Search

The complexity of data has been increasing not only in the dimension of the data, but also in the interrelated class relationships. Pre-defined hierarchical structure of classes helps to model such relationships in a way, which divides the whole classification problem into a set of subtasks. Feature selection, known as a data preprocessing technique, aims to selecting useful feature to improve the performance of learning algorithms. However, the issue of close class relationships, especially under the same parent node in the hierarchical structure, is adverse to the feature selection. In this paper, we introduce a novel distance metric based feature selection method with large margin nearest neighbor strategy for the hierarchical classification problems. Different from other algorithms, our proposed method is to project the data into a low-dimensional feature space for learning correct metrics, and the weights of features will be indicated in the transformation matrix. A sparsity regularization and iterative optimization ensure that superior feature subsets can be selected. Experimental evaluation on several datasets with hierarchical class structure demonstrates that the proposed approach is comparable or better than some other state-of-the-art hierarchical feature selection methods. [ABSTRACT FROM AUTHOR]

Intelligent fault diagnosis is a rapidly evolving field within power engineering. Using gas-in-oil data is a reliable method for transformer fault diagnosis that has been widely adopted in the power industry. However, traditional machine learning methods often suffer from low diagnostic accuracy due to the lack of a clear and effective feature set for gas-in-oil data as well as an imbalance between classes of sample size. To overcome this challenge, this paper proposes a novel transformer fault diagnosis model that utilizes a Filter-Wrapper Combined Feature Selection method and an AdaBoost integrated weighted broad learning system (AdaBoost-WBLS). More specifically, the original data is expanded to extract meaningful features, and the Filter-Wrapper combined feature selection method is used to eliminate preliminary redundancy, relevance, and significance of current features. The Wrapper algorithm is then used for precise screening to obtain the optimal feature subset, which effectively improves the quality of transformer features. Furthermore, to address the issue of imbalanced transformer samples, an improved BLS and AdaBoost integration method is introduced, and a fault diagnosis model based on AdaBoost-WBLS is proposed. Compared with existing power transformer fault diagnosis methods, the proposed method has a more accurate and balanced effect on fault classification. Overall, this paper provides a comprehensive and effective approach to transformer fault diagnosis, which has important implications for the reliability and safety of power systems. [ABSTRACT FROM AUTHOR]

Air pollution prediction is a burning issue, as pollutants can harm human health. Traditional machine learning models usually aim to improve the overall prediction accuracy but neglect the accuracy for peak values. Moreover, these models are not interpretable. They fail to explain the interactions between various determining factors and their impacts on air pollution. In this paper, we propose a new Hybrid Interpretable Predictive Machine Learning model for the Particulate Matter 2.5 prediction, which carries two novelties. First, a hybrid model structure is constructed with deep neural network and Nonlinear Auto Regressive Moving Average with Exogenous Input model. Second, automatic feature generation and feature selection procedures are integrated into this hybrid model. The experimental results demonstrate the superiority of our model over other models in prediction accuracy for peak values and model interpretability. The proposed model reveals how PM2.5 prediction is estimated by historical PM2.5, weather, and season. The accuracies (measured by correlation coefficients) of 1, 3 and 6-hour-ahead prediction are 0.9870, 0.9332 and 0.8587, respectively. More importantly, the proposed approach presents a new interpretable machine learning framework for time series data, enabling to explain complex dependence of multimode inputs, and to build reliable predictive models. [ABSTRACT FROM AUTHOR]

Fuzzy rough set is a theoretical framework of fuzzy uncertainty management, and discernibility matrix offers a mathematical foundation for algorithm construction of feature learning. The approaches of fuzzy rough set and discernibility matrix have been successfully applied in single-label learning. However, few works have been done on investigating the foundation of fuzzy rough discernibility matrix on multi-label data. There will be two pivotal problems to be addressed when using fuzzy rough discernibility matrix for multi-label data analysis. One is how to extract sample-level and label-level correlations; and the other is how to utilize the discernibility matrix for algorithm construction. For this reason, in this paper the fuzzy rough discrimination matrix is introduced to deal with the problem of multi-label feature selection. First, the significance of labels in the label space is captured based on the label correlation. Labels with different significances contribute to different weights for measuring the similarity between samples. Hence, a sample similarity matrix in the label space is computed based on the label weighting strategy. Then, a framework of a fuzzy decision system is formalized, in which the discernibility matrix of fuzzy rough sets is introduced as a foundation to evaluate the sample discrimination ability of features. Under the discernibility matrix criterion, a multi-label learning algorithm is developed to select discriminative features from multi-label data. A series of experimental analysis verifies the effectiveness and efficiency of the proposed method. [ABSTRACT FROM AUTHOR]

