prediction model (SMENET) based on multi-view semantics is proposed. This method uses protein large language model to extract semantic information. Subsequently, this semantic information is fed into multiple information extraction network modules, followed by using Biologic Sematic Attention to integrate these views' information. Finally, a multi-view adaptive fusion network is designed to extract the best common representation between multiple semantic views. Extensive experiments were conducted on multiple datasets to validate the effectiveness of SMENET.

Effectively extracting discriminative information from multi-view data remains a key challenge in multi-view learning. Existing methods typically focus on exploring inter-view consistency via linear or nonlinear transformation. While nonlinear methods tend to yield better performance, their limited transparency and interpretability limit their practical applications. Takagi-Sugeno-Kang Fuzzy Systems (TSK-FS), as a rule-based model with high interpretability, have been applied to multi-view tasks. However, prior methods either rely solely on antecedent components for nonlinear modeling or integrate deep neural networks into the consequent part, thereby compromising model interpretability. To address these challenges, we propose Fuzzy Rule-guided Multi-view Differentiable Representation Learning (FRMVDRL). Specifically, in our framework, antecedent parameters of TSK-FS are first used to map data into highdimensional fuzzy space. Then during the learning of consequent parameters, a dual information extraction mechanism is proposed to jointly capture shared knowledge across views and view-specific knowledge. Moreover, a second-order geometric structure preservation mechanism is constructed to exploit structural information at both the instance and the instance-pair level. To enhance discriminability of the learned representations, a biorthogonal constraint alongside a Shannon entropy mechanism is introduced. Finally, to balance model performance and interpretability, we introduce a novel multi-view differentiable optimization strategy that incorporates learnable parameters to expand the solution space while preserving the structure of traditional optimization. Extensive experiments on benchmark multi-view datasets demonstrate the effectiveness of the FRMVDRL.

Generative models (GMs), particularly large language models (LLMs), have garnered significant attention in machine learning and artificial intelligence for their ability to generate new data by learning the statistical properties of training data and creating data that resemble the original data. This capability offers a wide range of applications across various domains. However, the complex structures and numerous model parameters of GMs obscure the input-output processes and complicate the understanding and control of the outputs. Moreover, the purely data-driven learning mechanism limits GMs' abilities to acquire broader knowledge. There remains substantial potential for enhancing the robustness and generalization capabilities of GMs. In this work, we leverage fuzzy system, a classical modeling method, to combine both data-driven and knowledge-driven mechanisms for generative tasks. We propose a novel generative fuzzy system framework, named GenFS, which integrates the deep learning capabilities of GMs with the term-based interpretability and dual-driven mechanisms of fuzzy systems. Specifically, we propose an end-to-end GenFS-based model for sequence generation, called FuzzyS2S. A series of test studies were conducted on 12 datasets, covering three distinct categories of generative tasks: machine translation, code generation, and summary generation. The results demonstrate that FuzzyS2S outperforms the transformer in terms of accuracy and fluency. Furthermore, it exhibits better performance than state-of-the-art models T5 and CodeT5 for some application scenarios.

Representation learning is a key area in machine learning and deep learning, focusing on extracting meaningful features to support downstream tasks such as classification and clustering. Current mainstream representation learning methods primarily rely on nonlinear data mining techniques such as kernel methods and deep neural networks (DNNs) to extract abstract knowledge from complex datasets. However, most of them are "black-box" methods, lacking transparency and interpretability in the learning process, which constrain their practical utility. To this end, this article introduces a novel representation learning method called fuzzy rule-based differentiable representation learning (FRDRL), which is grounded in an interpretable fuzzy rule-based model. Specifically, it is built upon the Takagi-Sugeno-Kang fuzzy system (TSK-FS) to map input data to a high-dimensional fuzzy feature space through the antecedent part of the TSK-FS. Subsequently, a novel differentiable optimization method is proposed for learning in the consequent part, which preserves interpretability and transparency while effectively capturing nonlinear relationships in the data. By retaining the essence of traditional optimization and parameterizing key components as differentiable modules, the method improves performance without sacrificing interpretability. Moreover, a second-order geometry preservation strategy is incorporated to further improve robustness. Extensive evaluations conducted on various benchmark datasets validate the superiority of the proposed method. The source codes are available at https://github.com/BBKing49/FEDRL.

