Search

In machine learning field, Graph-based Semi-Supervised Learning (GSSL) has recently attracted much attention and many researchers have proposed a number of different methods. GSSL generally constructs a k nearest neighbors graph to explore manifold structure which may improve learning performance of GSSL. If one uses an inappropriate graph to learn a semi-supervised classifier, the performance of the classifier may be worse than that of supervised learning (SL) only trained by labeled samples. Hence, it is worthy to design a safe version to broaden the application area of GSSL. In this paper, we introduce a Safety-aware GSSL (SaGSSL) method which can adaptively select the good graphs and learn a safe semi-supervised classifier simultaneously. The basic assumption is that a graph has a high quality if the sample margin obtained by GSSL with the graph is larger than that obtained by SL. By identifying the high-quality graphs and setting the corresponding weights large, the predictions of our algorithm will approach to those of GSSL with the graphs. Meanwhile, the weights of the low-quality graphs should be small and the predictions of our algorithm will be close to those of SL. Hence the degeneration probability will be reduced and our algorithm is expected to realize the goal of safe exploitation of different graphs. Experimental results on several datasets show that our algorithm can simultaneously implement the graph selection and safely exploit the unlabeled samples. [ABSTRACT FROM AUTHOR]

Feature selection plays an important role in the classification problems associated with expert and intelligent systems. The central idea behind feature selection is to identify important input features in order to reduce the dimensionality of the input space while maintaining or improving classification performance. Traditional feature selection approaches were designed to handle either categorical or numerical features, but not the mix of both that often arises in real datasets. In this paper, we propose a novel feature selection algorithm for classifying mixed-type data, based on a rough set model, called feature selection for mixed-type data with feature space decomposition (FSMSD). This can handle both categorical and numerical features by utilizing rough set theory with a heterogeneous Euclidean-overlap metric, and can be applied to mixed-type data. It also uses feature space decomposition to preserve the properties of multi-valued categorical features, thereby reducing information loss and preserving the features’ physical meaning. The proposed algorithm was compared with four benchmark methods using real mixed-type datasets and biomedical datasets, and its performance was promising, indicating that it will be helpful to users of expert and intelligent systems. [ABSTRACT FROM AUTHOR]

Inductive Logic Programming (ILP) and Relational Data Mining (RDM) address the task of inducing models or patterns from multi-relational data. One of the established approaches to RDM is propositionalization, characterized by transforming a relational database into a single-table representation. This paper presents a propositionalization technique called wordification which can be seen as a transformation of a relational database into a corpus of text documents. Wordification constructs simple, easy to understand features, acting as words in the transformed Bag-Of-Words representation. This paper presents the wordification methodology, together with an experimental comparison of several propositionalization approaches on seven relational datasets. The main advantages of the approach are: simple implementation, accuracy comparable to competitive methods, and greater scalability, as it performs several times faster on all experimental databases. Furthermore, the wordification methodology and the evaluation procedure are implemented as executable workflows in the web-based data mining platform ClowdFlows. The implemented workflows include also several other ILP and RDM algorithms, as well as the utility components that were added to the platform to enable access to these techniques to a wider research audience. [ABSTRACT FROM AUTHOR]

Distance metric learning is the task that aims to automate this process of learning task-specific distance functions in a supervised manner. In this paper, we study how to learn a Mahalanobis distance metric that can improve nearest neighbor classification. Our paper makes two contributions. First, we propose a novel framework named CLE_LMNN for Mahalanobis distance learning. CLE_LMNN builds on a recently proposed framework known as large margin nearest neighbor (LMNN) classification. Compared with LMNN, CLE_LMNN learns a Mahalanobis distance in a fine-grained way by first partitioning the labeled examples into subsets. As shown by our experiments, this fine-grained learning way is inclined to obtain a Mahalanobis distance more suitable for classification. Second, we present a novel algorithm named CLE for clustering labeled examples. Different from traditional unsupervised clustering algorithms, CLE fully employ the class information of labeled examples to effectively partition the set of examples with the same class into different subsets. To evaluate our proposed framework, we conduct extensive experiments on three real datasets. The experimental results show the effectiveness of CLE_LMNN when applied to classification. [ABSTRACT FROM AUTHOR]

Abstract: Even though rapid advances in intelligent firearm identification have been made in recently years, the major practical and theoretical problems are still unsolved. From the practical point of view, capturing high quality images from ballistics specimen is a difficult task. From the theoretical point of view, extracting the descriptive features from projectile and cartridge images is an open research question in firearm identification. The aim of this paper is to address the research issues with respect to feature extraction and intelligent ballistics recognition. In this paper, different image processing techniques are employed for digitizing the ballistics images. Due to some segments in an image systematically distributed by the image’s geometrical circular center, the existing moment invariants however cannot extract the required pattern features for intelligent recognition. This paper presents the novel feature set called circle moment invariants to overcome the shortcoming of existing moment invariants. In addition, an intelligent system is designed for classifying and evaluating the extracted features of ballistics images. The experimental results indicate that the proposed approach and feature criteria are capable of classifying the cartridge images very efficiently and effectively. Consequently, the circle moment invariants are proved to be the adequate descriptors for describing the pattern features in cartridge images. [Copyright &y& Elsevier]

Multi-label classification has a fundamental challenge of finding the interdependence of labels. Stacking is one of the most prominent methods for the same. It uses the full dependency, i.e., it utilizes dependence among all the labels. The stacking method uses classifiers in two layers, where the first layer's classifiers assume no interdependence among labels. It leads to the first layer classifiers being error-prone. Also, it diminishes the performance of the second layer's classifiers because they are fed with the first layer's classifiers' outputs. Further, there is no correlation utilization among the second layer's classifiers. This paper addresses the poor performance of the first layer's classifiers and no correlation utilization at the second layer. It proposes a method called Stacked Chaining (StaC) that uses interdependence of labels in a chaining fashion at both layers. This way, it resolves the poor performance of the first layer and also incorporates label correlation at the second layer. The StaC method is compared with other prevalent state-of-the-art methods, including Classifier Chain and Stacking, using ten benchmark multi-label datasets. Experimental results show that the proposed StaC method achieves better performance than other state-of-the-art methods for different performance evaluation parameters. • A novel method StaC with three variants StaCv1, StaCv2, StaCv3 is proposed. • The StaC ensures full label dependence and augments ground truths during training. • The labels' order is as per the prediction performance of classifiers for labels. • Experimentation approve that these methods improve the classification performance. [ABSTRACT FROM AUTHOR]