Sparse Bayesian learning (SBL) provides state-of-the-art performance in accuracy, sparsity and probabilistic prediction for classification. In SBL, the regularization priors are automatically determined that avoids an exhaustive hyperparameter selection by cross-validation. However, scalability to large problems is a drawback of SBL because of the inversion of a potentially enormous covariance matrix for updating the regularization priors in every iteration. This paper develops an approximate SBL algorithm called ARP-SBL, where the regularization priors are approximated without inverting the covariance matrix. Therefore, our approach can easily scale up to problems with large data size or feature dimension. It alleviates the long training time and high memory complexity in SBL. Based on ARP-SBL, two scalable nonlinear SBL models: scalable relevance vector machine (ARP-RVM) and scalable sparse Bayesian extreme learning machine (ARP-SBELM) are developed for problems of large data size and large feature size respectively. Experiments on a variety of benchmarks have shown that the proposed models are with competitive accuracy compared to existing methods while i) converging faster; ii) requiring thousands of times less memory; iii) without exhaustive regularized hyperparameter selection; and iv) easily scaling up to large data size and high dimensional features. [ABSTRACT FROM AUTHOR]

In recent years, deep learning techniques have significantly improved the performance of single image super-resolution (SISR). However, this improvement is often achieved at the cost of introducing a large amount of parameters, which limits the real-world applications for SISR. In this paper, we propose a lightweight SISR network called Multi-scale Channel Attention Network for Image Super-Resolution (MCSN). Our contributions are threefold. First of all, the multi-scale feature fusion block (MSFFB) can extract multi-scale features by filters with different receptive fields. Secondly, the channel shuffle attention mechanism (CSAM) encourages the flow of the information across feature channels and enhances the ability of feature selection. Thirdly, the global feature fusion connection (GFFC) can effectively improve feature utilization. Extensive experiments demonstrate that the parameter amount of our method is reduced by 3/4 compared with the current state-of-the-art MSRN method, while both the subjective visual effect and objective quality of the reconstructed high-resolution images are significantly better. [ABSTRACT FROM AUTHOR]

This paper investigates a new bacterial colony-based feature selection algorithm to improve the classification accuracy of customers for personalized products recommendation. An attribute learning strategy is developed in this study to update the feature related population. Specifically, the features can be weighted according to their historic contributions to both the individual- and group-based subsets. Additionally, the frequency of appearance is also recorded for the feature candidates to improve the diversity of feature distribution and avoid the over-fitting. Based on the weight-based feature indexes and frequency of appearance records, the performance of feature subsets are enhanced by replacing the features being repeatedly appeared in a same vector. To explore the feasibility of the proposed method for the missing feature problems, the objective of the optimization is to minimize the classification error using the acceptable number of features. K-Nearest Neighbor is employed as the learning technique to cooperate with the proposed feature selection method. The effectiveness of the proposed feature selection method is demonstrated by performing test on the datasets from UCI machine learning repository and real-world data from Amazon customer reviews of products. Compared with other seven feature selections methods, the proposed feature selection algorithm outperforms the other algorithms by achieving higher classification accuracy rate using smaller features. [ABSTRACT FROM AUTHOR]

Breast cancer is one of the most frequently occurred cancers for females, and thus diagnosing breast cancer is very important. Neural dynamic algorithm (NDA) has been successfully applied in many fields with the characteristics of parallel computing and exponential convergence. However, there is no research on applying NDA-based neural network for pattern classification. In this paper, a novel voting convergent difference neural network (V-CDNN) is proposed. To do so, samples are firstly handled by feature selection, feature weighting and sample normalization. Secondly, the preprocessed samples are used to simultaneously and independently train several convergent difference neural networks in different types of mapping functions. Thirdly, in the testing process, voting strategy for these networks is applied to make diagnosis results more accurate and convincing. Being different from most existing neural networks, the proposed V-CDNN adopts neural dynamic learning algorithm, which greatly improves computation efficiency and increases accuracy rate of diagnosis. Experimental results verify that the proposed V-CDNN can achieve 100 % average diagnosis accuracy, which is the highest among existing state-of-the-art methods on the open data set. [ABSTRACT FROM AUTHOR]