Urban spatial tasks span multiple levels, ranging from area-level analysis, crime prediction, and taxi demand forecasting to POI-level tasks such as new store recommendation. Urban knowledge graphs (UrbanKGs) can enhance these tasks by integrating structured urban knowledge. However, existing studies face two main issues: most research uses task-specific UrbanKGs for corresponding single-level predictions, and public UrbanKGs contain only coarse-grained administrative areas, lacking the rich semantic and spatial relationships required for multi-level tasks. We propose a Hierarchically Structured UrbanKG Dataset (HUSK) with an intermediate functional zone layer that bridges and enriches the understanding across multiple levels, and evaluate it on three area-level and three POI-level tasks, showing accuracy improvements over single-view baselines.

Explainable Artificial Intelligence (XAI) methods, such as Local Interpretable Model-Agnostic Explanations (LIME), have advanced the interpretability of black-box machine learning models by approximating their behavior locally using interpretable surrogate models. However, LIME's inherent randomness in perturbation and sampling can lead to locality and instability issues, especially in scenarios with limited training data. In such cases, data scarcity can result in the generation of unrealistic variations and samples that deviate from the true data manifold. Consequently, the surrogate model may fail to accurately approximate the complex decision boundary of the original model. To address these challenges, we propose a novel Instance-based Transfer Learning LIME framework (ITL-LIME) that enhances explanation fidelity and stability in data-constrained environments. ITL-LIME introduces instance transfer learning into the LIME framework by leveraging relevant real instances from a related source domain to aid the explanation process in the target domain. Specifically, we employ clustering to partition the source domain into clusters with representative prototypes. Instead of generating random perturbations, our method retrieves pertinent real source instances from the source cluster whose prototype is most similar to the target instance. These are then combined with the target instance's neighboring real instances. To define a compact locality, we further construct a contrastive learning-based encoder as a weighting mechanism to assign weights to the instances from the combined set based on their proximity to the target instance. Finally, these weighted source and target instances are used to train the surrogate model for explanation purposes. Experimental evaluation with real-world datasets demonstrates that ITL-LIME greatly improves the stability and fidelity of LIME explanations in scenarios with limited data. Our code is available at https://github.com/rehanrazaa/ITL-LIME.

We propose a novel framework for the statistical analysis of genus-zero 4D surfaces, i.e., 3D surfaces that deform and evolve over time. This problem is particularly challenging due to the arbitrary parameterizations of these surfaces and their varying deformation speeds, necessitating effective spatiotemporal registration. Traditionally, 4D surfaces are discretized, in space and time, before computing their spatiotemporal registrations, geodesics, and statistics. However, this approach may result in suboptimal solutions and, as we demonstrate in this paper, is not necessary. In contrast, we treat 4D surfaces as continuous functions in both space and time. We introduce Dynamic Spherical Neural Surfaces (D-SNS), an efficient smooth and continuous spatiotemporal representation for genus-0 4D surfaces. We then demonstrate how to perform core 4D shape analysis tasks such as spatiotemporal registration, geodesics computation, and mean 4D shape estimation, directly on these continuous representations without upfront discretization and meshing. By integrating neural representations with classical Riemannian geometry and statistical shape analysis techniques, we provide the building blocks for enabling full functional shape analysis. We demonstrate the efficiency of the framework on 4D human and face datasets. The source code and additional results are available at https://4d-dsns.github.io/DSNS/.