• Presented an extensive overview on different stages of electrocardiogram analysis. • Summarized the most notable developments of smart health monitoring systems. • Discussed the latest hardware implementations of real-time ECG monitoring systems. • Identified challenges and limitations for various automatic ECG diagnosis systems. • Outlines the future vision of next generation ECG monitoring systems for healthcare. Over the last decade, cardiovascular diseases (CVD's) are the leading cause of death globally. Early prediction of CVD's can help in reducing the complications of high-risk patients. The electrocardiogram (ECG) is an efficient aiding tool, provides accurate information about various cardiac conditions of the human heart. Evaluation and interpretation of ECG signal has become the major goal in current research to identify and mitigate risky cardiovascular conditions. The ECG signal is efficiently analysed and classified using signal processing techniques, ranging from traditional to machine learning approaches and its subbranches, such as deep learning are used for the early detection and diagnosis of cardiac conditions and arrhythmias. The development of novel types of body sensors increases the need for automated, low-cost, real-time, and efficient ECG monitoring systems that can be used at home or in ambulatory settings. IoT in healthcare industry enables the doctors and specialists to diagnose the patient status remotely in a smart and efficient manner. Motivated with all the research done in this area for early diagnosis of cardiac abnormalities and save the life of high-risk patients, we surveyed numerous research works reported in the literature and provide a glance on several aspects of ECG signal monitoring systems. We first introduce the step-by-step framework of ECG signal analysis starting data acquisition to classification, describes each stage analysis of both traditional and advanced machine learning models that have been reported in the literature. We provide a deep discussion on ECG signal acquisition, pre-recorded clinical ECG data taken from the databases, signal processing and denoising, detection of feature based on feature engineering, and signal classification along with the comparative assessment amongst reviewed studies. This work also presents the detailed analysis of smart health care system that employs portable and wearable devices, as well as the Internet of Things (IoT) and wireless technologies, as an innovative medium for data transmission that allows for remote access to a patient's health status at lower costs and with greater efficiency. Moreover, this work provides an extensive knowledge on various hardware platforms that have been adopted for the development of biomedical processors including microcontrollers, FPGA and ASIC. Additionally, the challenges and limitations are discussed in this research field and directions for future work are suggested. This is the first review paper of its kind to present a comprehensive discussion on all aspects of cardiovascular disease monitoring systems in one place. [ABSTRACT FROM AUTHOR]

Classification plays a crucial role in mining the evolving data streams. The concept drift and concept evolution are the major issues of data streams classification, which greatly affect the classification performance. Most existing works of concept drift and evolution are supervised in nature, where labeling the data is time and resource consuming. In this paper, for the evolving data streams, a semi-supervised classification approach using partially labeled data is proposed. Firstly, an ensemble model dynamically maintains a series of micro-clusters to capture the concept drift. The ensemble model processes the instances in an online fashion rather than chunk-based. Secondly, the concept evolution detection module is constructed to detect the outliers by the local density. The module examines the current buffer to capture the class emergence in data streams with complex class distribution. For improving the execution efficiency of emerging class detection without compromising performance, several constructive strategies are adopted, including removing part of the buffer and selectively executing sample generation. The extensive experiments are constructed about the popular data streams sets and the processed industry data streams, whose results indicate the practicality and effectiveness of the proposed approach for the classification of evolving data streams. • For the evolving data streams, a semi-supervised classification approach using partial labeled data is proposed. • The ensemble model processes the instances in an online fashion rather than chunk-based. • An ensemble model dynamically maintains a series of micro-clusters to capture the concept drift. • The concept evolution detection module is constructed to detect the outliers by local density. [ABSTRACT FROM AUTHOR]

The accurate classification of Typhoon Intensity (TI) based on cloud patterns in satellite images is crucial for effective disaster warning and management. However, this task is particularly challenging due to the striking visual similarities among different sub-classes of typhoons, coupled with a long-tailed distribution of class occurrences. Existing deep learning methods often struggle with biased classification stemming from imbalanced datasets and face difficulty discerning subtle differences between categories. This paper proposes a novel solution to these challenges through a hybrid framework that integrates contrastive learning and classifier learning. Contrastive learning is employed to increase the separation of similar sub-classes within the feature space, while classifier learning aims to train a discriminative and unbiased classifier. The proposed approach is evaluated extensively on the publicly available DeepTI dataset, demonstrating enhanced performance for both prominent and less frequent classes. The model achieves the 70.88% accuracy for seven categories typhoon intensity classification. The code implementation is at https://github.com/chen-huanxin/contrastive-hybrid-network. • The proposed network combines contrastive learning with classifier learning. • The proposed network can learn the relation between samples. • Contrastive learning provides the network with more supervision. • The proposed network can learn an unbiased classifier from imbalanced dataset. [ABSTRACT FROM AUTHOR]

Images have become a significant medium for information transmission, while image forensics has garnered widespread attention from researchers. Due to the scarcity of finely annotated images in the field of image manipulation detection and localization, existing methods for locating manipulated images can only utilize a limited amount of data for model training. However, deep learning models typically require a large number of finely annotated images to fully leverage their powerful fitting capabilities. In this paper, we propose a semi-supervised image manipulation localization framework, employing semi-supervised learning to use unannotated images for deep learning model training. To achieve this goal, we first design a residual enhancement module that contains an encoding-decoding structure. The regression target of this branch is the pristine regions in the images to generate the reconstructed images. Next, we perform a residual operation between the reconstructed and original images to roughly locate the anomalous regions and refine the features extracted by the encoder. Secondly, we employ weakly-supervised learning by adding an image-level classification head to the final layer of the encoder. The classification head generates a class active map based on image-level classification results, which further guides the model's feature augmentation. Finally, we formulate a specialized semi-supervised framework tailored for image manipulation detection, enabling the utilization of a large volume of unannotated data for model training. Extensive experimental results highlight the notable efficacy of the proposed approach, and outperforms state-of-the-art approaches. • Unannotated images are utilized for image manipulation detection. • Semi-supervised framework tailored for image manipulation detection. • Weakly-supervised framework for features augmentation. • Residual enhancement module refines the features extracted by encoder. • Approach shows state-of-the-art performance. [ABSTRACT FROM AUTHOR]

With rapid growth of online healthcare systems, chief complaint classification plays an important role in areas such as triage or doctor recommendation. Existing medical text classification techniques such as rule-based or learning-based methods fail to effectively utilize the inherent hierarchical structure of label descriptions that contain strong domain knowledge. In this paper, we propose a novel text classification framework for chief complaint by embedding both input text and hierarchical structure of label descriptions based on deep neural networks. The proposed framework makes use of not only three branches (i.e. chief complaint branch, main-category branch, and sub-category branch) with a Sequence Information Encoder to encode semantics from chief complaint and hierarchical structure of label descriptions but also a Hierarchical Relational Network with Attention module to capture complex relationships among them focusing on informative words with attentional scores. We evaluate our framework on two public medical datasets with label descriptions extracted from medical books and websites. Experimental results show that the proposed method outperforms other baseline techniques by a significant margin. The source code of our framework is available at ANONYMISED. [Display omitted] • Novel chief complaint classification model using hierarchical label descriptions. • SIE effectively embeds input text to capture sequential information with context. • HRNA captures relationships among chief complaint and category descriptions. • The framework greatly outperforms SoTA methods on two real world datasets. [ABSTRACT FROM AUTHOR]

Flow field segmentation and classification help researchers to understand vortex structure and thus turbulent flow. Existing deep learning methods are mainly based on global vorticity information and focused on 2D circumstance. In this paper, based on flow field theory, criteria for flow field segmentation and classification in 3D space using local velocity information and the physical relationship between local vorticity and vortex wake are conducted, and then novel physical local information-embedded flow field segmentation and classification methods in 3D space using Multilayer Perceptron (MLP) are proposed. Our methods consist of three modules: a preprocessing module, a flow field segmentation module and a flow field classification module. The preprocessing module calculates input for succeeding modules based on the above criteria. The segmentation and classification modules process corresponding input to identify vortex structure, and further classify the type of vortex wakes in a 3D flow field. Experimental results show that our method can identify vortex area with an accuracy rate that reaches 90% while reducing time consumption by more than 90% compared with existing segmentation methods; our classification method reaches an accuracy rate of 99.7% while reducing time consumption by more than 90% compared with existing classification methods. Our proposed methods proved to achieve good performance both on accuracy and efficiency. • A novel classification criterion based on Re-vorticity relationship is constructed. • 3D flow field classification learning method using above criterion is proposed. • Explore new 3D flow field segmentation criterion using local velocity information. • A deep learning method for efficient 3D flow field segmentation is introduced. [ABSTRACT FROM AUTHOR]