Abstract Dealing with high-dimensional dataset has always been an important problem and feature selection is one of useful tools. In this paper, we develop a new filter based supervised feature selection method by combining Trace Ratio Criterion of Linear Discriminant Analysis (TRC-LDA) and group sparsity regularization. The filter based supervised feature selection method is a classifier-independent method while the TRC-LDA criterion is a recently developed criterion for dimensionality reduction that can well preserve discriminative information of dataset. However, there are seldom methods by utilizing TRC-LDA criterion for feature selection. On the other hand, imposing the l2,0-norm to the projection matrix of TRC-LDA will force some rows in it to be zero while keep other rows nonzero making the index of nonzero rows to be the selected features, however, l 2,0 -nom minimizing problem is NP-hard and intractable. To solve the above problem, in this paper, we develop a new method, namely, Trace Ratio Criterion Discriminative Feature Selection (TRC-DFS), for feature selection. The proposed TRC-DFS has imposed l 2,1 -norm, i.e. an approximation of l 2,0 -norm, to the projection matrix W of TRC-LDA to achieve feature selection. As a result, the proposed TRC-DFS can both achieve feature selection as well as capture the discriminative structure of data. We also extend the proposed method with l 2, p -norm (0 < p < 1) regularization to grasp more sparse property and develop an iterative approach to calculate the optimal solution, which is rigorously proved to be converged. Extensive simulations based on several real-world datasets verify the effectiveness of the proposed methods. [ABSTRACT FROM AUTHOR]

Classification and distinguishing of malware is key to predict the malicious attack, which is essential in self-driving systems. In order to handle large number of malware variants, many machine learning methods have been proposed. However, the accuracy and efficiency of multiple class classification of malware still remained inadequate to meet demand. In this paper, we propose a 4-LFE method to deal with the issues above. We extract multi-features from malicious programs by combining pixel and n-gram features. In the process of feature selection, we apply L1-L2 penalty into the Logistic Regression, then use LDA to reduce dimensions of malware features. Based on the selected features, we study the performance of classification on ten machine learning algorithms. We assess our approach's precision on a public dataset consisting 10,868 malware samples. Experimental results show our method could classify malware to their family with accuracy of 99.99%. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) has incorporated various automatic systems and frameworks to diagnose the severity of depression using hand-crafted features. However, process of feature selection needs domain knowledge and is still time-consuming and subjective. Deep learning technology has been successfully adopted for depression recognition. Most previous works pre-train the deep models on large databases followed by fine-tuning with depression databases (i.e., AVEC2013, AVEC2014). In the present paper we propose an integrated framework – Deep Local Global Attention Convolutional Neural Network (DLGA-CNN) for depression recognition, which adopts CNN with attention mechanism as well as weighted spatial pyramid pooling (WSPP) to learn a deep and global representation. Two branches are introduced: Local Attention based CNN (LA-CNN) focuses on the local patches, while Global Attention based CNN (GA-CNN) learns the global patterns from the entire facial region. To capture the complementary information between the two branches, Local–Global Attention-based CNN (LGA-CNN) is proposed. After feature aggregation, WSPP is used to learn the depression patterns. Comprehensive experiments on AVEC2013 and AVEC2014 depression databases have demonstrated that the proposed method is capable of mining the underlying depression patterns of facial videos and outperforms the most of the state-of-the-art video-based depression recognition approaches. [ABSTRACT FROM AUTHOR]

Hyperparameter optimization is a challenging process that has the potential to improve machine learning algorithms. Since it creates a remarkable computational burden for machine learning tasks, there have been few works coping with tuning strategies of a specific algorithm. In this paper, an improved Stochastic Gradient Descent (SGD) based on Fisher Maximization is developed for tuning hyperparameters of an Echo State Network (ESN) which has a wide range of applications. The results of the method are then compared with those of traditional Gradient Descent and Grid Search. According to the obtained results; 1) The scale of the data sets greatly affects the reliability of hyperparameter optimization results; 2) Feature selection is critical in terms of mean error of training when hyperparameter optimization is applied on some methods such as ESN; 3) SGD falls in a good local minima if Fisher Maximization is performed to find a good starting point. [ABSTRACT FROM AUTHOR]