The rapid advancement of medical imaging technologies has led to an exponential increase in medical image data, making efficient retrieval from large-scale datasets critical for improving diagnostic accuracy and speed. However, two key challenges hinder this process: first, the presence of uncertain and subtle lesions in medical images that are often difficult to discern, and second, class imbalance across different case types within medical image databases. These inherent challenges significantly degrade the performance of existing hashing algorithms. In recent years, methods based on the Takagi–Sugeno–Kang fuzzy system (TSK-FS) have shown promising performance in medical image modeling. Inspired by these advances, this article proposes a novel fuzzy hashing network (FHN) based on TSK-FS to enhance retrieval performance by effectively handling both uncertainty and data imbalance in medical imaging. The FHN first introduces a novel fuzzification mechanism that incorporates the concept of a self-attention mechanism to effectively capture the complex underlying features in medical images, thereby enhancing the data discriminability in fuzzy spaces. Meanwhile, a new consequent parameter learning mechanism is developed for defuzzification by introducing the Transformer network, which aims to improve the inference efficiency and generalization capability of the FHN. Based on these two mechanisms, FHN's capability of analyzing and handling uncertain data is significantly enhanced. Furthermore, a novel hash center loss is designed to capture global relationships while emphasizing local structural information, thereby improving the handling of imbalanced data and significantly enhancing retrieval performance.

The federated Fuzzy C-means (federated FCM) extends the traditional Fuzzy C-means (FCM) to the federated learning (FL) scenario, aiming to address the data privacy preservation issue of soft clustering in distributed environments. However, a significant challenge persists with existing federated FCM algorithms, i.e., they struggle to converge effectively in complex heterogeneous scenarios, leading to unstable clustering outcomes. Here the complex heterogeneous scenarios stem from the combination of non-independently and identically distributed (non-IID) data across different clients (statistical heterogeneity), coupled with the involvement of only some clients in each iteration (systematic heterogeneity). While prior research has attempted to address the impact of statistical heterogeneity in FL scenarios, it has overlooked the issue of system heterogeneity. In response, this paper proposes a novel federated FCM algorithm (SC-FFCM) that remains robust even in such complex heterogeneous scenarios. Firstly, the client-side clustering module of SC-FFCM adopts a Gradient-Based FCM algorithm, facilitating corrections to the direction of local optimization. Secondly, the algorithm introduces a control variates technique to rectify update bias during the iteration process, thereby mitigating the adverse effects of random client sampling and non-IID data distribution on the algorithm convergence. Finally, the proposed algorithm approximates the ideal federated FCM algorithm. Experimental studies verify the effectiveness of the proposed method. The source code of the proposed SC-FFCM algorithm is available from the following website https://github.com/Creazy-MR/SC-FFCM.

Smartphone sensor-based human activity recognition (SSHAR) generally deals with three main steps: 1) raw signal collection; 2) feature extraction; and 3) human activity classification. This study focuses on an interpretable human activity classification method to enhance SSHAR's very applicability for the application scenarios like healthcare services and personal biometric signature. To this end, by taking Takagi–Sugeno–Kang fuzzy classifiers as the subclassifiers, a novel interpretable ensemble fuzzy classifier FINE is proposed to provide linguistically interpretable fuzzy rules for classification, strong generalization and scalability for SSHAR. Since each subclassifier of FINE works on its bootstrapping subspace of original features and then is combined without an explicit aggregation, FINE has the following characteristics: 1) the diversities among all the subclassifiers are assured; 2) more generalization capabilities than the corresponding structure of each subclassifier on all the input features is justified; 3) its incremental learning can be implemented through only training an incremental subclassifier or training FINE only on incremental data. The experimental results demonstrate that FINE not only keeps at least comparable to and even better than most of the comparative methods in terms of testing performance and training time but also has both linguistically interpretable fuzzy rules and fast incremental learning capability.