With the rapid advancement of the internet, there has been a dramatic increase in short-text data. Due to the brevity of short texts, sparse features, and limited contextual information, short-text classification has become a challenging task in natural language processing. However, current methods primarily capture semantic information from locally-sequenced words in short text, which ignores the intricate feature relationships that pervade both the intra-text and inter-text. Therefore, this paper proposes a novel Edge-Enhanced Minimum-Margin Graph Attention Network (EMGAN) for short text classification to address this issue. Specifically, we construct a Heterogeneous Information Graph (HIG) to represent complex relationships among short text features. HIG mainly considers the relationship between document features and three attribute features, such as entities, topics, and keywords, and can represent short text features from multiple dimensions and levels. Then, to enhance the connectivity and expressiveness of the HIG for more effective propagation of feature information within it, we present a novel X-shaped structure edge-enhancement method. It enriches their relationships by reconstructing the edge structures. Furthermore, we design a Minimum Margin Graph Attention Network (MMGAN) for short text classification. Specifically, this method aims to explore the minimum margin between high-order neighbors and central nodes at the minimum cost, efficiently extracting and aggregating feature information. Extensive experimental results demonstrate that our proposed EMGAN model outperforms existing methods on five datasets, validating its effectiveness in short-text classification. Our code is submitted at https://github.com/w123yy/EMGAN. • A new method for efficient classification of short texts. • Heterogeneous infographics represent complex relationships among short texts. • The edge enhancement method enriches edge relations and enhances graph structure. • Minimum margin attention network aggregates feature to solve feature sparsity. [ABSTRACT FROM AUTHOR]

• We analyze the limitations in the Enhanced Fuzzy Min-Max (EFMM) model. • Proposing a new FEFMM model to overcome the EFMM limitations. • The performance of FEFMM is evaluated using benchmark data sets. • Results show FEFMM's superior classification accuracy. In this paper, the existing enhanced fuzzy min–max (EFMM) neural network is improved with a flexible learning procedure for undertaking pattern classification tasks. Four new contributions are introduced. Firstly, a new training strategy is proposed for avoiding the generation of unnecessary overlapped regions between hyperboxes of different classes. The learning phase is simplified by eliminating the contraction procedure. Secondly, a new flexible expansion procedure is introduced. It eliminates the use of a user-defined parameter (expansion coefficient) to determine the hyperbox sizes. Thirdly, a new overlap test rule is applied during the test phase to identify the containment cases and activate the contraction procedure (if necessary). Fourthly, a new contraction procedure is formulated to overcome the containment cases and avoid the data distortion problem. Both the third and fourth contributions are important for preventing the catastrophic forgetting issue and supporting the stability-plasticity principle pertaining to online learning. The performance of the proposed model is evaluated with benchmark data sets. The results demonstrate its efficiency in handling pattern classification tasks, outperforming other related models in online learning environments. [ABSTRACT FROM AUTHOR]

Feature selection refers to the process of identifying the most relevant and valuable features or characteristics in a dataset, and it is an important step that can improve the performance and accuracy of predictive models for machine learning and data mining. Feature selection is crucial in efficiently reducing the number of features, thereby enhancing classification accuracy. This, in turn, alleviates computational burdens and elevates the overall performance of machine learning algorithms. This paper extended the conventional manta ray foraging optimization (MRFO) algorithm to an improved binary MRFO (IBMFO) algorithm. IBMRFO serves as a search strategy for designing a wrapper-based feature selection approach. To commence, we introduce a novel mechanism, termed the chaotic tent map (CTM), aimed at initializing the population to fortify both the algorithm's exploitation capabilities and population diversity. Second, an adaptive somersault foraging (ASF) factor is tuned to avoid premature convergence. Finally, a binary mechanism is introduced to adapt the algorithm for binary feature selection dilemmas. Experiments are performed on 30 renowned UC Irvine machine learning repository datasets. The results of the student's t-test show that the classification algorithm has statistical significance. The findings unequivocally exhibit the superior performance of the proposed IBMRFO compared to other algorithms under scrutiny. Notably, the efficiency of the enhanced factors is assessed through rigorous testing. For post-publications supporting this work, readers can refer to https://github.com/ELF-CHEUNG/IBMRFO. • Novel chaotic tent map for population initialization from evolutionary computation. • Optimization methods based on enhanced evolutionary computation. • Adaptive factor enhances convergence rate and search efficiency of algorithm. • Extensive numerical and scientific simulations for verifications with UCI dataset. [ABSTRACT FROM AUTHOR]

Many deep graph neural networks perform well in tackling semi-supervised node classification tasks. However, they often focus on learning a holistic representation for nodes, which lacks interpretability and compactness for classification purposes. In this paper, we propose a Multi-Relation Enhanced Graph Disentangled Network (MREGDN), which can effectively disentangle the latent factors that affect graph topology, while enhancing the interpretability and low-redundancy of the learned node representations for classification. There are three types of relation constraints designed elaborately to optimize the disentangled representation learning process: (a) constraints of intra-factor compactness to encourage a factor to represent a distribution, (b) constraints of inter-factor irrelevance to ensure diversity among factors, and (c) constraints of keeping inter-node similarity in different spaces to retain necessary information related to category. Extensive experiments verify that MREGDN significantly outperforms its best-known competitors on various semi-supervised node classification tasks. • MREGDN improves the interpretability and discrimination of node representations. • MREGDN learns more compact and category-relevant node representations. • MREGDN greatly outperforms existing competitors on various classification tasks. [ABSTRACT FROM AUTHOR]

Leveraging multi-modal information sources has attracted the attention of researchers and practitioners for developing resources and technologies in the broad areas of applied Artificial Intelligence (AI). During the occurrence of a natural disasters, people heavily use social media for communication by posting multimedia information in the form of texts and images. In such a critical situations, it becomes imperative to use all modalities of information sources to better capture vital knowledge related to the crisis. In this paper, we propose an effective deep learning model to leverage multi-modal information sources in the form of both texts and images, and then disseminate useful information at the time of natural disasters. Our proposed model classifies the tweets into seven critical and potentially actionable categories, such as the reports of 'injured or dead people', 'infrastructure damage' etc. Experiments on a benchmark dataset show that fusion of multi-modal information sources, viz. text and image both, are more effective compared to uni-modal (i.e. either text or image) source in extracting meaningful information generated during disaster situations. By using information from both the modalities (text and image), we obtain a macro F1-Score of 0.51, which is a significant improvement over the baseline models that make use of only text or image for the classification. We supplement our results with a thorough analysis exploring the reasons for this phenomenon, thus, further demonstrating the utility of exploiting multiple modalities. The primary contribution of this paper lies in developing an attentive deep learning model that uses social media text and image to help classify images into crucial classes for disaster domain. The major findings of our research are that using textual features while classifying the corresponding images, improves the classification performance. We also explore different methods of fusion of multi-modal features and conclude that fusion through attention mechanism works best for image classification in disaster domain. • A multi-modal deep learning model for image classification in disaster domain. • Devising an attention mechanism for combination of the two modalities. • Showing how adding text can help in image classification. [ABSTRACT FROM AUTHOR]