The recently proposed Minimal Complexity Machine (MCM) learns a hyperplane classifier by minimizing a bound on the Vapnik-Chervonenkis (VC) dimension. Both the linear and kernel versions of the MCM solve a linear programming problem, in order to minimize a bound on the VC dimension. This paper proposes a new quadratic programming formulation, termed as the Quadratic MCM (QMCM), that minimizes a tighter bound on the VC dimension. We present two variants of the QMCM, that differ in the norm of the error vector being minimized. We also explore a scalable variant of the QMCM for large datasets using Stochastic Gradient Descent (SGD), and present the use of the QMCM as a viable featureselection method, in view of the the sparse nature of the models it learns. We compare the performance of the QMCM variants with LIBLinear, a linear Support Vector Machine (SVM) library; as well as against Pegasos and the linear MCM for large datasets, along with sequential feature selection methods and ReliefF. Our results validate the superiority of the QMCM in terms of statistically significant improvements on benchmark datasets from the UCI Machine Learning repository. [ABSTRACT FROM AUTHOR]

Data-driven facial animation has attracted considerable attention from both academic and industrial communities in recent years. Typically, the motion data used to animate the faces are derived from either 3D or 2D facial features whose positions on the face are determined according to experience. There still lacks an automatic approach to determine the optimal positions of the features to face deformation, and current face deformation methods are incapable of providing fine-grained local geometric characteristics. This paper proposes a novel scheme for face animation in which an active learning method based on Locally Linear Reconstruction algorithm is exploited to determine the optimal feature positions on the face for face deformation, and the Semi-Supervised Local Patch Alignment algorithm is subsequently used to deform the face with the selected features according to the optimal feature positions. The active learning model can be solved by a sequential approach, and the Semi-Supervised Local Patch Alignment model can be addressed by a least-square method. Experimental results on various types of faces demonstrate the superiority of the proposed scheme to existing approaches in both feature points selection and fine-grained local characteristics preservation. [ABSTRACT FROM AUTHOR]

For data with various complicated distribution in the original feature space, it is difficult to find the clusters of the data. Extreme learning machine (ELM) is famous for its universal approximation capability and the hidden space created by random nonlinear feature mapping. Existing ELM based clustering methods address this by constructing an embedding space, in which the cluster are easily revealed. A commonality of them is the final results have to be subsequently derived by k -means clustering. In this paper, we propose an unsupervised feature selection based extreme learning machine (UFSELM) for clustering, which integrates ELM with L2,1 norm regularization to remove the worthless hidden neurons and clusters the data directly without building an embedding. Specifically, the proposed method conducts feature selection by minimizing the L2,1 norm of output weights, and the clustering results is computed by eigendecomposition. By solving the formulated optimization problem in an iterative fashion, we improved the accuracy of clustering. We conducted experiments on several public datasets to demonstrate the effectiveness of the proposed method and further analyzed the properties of the proposed method. [ABSTRACT FROM AUTHOR]

High-dimensional data in many machine learning applications leads to computational and analytical complexities. Feature selection provides an effective way for solving these problems by removing irrelevant and redundant features, thus reducing model complexity and improving accuracy and generalization capability of the model. In this paper, we present a novel teacher-student feature selection (TSFS) method in which a 'teacher' (a deep neural network or a complicated dimension reduction method) is first employed to learn the best representation of data in low dimension. Then a 'student' network (a simple neural network) is used to perform feature selection by minimizing the reconstruction error of low dimensional representation. Although the teacher-student scheme is not new, to the best of our knowledge, it is the first time that this scheme is employed for feature selection. The proposed TSFS can be used for both supervised and unsupervised feature selection. This method is evaluated on different datasets and is compared with state-of-the-art existing feature selection methods. The results show that TSFS performs better in terms of classification and clustering accuracies and reconstruction error. Moreover, experimental evaluations demonstrate a low degree of sensitivity to parameter selection in the proposed method. [ABSTRACT FROM AUTHOR]