Fuzzy systems (FSs) have been widely utilized in diverse domains, such as pattern recognition, intelligent control, data mining, and bioinformatics due to their strong interpretation and learning abilities. Traditionally, FSs have mainly been applied to model Euclidean data. However, with the emergence of scenarios involving graph data, such as social networks and traffic route maps, which inherently possess non-Euclidean structures, there is a need to develop FS modeling methods suitable for graph data while retaining the advantages of traditional FSs. This article presents a novel FS called graph fuzzy system (GFS) specifically designed for modeling whole graph data. The concepts, modeling framework, and construction algorithms are systematically developed. First, the article defines GFS-related concepts, including the graph fuzzy rule base, graph fuzzy sets, and graph consequent processing unit (GCPU). Second, the learning framework for GFS is proposed. It includes a novel K-Means with graph similarity measure clustering approach (KM-GSM) for generating antecedents in GFS and a new consequent parameters learning algorithm based on graph neural network (GNN). Moreover, three different versions of the GFS implementation algorithm are developed and thoroughly evaluated through experiments on various graph prediction datasets. The results demonstrate that the proposed GFS inherits the advantages of mainstream GNNs methods and conventional FSs methods while achieving superior performance in whole graph prediction compared to existing approaches.

Stacking-based takagi-sugeno-kang (TSK) fuzzy system ensemble has been successfully applied to imbalanced data classification. However, there still exist many challenges that need to be further addressed. For example, during stacking, augmenting output variables into the input feature space reduces the interpretability of antecedents of fuzzy rules. During sampling for balancing, discovering informative samples usually only relies on training samples, which may reduce generalizability. More importantly, there is no theory to support the reliability of stacking. To address the aforementioned challenges, in this study, we propose a deep reconciled and self-paced TSK fuzzy system ensemble framework termed D-RSP-TSKE for imbalanced data classification. Compared with the existing ensemble frameworks, its superiorities can be exhibited from the following three aspects. First, in the first layer, we use random undersampling to generate a class-balanced training set to train an initial zero-order TSK fuzzy classifier. Based on the TSK fuzzy classifier, then we define classifier-specific and testing-compatible sample sensitivity to discover informative (high-sensitive) samples and design a reconciled and self-paced sampling approach to balance the minority class for the training of the following layers. Second, to improve the interpretability of antecedents of fuzzy rules, we propose to transfer the output variables from antecedents to consequents through equivalent mathematical transformations while keeping the final output unchanged. These transferred output variables are interpreted as the dynamic fuzzy rule confidence. Third, furthermore, we engage in a comprehensive theoretical examination of our stacking-based ensemble to elucidate the underlying mechanisms that enable the stacking strategy to consistently deliver superior performance. We conduct tests and comparisons on 7 artificial datasets and 30 real-world datasets to evaluate D-RSP-TSKE. The experimental results demonstrate the effectiveness and interpretability of D-RSP-TSKE for imbalanced data classification.

Multiview fuzzy c-means clustering has garnered significant attention in recent years, leading to the development of various multiview fuzzy clustering algorithms. However, existing algorithms still exhibit room for improvement. First, most existing algorithms only utilize the shallow information of the view data and fail to delve into the mining and utilization of deeper representations. Second, existing algorithms tend to extract common representations among the views first and then implement clustering separately, which may lack a collaborative linkage between two tasks. Finally, multiview clustering algorithms based on representation learning often overlook the importance of effectively preserving similarity information within the views. To address these limitations, we propose a novel algorithm called pseudolabel enhanced multiview deep concept factorization fuzzy clustering (PE-MV-DCFCM). The algorithm first introduces a deep concept factorization method to uncover the deep information of the view data. Subsequently, it employs pseudolabel learning to preserve intraview similarity information during the learning of common representations among the views, based on non-negative matrix factorization. Finally, this algorithm integrates deep concept factorization, representation learning, and fuzzy clustering into a unified framework to enhance the collaboration among the various substeps of the algorithm. Experiments on several benchmark datasets show that the proposed PE-MV-DCFCM algorithm outperformed other state-of-the-art algorithms.