• Transductive and Inductive conformal isotonic separation techniques are proposed. • A new Non-Conformity measure is introduced to find uniqueness of data point. • Performance of proposed techniques is evaluated on various data sets. • Statistical and time complexity analysis of the proposed work are done. Isotonic Separation (IS) is a non-parametric classification algorithm which constructs an isotonic function as a model from given data with monotonic nature. This paper proposes an approach in which a novel conformal prediction framework is deployed in two setups to improve the classification performance of isotonic separation: Transductive and Inductive method. This paper introduces a new Non-Conformity Measure (NCM) which is specific to isotonic separation to compute the uniqueness of a data point to a set of data points. This measure is deployed on two different forms of conformal prediction methods: Transductive Conformal Isotonic Separation (TCIS) and Inductive Conformal Isotonic Separation (ICIS). Experiments are conducted on different monotonic data sets using transductive and inductive isotonic separation techniques. Experimental and statistical results reveal that transductive and inductive conformal isotonic separation give significant results in terms of performance measures. [ABSTRACT FROM AUTHOR]

The paper presents application of data mining methods for recognizing the most significant genes and gene sequences (treated as features) stored in a dataset of gene expression microarray. The investigations are performed for autism data. Few chosen methods of feature selection have been applied and their results integrated in the final outcome. In this way we find the contents of small set of the most important genes associated with autism. They have been applied in the classification procedure aimed on recognition of autism from reference group members. The results of numerical experiments concerning selection of the most important genes and classification of the cases on the basis of the selected genes will be discussed. The main contribution of the paper is in developing the fusion system of the results of many selection approaches into the final set, most closely associated with autism. We have also proposed special procedure of estimating the number of highest rank genes used in classification procedure. [ABSTRACT FROM AUTHOR]

The work presented in this papers is directed at mechanisms where by 3D surfaces can be represented to support the generation and application of classification techniques. Three different mechanisms are presented to allow for the representation of 3D surfaces in such a way that key features are retained while at the same time ensuring compatibility with prediction (classification) techniques. The three representation techniques are: (i) Local Geometry Matrices (LGMs) founded on the concept of local binary patterns, (ii) Local Distance Measure (LDM) founded on the idea that distances from edges (critical points) may be significant, and (iii) Point Series (PS) whereby local geometries are represented in terms of a linearisation of space. The representations are designed to capture the nature of 3D surfaces in terms of their local geometry and predict class labels associated with such local geometries. To act as a focus for the work the prediction of “springback” within the context of sheet metal forming is considered, where springback is a form of deformation that occurs (in a non-uniform manner) across a manufactured 3D surface as a result of the application of some sheet metal forming process. The evaluation of each of the techniques, and variations thereof, using sheet metal parts that have been manufactured especially for the purpose, is fully described. The paper also reports on a statistical significance test concerning the results. [ABSTRACT FROM AUTHOR]

Classification is an important task in various areas. In many real-world applications, class imbalance and overlapping problems have been reported as major issues in the application of traditional classification algorithms. An imbalance problem occurs when training data contain considerably more representatives of one class than of other classes. Class overlap occurs when a region in the data space contains a similar number of data for each class. When a class overlap occurs in imbalanced data sets, classification becomes even more complicated. Although various approaches have been proposed to deal separately with class imbalance and overlapping problems, only a few studies have attempted to address both problems simultaneously. In this paper, we propose an overlap-sensitive margin (OSM) classifier based on a modified fuzzy support vector machine and k -nearest neighbor algorithm to address imbalanced and overlapping data sets. The main idea of the proposed OSM classifier is to separate the data space into soft- and hard-overlap regions using the modified fuzzy support vector machine algorithm. The separated spaces are then classified using the decision boundaries of the support vector machine and 1-nearest neighbor algorithms. Furthermore, by separating a data set into soft- and hard-overlap regions, one can determine which part of the data is to be examined more closely for classification in real-world situations. Experiments using synthetic and real-world data sets demonstrated that the proposed OSM classifier outperformed existing methods for imbalanced and overlapping situations. [ABSTRACT FROM AUTHOR]

Classifiers operating in a dynamic, real world environment, are vulnerable to adversarial activity, which causes the data distribution to change over time. These changes are traditionally referred to as concept drift, and several approaches have been developed in literature to deal with the problem of drift detection and handling. However, most concept drift handling techniques approach it as a domain independent task, to make them applicable to a wide gamut of reactive systems. These techniques are developed from an adversarial agnostic perspective, where they naively assume that adversarial activity is like any other change to the data, which can be fixed by retraining the models. However, this is not the case when a malicious agent is trying to evade the deployed classification system. In such an environment, the properties of concept drift are unique, as the drift is intended to degrade the system and at the same time designed to avoid detection by traditional concept drift detection techniques. This special category of drift is termed as adversarial drift, and this paper analyzes its characteristics and impact in a streaming environment. A novel framework for dealing with adversarial concept drift is proposed, called the Predict-Detect streaming framework. This framework uses adversarial forethought and incorporates the context of classification into the drift detection task, to provide leverage in dynamic-adversarial domains. Experimental evaluation of the framework, on generated adversarial drifting data streams, demonstrates that this framework is able to provide early and reliable unsupervised indication of drift, and is able to recover from drifts swiftly. While traditional drift detectors can be evaded by intelligent adversaries, the proposed framework is especially designed to capture adversaries by misdirecting them into revealing themselves. In addition, the framework is designed to work on imbalanced and sparsely labeled data streams, as a limited-memory, incremental algorithm. The generic design and domain independent nature of the framework makes it applicable as a blueprint for developers wanting to implement reactive security to their classification based systems. [ABSTRACT FROM AUTHOR]

Expert and intelligent systems understand the underlying information behind the data by relying on a wide range of machine learning techniques. The interpretation of machine learning models is often the key to success in research areas such as business, finance, medical and health science, and bioinformatics; such research areas demand human understanding of the obtained model. The logistic model tree (LMT) algorithm is a popular classification method that combines a decision tree and logistic regression models. The combination of two complementary algorithms produces an accurate and interpretable classifier by combining the advantages of both logistic regression and tree induction. However, LMT has the disadvantage of high computational cost, which makes the algorithm undesirable in practice. In this paper, we propose an efficient method to learn the logistic regression models in the tree. We employ least angle regression to update the regression model in LogitBoost so that the algorithm efficiently learns sparse logistic regression models composed of relevant input variables. We compare the performance of our proposed method with the original LMT algorithm using 14 benchmark datasets and show that the training time dramatically decreases while the accuracy is preserved. Our proposed algorithm is not only accurate and intuitively interpretable but also computationally efficient. It helps users in making the best possible use of the data that are included in expert and intelligent systems. [ABSTRACT FROM AUTHOR]