Keywords can express the main content of an article or a sentence. Keywords extraction is a critical issue in many Natural Language Processing (NLP) applications and can improve the performance of many NLP systems. The traditional methods of keywords extraction are based on machine learning or graph model. The performance of these methods is influenced by the feature selection and the manually defined rules. In recent years, with the emergence of deep learning technology, learning features automatically with the deep learning algorithm can improve the performance of many tasks. In this paper, we propose a deep neural network model for the task of keywords extraction. We make two extensions on the basis of traditional LSTM model. First, to better utilize both the historic and following contextual information of the given target word, we propose a target center-based LSTM model (TC-LSTM), which learns to encode the target word by considering its contextual information. Second, on the basis of TC-LSTM model, we apply the self-attention mechanism, which enables our model has an ability to focus on informative parts of the associated text. In addition, we also introduce a two-stage training method, which takes advantage of large-scale pseudo training data. Experimental results show the advantage of our method, our model beats all the baseline systems all across the board. And also, the two-stage training method is of great significance for improving the effectiveness of the model. [ABSTRACT FROM AUTHOR]

Clustering is an important method for data analysis. Up to now, how to develop an efficient clustering algorithm is still a critical issue. Unsupervised extreme learning machine is an effective neural network learning method which has a fast training speed. In this paper, a fuzzy granularity neighborhood extreme clustering algorithm which is based on extreme learning machine is proposed. We use fuzzy neighborhood rough set to develop a new feature selection method to eliminate redundant attributes and introduce the adaptive adjustment mechanism to solve the parameters of unsupervised extreme learning machine. Different from the existing clustering algorithms, the proposed algorithm can obtain a clustering result with minimum intra-cluster distance and maximum inter-cluster distance. The proposed algorithm and comparison algorithms are executed on the synthetic data sets and real data sets. The experimental results show that the proposed algorithm outperforms the comparison algorithms on the most data sets and the proposed algorithm is effective for clustering task. [ABSTRACT FROM AUTHOR]

Both subspace learning methods and feature selection methods are often used for removing irrelative features from high-dimensional data. Studies have shown that feature selection methods have interpretation ability and subspace learning methods output stable performance. This paper proposes a new unsupervised feature selection by integrating a subspace learning method (i.e., Locality Preserving Projection (LPP)) into a new feature selection method (i.e., a sparse feature-level self-representation method), aim at simultaneously receiving stable performance and interpretation ability. Different from traditional sample-level self-representation where each sample is represented by all samples and has been popularly used in machine learning and computer vision. In this paper, we propose to represent each feature by its relevant features to conduct feature selection via devising a feature-level self-representation loss function plus an ℓ 2 , 1 -norm regularization term. Then we add a graph regularization term (i.e., LPP) into the resulting feature selection model to simultaneously conduct feature selection and subspace learning. The rationale of the LPP regularization term is that LPP preserves the original distribution of data after removing irrelative features. Finally, we conducted experiments on UCI data sets and other real data sets and the experimental results showed that the proposed approach outperformed all comparison algorithms. [ABSTRACT FROM AUTHOR]

Recent years, domain adaptation has attracted much attention in the community of machine learning. In this paper, we mainly focus on the tasks of Joint Domain Matching and Classification (JDMC) under the framework of extreme learning machine (ELM). Specifically, our JDMC method is formulated by optimizing both the output-adapted transformation and the cross-domain classifier, which allows us to simultaneously (1) align the source domain and target domain in the feature space with correlation alignment, (2) minimize the discrepancy between the source and target domain, measured in terms of both marginal and conditional probability distribution in the mapped feature space, (3) select informative features which behave similarly in both domains for knowledge transfer by imposing ℓ 2,1 -norm on the output weights of ELM. In this respect, the proposed JDMC integrates the feature matching, feature selection and classifier design in a unified framework. Besides, an efficient alternative optimization strategy is exploited to solve the joint learning model. To evaluate the effectiveness of the proposed method, extensive experiments on several commonly used domain adaptation datasets are presented, the results show that the proposed method significantly outperforms the non-transfer ELM networks and consistently outperforms several state-of-art domain adaptation methods. [ABSTRACT FROM AUTHOR]