Extreme learning machines (ELMs) are commonly adopted as base learners in ensemble methods due to unstable results and fast learning speed. However, most existing ensemble structures require explicit aggregation of ELM base learners’ results before the final decision. This study proposes a novel Biological Ensemble ELMs called BE-ELM from a biological perspective for the first time and exploit the superiority of assembling multiple ELM base learners in parallel without the need for explicit aggregation, and thus simplifying the learning procedure. BE-ELM’s structure is inspired by recent MIT neuroscience findings of brain learning mechanisms. The expression of the analytical solution of BE-ELM is similar to that of basic ELMs, which allows it to inherit their fast learning speed in the ensemble structure. Moreover, BE-ELM also has the added advantage of superior performance. Our theoretical analysis shows that BE-ELM consisting of multiple ELM base learners built on subsets of the original features is equivalent to a more complex ELM on the original feature space. We prove that BE-ELM, even without any explicit aggregation, can guarantee enhanced generalization capabilities. To confirm our findings, we conducted extensive experiments on various datasets. The results demonstrate that BE-ELM outperforms traditional ELMs and other state-of-the-art ensemble ELMs in terms of generalization performance on most datasets. These findings suggest that BE-ELM has the potential to improve prediction outcomes in practical applications.

Partially labeled and skewed datasets are common in many applications including healthcare, due to the high costs and time constraints of data collection and annotation. However, training machine learning classifiers on such data can undermine their prediction performances. In this paper, we propose a novel classifier to address this problem by focusing on the Area Under the Curve (AUC), which is widely recognized as a more robust performance metric for skewed datasets than other metrics such as accuracy and error rate. We introduce a new classifier called PSVM-AUC Maximizer (PSVM-AUCMax) which is based on Proximal Support Vector Machines (PSVM) and directly maximizes a new AUC-based metric in its learning objective. PSVM-AUCMax has several merits. First, by directly integrating the maximization of the proposed AUC-based metric, PSVM-AUCMax can be proved to have the enhanced generalization capability on the partially labeled and skewed dataset. Second, it simplifies the model selection process with fewer tuning hyperparameters. Third, PSVM-AUCMax’s analytical solution remains the same form as traditional PSVM, preserving its advantages such as fast incremental updating in incremental learning scenarios. The efficacy of PSVM-AUCMax has been demonstrated through extensive experiments on several public datasets and a healthcare case study using data collected at the US Mayo Clinic. In the healthcare case study, we utilized PSVM-AUCMax to develop a clinical prediction model for forecasting composite outcomes in hospitalized COVID-19 patients which yielded promising results.

Abstract Gated Recurrent Neural Networks (RNNs) such as LSTM and GRU have been highly effective in handling sequential time series data in recent years. Although Gated RNNs have an inherent ability to learn complex temporal dynamics, there is potential for further enhancement by enabling these deep learning networks to directly use time information to recognise time-dependent patterns in data and identify important segments of time. Synonymous with time series data in real-world applications are missing values, which often reduce a model’s ability to perform predictive tasks. Historically, missing values have been handled by simple or complex imputation techniques as well as machine learning models, which manage the missing values in the prediction layers. However, these methods do not attempt to identify the significance of data segments and therefore are susceptible to poor imputation values or model degradation from high missing value rates. This paper develops Cyclic Gate enhanced recurrent neural networks with learnt waveform parameters to automatically identify important data segments within a time series and neglect unimportant segments. By using the proposed networks, the negative impact of missing data on model performance is mitigated through the addition of customised cyclic opening and closing gate operations. Cyclic Gate Recurrent Neural Networks are tested on several sequential time series datasets for classification performance. For long sequence datasets with high rates of missing values, Cyclic Gate enhanced RNN models achieve higher performance metrics than standard gated recurrent neural network models, conventional non-neural network machine learning algorithms and current state of the art RNN cell variants.