Identification of the material from which an object is made is of significant value for effective robotic grasping and manipulation. Characteristics of the material can be retrieved using different sensory modalities: vision based, tactile based or sound based. Compressibility, surface texture and thermal properties can each be retrieved from physical contact with an object using tactile sensors. This paper presents a method for collecting data using a biomimetic fingertip in contact with various materials and then using these data to classify the materials both individually and into groups of their type. Following acquisition of data, principal component analysis (PCA) is used to extract features. These features are used to train seven different classifiers and hybrid structures of these classifiers for comparison. For all materials, the artificial systems were evaluated against each other, compared with human performance and were all found to outperform human participants’ average performance. These results highlighted the sensitive nature of the BioTAC sensors and pave the way for research that requires a sensitive and accurate approach such as vital signs monitoring using robotic systems. [ABSTRACT FROM AUTHOR]

The application of classifiers on data represents an important help in a process of decision making. Any classifier, or other method used for knowledge extraction, suffers a deterioration when it is applied on data with noise. Credal C4.5 (CC4.5) is a recent method of classification, that introduces imprecise probabilities in the algorithm of the classic C4.5. It is very suitable in classification noise tasks, but it has a clear dependency of a parameter. It has been proved that this parameter is related with the level of overfitting of the model on the data used for training. In noisy domains, this characteristic is important in the sense that variations of this parameter can reduce the variance of the model. Depending on the degree of noise that a data set has, the application of different values of this parameter can produce different performance of the CC4.5 model. Hence, the use of the correct parameter is fundamental to attain a high level of performance for this model. In this paper, that problem is solved via a rough procedure to estimate the level of class noise in the training data. Combining this new noise estimation process with the CC4.5, it is presented a direct method that has an equivalent performance than the one of the CC4.5 when it is used with the best value of its parameter for each level of class noise. [ABSTRACT FROM AUTHOR]

Radio frequency identification (RFID) has been widely used for the automatic identification, tracking and tracing of goods throughout the supply chain from the manufacturer to the customer. However, one technological problem that impedes the productive and reliable use of RFID is the constraint of false positive readings, which refers to tags that are detected accidentally by the reader but not the ones of interest. This paper focuses on the use of machine learning algorithms to identify such RFID readings. A total of 11 statistical features are extracted from received signal strength (RSS) and phase rotations derived from the raw RFID data. Each of the features is highly statistically different to distinguish the false positive readings, but satisfactory classification cannot be achieved when these features are considered individually. Classifiers based on logistic regression (LR), support vector machine (SVM) and decision tree (DT) are constructed, which combine all of the extracted features to classify the RFID readings more effectively. The performance of the classifiers is evaluated in a real-world factory. Results show that SVM provides the highest accuracy of up to 95.3%. DT shows slightly better accuracy (92.85%) than LR (92.75%), while LR has the larger area under the curve (0.976) than DT (0.949). Overall, machine learning algorithms could achieve accuracy of 93% on average. The proposed methodology provides a much more reliable RFID application as false-positive readings are detected immediately without human intervention, which enables a significant potential of fully automatic identification and tracking of goods throughout the supply chain. [ABSTRACT FROM AUTHOR]

The number of applications for smartphones and tablets is growing exponentially in the last years. Many of these applications are supported by the so-called Location Based Services, which are expected to provide reliable real-time localization anytime and anywhere, no matter either outdoors or indoors. Even though outdoors world-wide localization has been successfully developed through the well-known Global Navigation Satellite System technology, its counterpart large-scale deployment indoors is not available yet. In previous work, we have already introduced a novel technology for indoor localization supported by a WiFi fingerprint approach. In this paper, we describe how to enhance such approach through the combination of hierarchical localization and fuzzy classifier ensembles. It has been tested and validated at the University of Edinburgh, yielding promising results. [ABSTRACT FROM AUTHOR]

The purpose of this paper is to introduce a novel similarity measure of intuitionistic fuzzy sets (IFSs). The proposed measure is based on the equivalence relation in the IFS-IBA approach. Due to the logic-based background, this measure compares IFS from a different viewpoint than the standard measures, emphasizing comprehension of intuitionism. The IFS-IBA similarity measure has a solid mathematical background and can be combined with various IF aggregation operators. Additionally, we define IFS-IBA distance function as a complement of IFS-IBA similarity. Both IFS-IBA similarity and distance functions may have different realizations that are easy to interpret. Hence, the measures are offering great descriptive power and the ability to model various problems. The benefits of the proposed measure are illustrated on the problem of pattern recognition and classification within k -NN algorithm. Finally, we show that the proposed measure is appropriate for IF hierarchical clustering on the problem of clustering Serbian medium-sized companies according to their financial ratios. Results obtained using the IFS-IBA measure are clear-cut and more meaningful compared to a standard IF distances regardless of the I-fuzzification method used. [ABSTRACT FROM AUTHOR]

In this paper, we propose a new decision forest algorithm that builds a set of highly accurate decision trees by exploiting the strength of all non-class attributes available in a data set, unlike some existing algorithms that use a subset of the non-class attributes. At the same time to promote strong diversity, the proposed algorithm imposes penalties (disadvantageous weights) to those attributes that participated in the latest tree in order to generate the subsequent trees. Besides, some other weight-related concerns are taken into account so that the trees generated by the proposed algorithm remain individually accurate and retain strong diversity. In order to show the worthiness of the proposed algorithm, we carry out experiments on 20 well known data sets that are publicly available from the UCI Machine Learning Repository. The experimental results indicate that the proposed algorithm is effective in generating highly accurate and more balanced decision forests compared to other prominent decision forest algorithms. Accordingly, the proposed algorithm is expected to be very effective in the domain of expert and intelligent systems. [ABSTRACT FROM AUTHOR]

In case-based reasoning (CBR) classification systems, the similarity metrics play a key role and directly affect the system's performance. Based on our previous work on the learning pseudo metrics (LPM), we propose a case-based reasoning method for pattern classification, where the widely used Euclidean distance is replaced by the LPM to measure the closeness between the target case and each source case. The same type of case as the target case can be retrieved and the category of the target case can be defined by using the majority of reuse principle. Experimental results over some benchmark datasets and a fault diagnosis of the Tennessee-Eastman (TE) process demonstrate that the proposed reasoning techniques in this paper can effectively improve the classification accuracy, and the LPM-based retrieval method can substantially improve the quality and learning ability of CBR classifiers. [ABSTRACT FROM AUTHOR]

By present, many models of bankruptcy forecasting have been developed, but this area remains a field of research activity; little is known about the practical application of existing models. In our opinion, this is because the use of existing models is limited by the conditions in which they are developed. Another question concerns the factors that can be significant for forecasting. Many authors suggest that indicators of the external environment, corporate governance as well as firm size contain important information; on the other hand, the large number of factors does not necessary increase predictive ability of a model. In this paper, we suggest the genetic algorithm based two-step classification method (TSCM) that allows both selecting the relevant factors and adapting the model itself to application. Classifiers of various models are trained at the first step and combined into the voting ensemble at the second step. The combination of random sampling and feature selection techniques were used to ensure the necessary diversity level of classifiers at the first step. The genetic algorithms are applied at the step of features selection and then at the step of weights determination in ensemble. The characteristics of the proposed method have been tested on the balanced set of data. It included 912 observations of Russian companies (456 bankrupts and 456 successful) and 55 features (financial ratios and macro/micro business environment factors). The proposed method has shown the best accuracy (0.934) value among tested models. It has also shown the most balanced precision-recall ratio. It found bankrupts (recall = 0.953) and not bankrupts (precision = 0.910) rather accurately than other tested models. The ability of method to select the task-relevant features has been also tested. Excluding the features that are significant for less than 50% of the classifiers in the ensemble improved the all performance metrics (accuracy = 0.951, precision = 0.932, recall = 0.965). So, the proposed method allows to improve the advantages and alleviate the weaknesses inherent in ordinary classifiers, enabling the business decisions support with a higher reliability. [ABSTRACT FROM AUTHOR]

Feature selection, both for supervised as well as for unsupervised classification is a relevant problem pursued by researchers for decades. There are multiple benchmark algorithms based on filter, wrapper and hybrid methods. These algorithms adopt different techniques which vary from traditional search-based techniques to more advanced nature inspired algorithm based techniques. In this paper, a hybrid feature selection algorithm using graph-based technique has been proposed. The proposed algorithm has used the concept of Feature Association Map (FAM) as an underlying foundation. It has used graph-theoretic principles of minimal vertex cover and maximal independent set to derive feature subset. This algorithm applies to both supervised and unsupervised classification. The performance of the proposed algorithm has been compared with several benchmark supervised and unsupervised feature selection algorithms and found to be better than them. Also, the proposed algorithm is less computationally expensive and hence has taken less execution time for the publicly available datasets used in the experiments, which include high-dimensional datasets. [ABSTRACT FROM AUTHOR]

This research proposes a new hybrid approach for feature selection and Support Vector Machine (SVM) model selection based on a new variation of Cohort Intelligence (CI) algorithm. Feature selection can improve the accuracy of classification algorithms and reduce their computation complexity by removing the irrelevant and redundant features. SVM is a classification algorithm that has been used in many areas, such as bioinformatics and pattern recognition. However, the classification accuracy of SVM depends mainly on tuning its hyperparameters (i.e., SVM model selection). This paper presents a framework that is comprised of the following two major components. First, Self-Adaptive Cohort Intelligence (SACI) algorithm is proposed, which is a new variation of the emerging metaheuristic algorithm, Cohort Intelligence (CI). Second, SACI is integrated with SVM resulting in a new hybrid approach referred to as SVM–SACI for simultaneous feature selection and SVM model selection. SACI differs from CI by employing tournament-based mutation and self-adaptive scheme for sampling interval and mutation rate. Furthermore, SACI is both real-coded and binary-coded, which makes it directly applicable to both binary and continuous domains. The performance of SACI for feature selection and SVM model selection was examined using ten benchmark datasets from the literature and compared with those of CI and five well-known metaheuristics, namely, Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Differential Evolution (DE) and Artificial Bee Colony (ABC). The comparative results demonstrate that SACI outperformed CI and comparable to or better than the other compared metaheuristics in terms of the SVM classification accuracy and dimensionality reduction. In addition, SACI requires less tuning efforts as the number of its control parameters is less than those of the other compared metaheuristics due to adopting the self-adaptive scheme in SACI. Finally, this research suggests employing more efficient methods for high-dimensional or large datasets due to the relatively high training time required by search strategies based on metaheuristics when applied to such datasets. [ABSTRACT FROM AUTHOR]

Active learning is used for classification when labeling data are costly, while the main challenge is to identify the critical instances that should be labeled. Clustering-based approaches take advantage of the structure of the data to select representative instances. In this paper, we developed the active learning through density peak clustering (ALEC) algorithm with three new features. First, a master tree was built to express the relationships among the nodes and assist the growth of the cluster tree. Second, a deterministic instance selection strategy was designed using a new importance measure. Third, tri-partitioning was employed to determine the action to be taken on each instance during iterative clustering, labeling, and classifying. Experiments were performed with 14 datasets to compare against state-of-the-art active learning algorithms. Results demonstrated that the new algorithm had higher classification accuracy using the same number of labeled data. [ABSTRACT FROM AUTHOR]

It is well established that multiple EEG channels are required for various brain functionality studies, including classification tasks. Yet, due to the curse of dimensionality problem, the analysis of multiple channels may not lead to the desired performance. Accordingly, a number of static channel selection algorithms have been proposed to identify the most relevant subset of channels. However, static methods select a fixed subset of channels that is unchanged when processing new data, and hence cannot adapt to changes in data. In this paper, we propose a novel algorithm that utilizes the dynamic classification behaviour of channels in selecting the channel that is most relevant for each time segment of the signal. The main idea is to identify for each time segment of every channel of the signal (testing sample) the closest training samples. These training samples are used to estimate the local accuracy of each channel. The best performing channel for that time segment will then be identified as the relevant one. Results obtained using EEG data of a four-class alertness state classification problem, with two different feature sets, reveal that the proposed approach is capable of achieving competitive performance compared to a traditional static channel selection based method. More importantly, the evaluation of the selected channels reveals that our approach is able to select relevant channels for each of the four alertness states. The proposed algorithm is expected to make a valuable contribution to the field of multichannel biomedical signal classification. [ABSTRACT FROM AUTHOR]

Bacterial foraging optimization (BFO) algorithm is a new swarming intelligent method, which has a satisfactory performance in solving the continuous optimization problem based on the chemotaxis, swarming, reproduction and elimination-dispersal steps. However, BFO algorithm is rarely used to deal with feature selection problem. In this paper, we propose two novel BFO algorithms, which are named as adaptive chemotaxis bacterial foraging optimization algorithm (ACBFO) and improved swarming and elimination-dispersal bacterial foraging optimization algorithm (ISEDBFO) respectively. Two improvements are presented in ACBFO. On the one hand, in order to solve the discrete problem, data structure of each bacterium is redefined to establish the mapping relationship between the bacterium and the feature subset. On the other hand, an adaptive method for evaluating the importance of features is designed. Therefore the primary features in feature subset are preserved. ISEDBFO is proposed based on ACBFO. ISEDBFO algorithm also includes two modifications. First, with the aim of describing the nature of cell to cell attraction-repulsion relationship more accurately, swarming representation is improved by means of introducing the hyperbolic tangent function. Second, in order to retain the primary features of eliminated bacteria, roulette technique is applied to the elimination-dispersal phase. In this study, ACBFO and ISEDBFO are tested with 10 public data sets of UCI. The performance of the proposed methods is compared with particle swarm optimization based, genetic algorithm based, simulated annealing based, ant lion optimization based, binary bat algorithm based and cuckoo search based approaches. The experimental results demonstrate that the average classification accuracy of the proposed algorithms is nearly 3 percentage points higher than other tested methods. Furthermore, the improved algorithms reduce the length of the feature subset by almost 3 in comparison to other methods. In addition, the modified methods achieve excellent performance on wilcoxon signed-rank test and sensitivity-specificity test. In conclusion, the novel BFO algorithms can provide important support for the expert and intelligent systems. [ABSTRACT FROM AUTHOR]

Individuals use Twitter for personal communication, whereas businesses, politicians and celebrities use Twitter for branding purposes. Distinguishing Personal from Branding Twitter accounts is important for Twitter analytics. Existing studies of Twitter account classification apply classical supervised learning, which requires intensive manual annotation for training. In this paper, we propose CDS (Collaborative Distant Supervision), a novel learning scheme for Twitter account classification that does not require intensive manual labelling. Twitter accounts are automatically labelled using heuristics for distant supervision learning. To achieve effective learning from heuristic labels, active learning is applied to identify and correct false positive labels, and semi-supervised learning is applied to further use false negatives missed by labelling heuristics for learning. Extensive experiments on Twitter data showed that CDS achieved high classification accuracy. [ABSTRACT FROM AUTHOR]