The COVID-19 patient data for composite outcome prediction often comes with class imbalance issues, i.e., only a small group of patients develop severe composite events after hospital admission, while the rest do not. An ideal COVID-19 composite outcome prediction model should possess strong imbalanced learning capability. The model also should have fewer tuning hyperparameters to ensure good usability and exhibit potential for fast incremental learning. Towards this goal, this study proposes a novel imbalanced learning approach called Imbalanced maximizing-Area Under the Curve (AUC) Proximal Support Vector Machine (ImAUC-PSVM) by the means of classical PSVM to predict the composite outcomes of hospitalized COVID-19 patients within 30 days of hospitalization. ImAUC-PSVM offers the following merits: (1) it incorporates straightforward AUC maximization into the objective function, resulting in fewer parameters to tune. This makes it suitable for handling imbalanced COVID-19 data with a simplified training process. (2) Theoretical derivations reveal that ImAUC-PSVM has the same analytical solution form as PSVM, thus inheriting the advantages of PSVM for handling incremental COVID-19 cases through fast incremental updating. We built and internally and externally validated our proposed classifier using real COVID-19 patient data obtained from three separate sites of Mayo Clinic in the United States. Additionally, we validated it on public datasets using various performance metrics. Experimental results demonstrate that ImAUC-PSVM outperforms other methods in most cases, showcasing its potential to assist clinicians in triaging COVID-19 patients at an early stage in hospital settings, as well as in other prediction applications.

This article presents a new deep cross-output knowledge transfer approach based on least-squares support vector machines, called DCOT-LS-SVMs. Its aim is to improve the generalizability of least-squares support vector machines (LS-SVMs) while avoiding the complicated parameter tuning process that occurs in many kernel machines. The proposed approach has two significant characteristics: 1) DCOT-LS-SVMs is inspired by a stacked hierarchical architecture that combines several layer-by-layer LS-SVMs modules. The module that forms the higher layer has additional input features that consider the predictions from all previous modules and 2) cross-output knowledge transfer is used to leverage knowledge from the predictions of the previous module to improve the learning process in the current module. With this approach, the model's parameters, such as a tradeoff parameter C and a kernel width δ , can be randomly assigned to each module in order to greatly simplify the learning process. Moreover, DCOT-LS-SVMs is able to autonomously and quickly decide the extent of the cross-output knowledge transfer between adjacent modules through a fast leave-one-out cross-validation strategy. In addition, we present an imbalanced version of DCOT-LS-SVMs, called IDCOT-LS-SVMs, given that imbalanced datasets are common in real-world scenarios. The effectiveness of the proposed approaches is demonstrated through a comparison with five comparative methods on UCI datasets and with a case study on the diagnosis of prostate cancer.

Existing research reveals that the misclassification rate for imbalanced data depends heavily on the problematic areas due to the existence of small disjoints, class overlap, borderline, and rare data samples. In this study, by stacking zero-order Takagi-Sugeno-Kang (TSK) fuzzy subclassifiers on the minority class and its problematic areas in the deep ensemble, a novel deep-ensemble-level-based TSK fuzzy classifier (IDE-TSK-FC) for imbalanced data classification tasks is presented to achieve both promising classification performance and high interpretability of zero-order TSK fuzzy classifiers. Simultaneously, according to the stacked generalization principle, the proposed classifier lifts up oversampling from the data level to the deep ensemble level with a guarantee of enhanced generalization capability for class imbalance learning. In the structure of IDE-TSK-FC, the first interpretable zero-order TSK fuzzy subclassifier is built on the original training dataset. After that, several successive zero-order TSK fuzzy subclassifiers are stacked layer by layer on the newly identified problematic areas from the original training dataset plus the corresponding interpretable predictions obtained by the averaging strategy on all previous layers. IDE-TSK-FC simply takes the classical K -nearest neighboring algorithm at each layer to identify its problematic area that consists of the minority samples and its surrounding K majority neighbors. After randomly neglecting certain input features and randomly selecting the five Gaussian membership functions for all the chosen input features and the augmented feature in the premise of each fuzzy rule, each subclassifier can be quickly obtained by using the least learning machine to determine the consequent part of each fuzzy rule. The experimental results on both the public datasets and a real-world healthcare dataset demonstrate IDE-TSK-FC's superiority in class imbalanced learning.

Extreme learning machines (ELMs) has been theoretically and experimentally proved to achieve promising performance at a fast learning speed for supervised classification tasks. However, it does not perform well on imbalanced binary classification tasks and tends to get biased toward the majority class. Besides, since a large amount of training data with labels are not always available in the real world, there is an urgent demand to develop an efficient semi-supervised version of ELM for imbalanced binary classification tasks. In this article, owing to the distinct insensitivity of area under the ROC curve (AUC) to both class skews and changes of class distributions, we focus the study on integrating AUC maximization into the ELM framework to tackle with imbalanced binary classification tasks well. By demystifying the AUC metric with the ELM framework, we develop a new AUC-based ELM called AUC-ELM for imbalanced binary classification, which essentially is revealed to be equivalent to an ELM on another transformed data space. Accordingly, its semi-supervised version called SAUC-ELM is also developed. Both AUC-ELM and SAUC-ELM have the distinctive merits: 1) they share the advantage of ELM in both generalization capability and training efficiency, and further uniquely tailored for imbalanced binary classification tasks and 2) in contrast to the existing imbalanced variants of ELM, such as class-specific cost regulation ELM and semi-supervised ELM, they have fewer parameters to tune, thereby reducing the computational cost for model selection. Experiments on a heap of datasets show that both AUC-ELM and SAUC-ELM outperform the other comparative methods in terms of both classification performance and training speed.

The performance of a classifier might greatly deteriorate due to missing data. Many different techniques to handle this problem have been developed. In this paper, we solve the problem of missing data using a novel transfer learning perspective and show that when an additive least squares support vector machine (LS-SVM) is adopted, model transfer learning can be used to enhance the classification performance on incomplete training datasets. A novel transfer-based additive LS-SVM classifier is accordingly proposed. This method also simultaneously determines the influence of classification errors caused by each incomplete sample using a fast leave-one-out cross validation strategy, as an alternative way to clean the training data to further improve the data quality. The proposed method has been applied to seven public datasets. The experimental results indicate that the proposed method achieves at least comparable, if not better, performance than case deletion, mean imputation, and k -nearest neighbor imputation methods, followed by the standard LS-SVM and support vector machine classifiers. Moreover, a case study on a community healthcare dataset using the proposed method is presented in detail, which particularly highlights the contributions and benefits of the proposed method to this real-world application.

The additive kernel least squares support vector machine (AK-LS-SVM) has been well used in classification tasks due to its inherent advantages. For example, additive kernels work extremely well for some specific tasks, such as computer vision classification, medical research, and some specialized scenarios. Moreover, the analytical solution using AK-LS-SVM can formulate leave-one-out cross-validation error estimates in a closed form for parameter tuning, which drastically reduces the computational cost and guarantee the generalization performance especially on small and medium datasets. However, AK-LS-SVM still faces two main challenges: 1) improving the classification performance of AK-LS-SVM and 2) saving time when performing a grid search for model selection. Inspired by the stacked generalization principle and the transfer learning mechanism, a layer-by-layer combination of AK-LS-SVM classifiers embedded with transfer learning is proposed in this paper. This new classifier is called deep transfer additive kernel least square support vector machine (DTA-LS-SVM) which overcomes these two challenges. Also, considering that imbalanced datasets are involved in many real-world scenarios, especially for medical data analysis, the deep-transfer element is extended to compensate for this imbalance, thus leading to the development of another new classifier iDTA-LS-SVM. In the hierarchical structure of both DTA-LS-SVM and iDTA-LS-SVM, each layer has an AK-LS-SVM and the predictions from the previous layer act as an additional input feature for the current layer. Importantly, transfer learning is also embedded to guarantee generalization consistency between the adjacent layers. Moreover, both iDTA-LS-SVM and DTA-LS-SVM can ensure the minimal leaveone-out error by using the proposed fast leave-one-out cross validation strategy on the training set in each layer. We compared the proposed classifiers DTA-LS-SVM and iDTA-LS-SVM with the traditional LS-SVM and SVM using additive kernels on seven public UCI datasets and one real world dataset. The experimental results show that both DTA-LS-SVM and iDTALS-SVM exhibit better generalization performance and faster learning speed.