A new nonlinear filter, which employs an adaptive spline function as the basis function is designed in this paper. The input signal to this filter is used to generate suitable parameters to update the control points in a spline function. The update rule for updating the control points have been derived and a mean square analysis has been carried out. The output of the spline functions are suitably combined together to obtain the filter response. This filter is called the generalized spline nonlinear adaptive filter (GSNAF). The proposed GSNAF is similar to a functional link artificial neural network (FLANN), considering a functional expansion using spline basis functions. GSNAF has been shown to offer improved accuracy in benchmark classification scenarios and provide enhanced modeling accuracy in single input single output as well as in multiple input multiple output dynamic system identification cases. [ABSTRACT FROM AUTHOR]

The aim of this paper is, first, to recall fuzzy relational compositions (products) and, to introduce an idea, how excluding features could be incorporated into the theoretical background. Apart from definitions, we provide readers with a theoretical investigation. This investigation addresses two natural questions. Firstly, under which conditions (in which underlying algebraic structures) the given three natural approaches to the incorporation of excluding symptoms coincide. And secondly, under which conditions, the proposed incorporation of excluding features preserves the same natural and desirable properties similar to those preserved by fuzzy relational compositions. The positive impact of the incorporation on reducing the suspicions provided by the basic “circlet” composition without losing the possibly correct suspicion is demonstrated on a real taxonomic identification (classification) of Odonata. Here, we demonstrate how the proposed concept may eliminate the weaknesses provided by the classical fuzzy relational compositions and, at the same time, compete with powerful machine learning methods. The aim of the demonstration is not to show that proposed concept outperforms classical approaches, but to show, that its potential is strong enough in order to complete them or in order to be combined with them and to use its different nature. [ABSTRACT FROM AUTHOR]

Supervised text classification methods are efficient when they can learn with reasonably sized labeled sets. On the other hand, when only a small set of labeled documents is available, semi-supervised methods become more appropriate. These methods are based on comparing distributions between labeled and unlabeled instances, therefore it is important to focus on the representation and its discrimination abilities. In this paper we present the ST LDA method for text classification in a semi-supervised manner with representations based on topic models. The proposed method comprises a semi-supervised text classification algorithm based on self-training and a model, which determines parameter settings for any new document collection. Self-training is used to enlarge the small initial labeled set with the help of information from unlabeled data. We investigate how topic-based representation affects prediction accuracy by performing NBMN and SVM classification algorithms on an enlarged labeled set and then compare the results with the same method on a typical TF-IDF representation. We also compare ST LDA with supervised classification methods and other well-known semi-supervised methods. Experiments were conducted on 11 very small initial labeled sets sampled from six publicly available document collections. The results show that our ST LDA method, when used in combination with NBMN, performed significantly better in terms of classification accuracy than other comparable methods and variations. In this manner, the ST LDA method proved to be a competitive classification method for different text collections when only a small set of labeled instances is available. As such, the proposed ST LDA method may well help to improve text classification tasks, which are essential in many advanced expert and intelligent systems, especially in the case of a scarcity of labeled texts. [ABSTRACT FROM AUTHOR]

Surface EMGs have been the primary sources for control of prosthetic hands due to their comfort and naturalness. The recent advances in the development of the prosthetic hands with many degrees of freedom and many actuators, requires many EMG channels to take the full advantage of the complex prosthetic terminals. Some EMG wearable devices were developed lately, that are able to detect several gestures. However, the main drawbacks of these systems are the cost, the size and the system complexity. In this paper, we suggest a simple, fast and low-cost system which can recognize up to 4 gestures with a single channel surface EMG signal. Gestures include hand closing, hand opening, wrist flexion and double wrist flexion. These gestures can be used to control a prosthetic terminal based on predefined grasp postures. We show that by using a high-dimensional feature space, together with a support vector machine algorithm, it is possible to classify these four gestures. Overall, the system showed satisfactory results in terms of classification accuracy, real time gesture recognition, and tolerance to hand movements through integration of a lock gesture. Calibration took only 30 seconds and session independence was demonstrated by high classification accuracy on different test sessions without repeating the calibration. As a case study we use this system to control a previously developed soft prosthetic hand. This is particularly interesting because we show that a simple hardware that has only a single channel EMG, can afford the control of a multi-DOF prosthetic hands. In addition, such system may be used as a general purpose Human Machine Interface for gaming,for controlling multimedia devices, or to control robots. [ABSTRACT FROM AUTHOR]

In this paper, starting from a comprehensive mathematical model of a Collaborative Reputation Systems (CRSes), we present a research study within the Cultural Heritage domain. The main goal of this study has been the evaluation and classification of the visitors’ behaviour during a cultural event. By means of mobile technological instruments, opportunely deployed within the environment, it is possible to collect data representing the knowledge to be inferred and give a reliable rate for both visitors and exposed artworks. Discussed results, confirm the reliability and the usefulness of CRSes for deeply understand dynamics related to people visiting styles. [ABSTRACT FROM AUTHOR]

Motorcycles are one of the most dangerous means of transportation. Its death toll is higher than in others, due to the inherent vulnerability of motorcycle drivers. The latest strategies in Advanced Driving Assistance Systems (ADAS) are trying to mitigate this problem by applying the advances of modern technologies to the road transport. This paper presents two different approaches on motorcycle protection, based on two of the most modern available technologies in ADAS, i.e. Computer Vision and Vehicle to Vehicle Communication (V2V). The first approach is based on data fusion of Laser Scanner and Computer Vision, providing accurate obstacle detection and localization based on laser scanner, and obstacle classification using computer vision and laser. The second approach is based on ad-hoc V2V technology and provides detection in case of occlusion for visual sensors. Both technologies have been tested in the presented work, and a performance comparison is given. Tests performed in different driving situations allows to measure the performance of every algorithm and the limitations of each of them based on empirical and scientific foundations. The conclusions of the presented work help foster of expert systems in the automotive sector by providing further discussion of the viability and impact from each of these systems in real scenarios. [ABSTRACT FROM AUTHOR]

Multi-instance learning (MIL) allows predictive algorithms to use complex data representation. The data in MIL is organized in the form of labeled bags of instances, and the labels of instances are not available in the training phase. The processing and classification of complex bag representation result in a complicated hypothesis space. Moreover, identifying essential instances inside is also important, as these instances trigger positive labels for the bag and play a vital role in the model interpretation and bag classification. The recent MIL algorithms are not robust to tackle hypothesis space complication. Additionally, the existing instance selection algorithms are based on explicit assumptions regarding the relationship of instances to the bag label. However, these assumptions may hold for a specific bag in the dataset but may not apply to the whole dataset. To deal with the hypothesis space complication and robust instance selection without any prior assumption, this paper proposes a fuzzy subspace clustering approach for robust instance selection and ensemble-based variant for Fisher vector (FV) encoding to solve MIL problems, named (FCBE-miFV). Specifically, the proposed algorithm uses a subspace fuzzy clustering approach to compute instance selection probabilities, selects essential instances from the bag, transforms the input bag into FV, and classifies the generated bag encodings using a stacking-based ensemble approach to obtain improved bag level classification performance. The proposed FCBE-miFV improves model performance by incorporating essential instances in the bag encoding process. The experimental results show that the FCBE-miFV obtained comparable performance to the state-of-the-art MIL problems. [ABSTRACT FROM AUTHOR]