Missing data is a common issue in community health and epidemiological studies. Direct removal of samples with missing data can lead to reduced sample size and information bias, which deteriorates the significance of the results. While data imputation methods are available to deal with missing data, they are limited in performance and could introduce noises into the dataset. Instead of data imputation, a novel method based on additive least square support vector machine (LS-SVM) is proposed in this paper for predictive modeling when the input features of the model contain missing data. The method also determines simultaneously the influence of the features with missing values on the classification accuracy using the fast leave-one-out cross-validation strategy. The performance of the method is evaluated by applying it to predict the quality of life (QOL) of elderly people using health data collected in the community. The dataset involves demographics, socioeconomic status, health history, and the outcomes of health assessments of 444 community-dwelling elderly people, with 5% to 60% of data missing in some of the input features. The QOL is measured using a standard questionnaire of the World Health Organization. Results show that the proposed method outperforms four conventional methods for handling missing data-case deletion, feature deletion, mean imputation, and K-nearest neighbor imputation, with the average QOL prediction accuracy reaching 0.7418. It is potentially a promising technique for tackling missing data in community health research and other applications.

Many medical applications face a situation that the on-hand data cannot fully fit an existing predictive model or on-line tool, since these models or tools only use the most common predictors and the other valuable features collected in the current scenario are not considered altogether. On the other hand, the training data in the current scenario is not sufficient to learn a predictive model effectively yet. In order to overcome these problems and construct an efficient classifier, for these real situations in medical fields, in this work we present an approach based on the least squares support vector machine (LS-SVM), which utilizes a transfer learning framework to make maximum use of the data and guarantee its enhanced generalization capability. The proposed approach is capable of effectively learning a target domain with limited samples by relying on the probabilistic outputs from the other previously learned model using a heterogeneous method in the source domain. Moreover, it autonomously and quickly decides how much output knowledge to transfer from source domain to the target one using a fast leave-one-out cross validation strategy. This approach is applied on a real-world clinical dataset to predict 5-year mortality of bladder cancer patients after radical cystectomy, and the experimental results indicate that the proposed method can achieve better performances compared to traditional machine learning methods, consistently showing the potential of the proposed method under the circumstances with insufficient data.

Recognition of epileptic electroencephalogram (EEG) signals using machine learning techniques is becoming popular. In general, the construction of intelligent epileptic EEG recognition system involves two steps. First, an appropriate feature extraction method is applied to obtain representative features from the original raw EEG signals. Second, an effective intelligent model is trained based on the extracted features. However, there exist two major challenges in the process: 1) it is nontrivial to determine the appropriate feature extraction method to be used; 2) although many classical machine learning methods have been used for epileptic EEG recognition, most of them are “black box” approaches and more interpretable methods are desirable. To address these two challenges, a new epileptic EEG recognition method based on a multiview learning framework and fuzzy system modeling is proposed. First, multiview EEG data are generated by employing different feature extraction methods to obtain the features from different views of the signals. Second, the classical Takagi-Sugeno-Kang fuzzy system (TSK-FS) is introduced as an easy-to-interpret recognition model to develop a multiview TSK-FS method, called MV-TSK-FS, to identify epileptic EEG signals. For the proposed MV-TSK-FS, the importance of each view, i.e., the importance of each feature extraction method, can be evaluated according to the weighting of each view, and consequently the final decision can be made based on the weighted outputs of different views. Experimental results indicate that the MV-TSK-FS is a promising method when compared with the state-of-the-art algorithms.