Accurate polyp segmentation is of great significance for the diagnosis and treatment of colon cancer. Deep convolution network can extract the common high level features of the target. However, most network models ignore some individual features, so the sample prediction results in complex space are fuzzy and lack of regularity. In this paper, a coarse-to-fine segmentation frame for polyp segmentation via deep and classification features is proposed. Firstly, batch schatten-p norms maximization is introduced into a network model to strengthen the predict map. Then, an automatic two classification mechanism is constructed and the prediction map is classified into two categories: simple and complex samples. Since the CNN prediction maps of simple samples are close to binary images, the prediction maps are not processed. Finally, an active contour model segmentation algorithm for saliency detection of complex samples is proposed. Experiments on Kvasir-SEG, CVC-300, CVC-ClinincDB, CVC-ColonDB and ETIS-LaribPolypDB datasets using multiple models verify the effectiveness of the framework. Code is available at https://doi.org/10.24433/CO.7821162.v1. • Batch schatten-p norms is introduced into the loss function. • An adaptive classification mechanism is constructed. • A coarse-to-fine segmentation framework is designed. • An ACM based on saliency information is proposed. [ABSTRACT FROM AUTHOR]

With rapid progress and significant successes in a wide domain of applications, deep learning has been extensively employed for solving complex problems. However, performance of deep learning has been vulnerable to well-designed samples, called adversarial samples. These samples are carefully designed to deceive the deep learning models without human perception. Therefore, vulnerability to adversarial attacks becomes one of the major concerns in life-critical applications of deep learning. In this paper, a novel approach to counter adversarial samples is proposed to strengthen the robustness of a deep learning model. The strategy is to filter the perturbation noise in adversarial samples prior to prediction. The proposed defense framework is based on DiscoGANs to discover the relation between attacker and defender characteristics. Attacker models are created to generate the adversarial samples from the training data, while the defender model is trained to reconstruct original samples from the adversarial samples. These two frameworks are trained to compete with each other in an alternating manner. The experimental results on different attack models are compared with popular defense mechanisms on three benchmark datasets. Our proposed method shows promising results and can improve the robustness on both white-box and black-box attacks including the computation time. • Adversarial samples without human perception degrade model prediction. • Filtering the perturbation noise can strengthen the robustness of the model. • The proposed method achieves impressive performance over various attacks. • Improvement of computation time of the reconstruction process is demonstrated. • Results illustrate the robustness to both white-box and black-box attacks. [ABSTRACT FROM AUTHOR]

Sentiment classification aims to predict the sentiment label for a given text. Recently, several research efforts have been devoted to incorporate matching clues between text words and class labels into the learning process of text representation. However, these methods heavily rely on the availability of label content. Moreover, they simply capture the label-specific signals to measure each word's contribution by either implicitly employing a learnable label representation or explicitly leveraging the interaction between text words and labels via the interaction mechanism. To deal with these issues, in this paper, we propose a novel framework called Label-Guided Dual-view Sentiment Classifier (LGDSC). We first introduce a new strategy for generating an effective label description and then design a novel Dual-Channel Label-guided Attention Network (DLAN) to learn a text representation via two different channels. DLAN will be further leveraged to learn label-guided text representations from two different views. Extensive experimental results on four real-world datasets demonstrate that LGDSC consistently outperforms the state-of-the-art baseline methods. • Propose an inverse label entropy based strategy for generating effective label descriptions. • Design a dual-channel label-guided attention network to learn text representation via two different channels. • Extensive experiments conducted on four widely used datasets demonstrate the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

Breast cancer is one of the leading killers of women around the world. Finding compounds with good bioactivity, metabolic dynamics and safety, including Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET for short), is a long and challenging task in breast cancer therapy. In the paper, molecular descriptor data of compounds was analyzed by the ensemble learning algorithm, and important features were selected for the development and validation of ADMET classification models. The overall process includes data cleaning, data splitting to training and testing sets, feature selection and classification model evaluation. A Two-Level Stacking Algorithm (TLSA) based on ensemble learning is proposed for ADMET classification. Various performance measures like classification accuracy, precision, recall, confusion matrix, F1-score, Receiver Operating Characteristic (ROC) curves and the Area Under the ROC Curves (AUC) are reported to show the superiority of the proposed method as compared to different classifiers. The experimental results show that the second level algorithm for TLSA utilizes Logistic Regression is better than other classifiers for the properties of Absorption, Distribution and Excretion, with accuracy of 94.6037%, 94.9410% and 88.1956% respectively. For the properties Metabolism and Toxicity, the second level algorithm utilizes Support Vector Machine to achieve the best classification performance, with accuracy of 88.8702% and 96.7960% respectively. The results show that the proposed approach works well with the classification of compound properties and can be a good alternative for the well-known machine learning program. • A novel TLSA model for ADMET classification of compounds is proposed. • The proposed method improves the accuracy of classification results. • Applying a wide range of evaluation metrics to ensure the model's reliability. • Most important molecular descriptors of compounds were selected using RF. [ABSTRACT FROM AUTHOR]

Real-world data exhibits a long-tailed label distribution, which leads to classification bias. Popular re-sampling or re-weighting methods usually require known category information. However, learning from long-tailed data with open categories is a challenging issue. In this paper, we propose an active distribution optimization algorithm (DALC) to handle the interesting issue. Through clustering, querying and classification iterations, we explore new categories and balance label distribution. For clustering, we present an exploration technique that adaptively obtains optimal data distribution with minimal total distance/cost. For each query, we design a critical instance selection strategy with the cluster information. For classification, we establish an ensemble model to continuously balance the label distribution. We conducted experiments on synthetic, benchmark and domain datasets. The results of the significance test verified the effectiveness of DALC and its superiority over state-of-the-art long-tailed data classification and open set classification algorithms. • We define the cost function and establish a cluster-based actual statistical model to obtain the best distribution of data. • We propose a strategy based on data distribution to resolve long-tailed data classification with open categories. • We design an algorithm called data distribution driven active learning algorithm (DALC). • Results of significance test verify the superiority of DALC to state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

• A novel method is presented to tackle classification problems with imbalanced data. • It is in the categories of under-sampling data-level and ensemble-based approaches. • It partitions majority samples into clusters by using the convex-hull concept. • Computational results confirm the performance of the proposed two-phase method. Classification problems with imbalanced data are challenging because traditional classifiers tend to misclassify minority samples. This paper introduces a novel two-phase method in the categories of under-sampling data-level and ensemble-based approaches to tackle these problems. Compared to classic k-means clustering, the main property of our method is that the majority class is partitioned into clusters so that there are no minority samples in the convex-hull of the majority samples of each cluster, and at the same time, the size of each cluster is controlled. Thus, it is expected that by using our method, the general pattern of data in the feature space is kept, and the possibility of changing the data distribution is reduced. Computational results over a variety of imbalanced classification datasets confirm the superiority of our method over the existing methods from different metrics. [ABSTRACT FROM AUTHOR]