Abstract Sparse subspace learning has been demonstrated to be effective in data mining and machine learning. In this paper, we cast the unsupervised feature selection scenario as a matrix factorization problem from the viewpoint of sparse subspace learning. By minimizing the reconstruction residual, the learned feature weight matrix with the l 2,1 -norm and the non-negative constraints not only removes the irrelevant features, but also captures the underlying low dimensional structure of the data points. Meanwhile in order to enhance the model's robustness, l 1 -norm error function is used to resistant to outliers and sparse noise. An efficient iterative algorithm is introduced to optimize this non-convex and non-smooth objective function and the proof of its convergence is given. Although, there is a subtraction item in our multiplicative update rule, we validate its non-negativity. The superiority of our model is demonstrated by comparative experiments on various original datasets with and without malicious pollution. [ABSTRACT FROM AUTHOR]

Abstract As one of the dimensionality reduction methods, unsupervised feature selection aims to select a subset of original features without the labels of data instances. It becomes more and more important since large amounts of unlabelled high dimensional data need to be processed in many machine learning and data mining tasks. In this paper, we propose a novel unsupervised feature selection method via compact feature representation, with the local geometrical structure of data being preserved by using a dual Laplacian graph regularization term. In detail, different to many previous representation based methods which use the original data as representation dictionary, we propose to learn a feature dictionary subspace for compact and robust feature representation. During the dictionary learning process, the l 2,1 -norm is imposed on the combination coefficient matrix to select discriminate features for constructing the latent feature dictionary. Meanwhile, the local geometrical structure of original data is preserved from the perspectives of subspace learning and feature representation. In general, our method conducts dictionary learning and unsupervised feature selection simultaneously. We develop an efficient optimization algorithm based on Alternating Direction Method of Multipliers to solve the proposed optimization problem and experiments on various types of real world datasets are conducted to demonstrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Highlights • Theoretical framework for the comparison of feature selection methods. • Derivation of upper and lower bounds for the target objective functions. • Linking objective function bounds with feature types. • Distributional setting to highlight deficiencies of feature selection methods. • Identification of feature selection methods to be avoided and preferred. Abstract Feature selection problems arise in a variety of applications, such as microarray analysis, clinical prediction, text categorization, image classification and face recognition, multi-label learning, and classification of internet traffic. Among the various classes of methods, forward feature selection methods based on mutual information have become very popular and are widely used in practice. However, comparative evaluations of these methods have been limited by being based on specific datasets and classifiers. In this paper, we develop a theoretical framework that allows evaluating the methods based on their theoretical properties. Our framework is grounded on the properties of the target objective function that the methods try to approximate, and on a novel categorization of features, according to their contribution to the explanation of the class; we derive upper and lower bounds for the target objective function and relate these bounds with the feature types. Then, we characterize the types of approximations taken by the methods, and analyze how these approximations cope with the good properties of the target objective function. Additionally, we develop a distributional setting designed to illustrate the various deficiencies of the methods, and provide several examples of wrong feature selections. Based on our work, we identify clearly the methods that should be avoided, and the methods that currently have the best performance. [ABSTRACT FROM AUTHOR]

Abstract Feature selection is used in many applications in machine learning and bioinformatics. As a popular approach, feature selection can be implemented in the filter-manner based on the sparse solution of the l 1 regularization. Most study of the l 1 regularization concentrates on investigating the iteration solution of the problem or focuses on adapting sparsity to different applications. It is necessary to explore more deeply about how the sparsity learned with the l 1 regularization contributes to feature selection. In this paper, we make an effort to analyze the role of the l 1 regularization in feature selection from the perspective of information theory. We discover that the l 1 regularization contributes to minimizing the redundancy in feature selection. To avoid the complex computation of the l 1 optimization, we propose a novel feature selection algorithm, i.e. the Laplacian regularization based on mutual information (LRMI), which realizes the minimization of the redundancy in a new way, and incorporates the l 2 norm to achieve automatic grouping. Extensive experimental results demonstrate the superiority of LRMI over several traditional l 1 regularization based feature selection algorithms with less time consumption. [ABSTRACT FROM AUTHOR]

Abstract Unsupervised feature selection has been widely applied to machine learning and pattern recognition, as it does not require class labels. The majority of the popular unsupervised feature selection methods focus on various forms of reconstruction, and minimize the reconstruction residual by discarding features with low contributions. However, they cannot effectively preserve the data distribution in multiple subspaces, because the sample structure information is not substantially utilized to constrain the selected features. In this paper, we propose a low-rank structure preserving method for unsupervised feature selection (LRPFS) to address this shortcoming. The data matrix consisting selected features is assumed as a dictionary, which is learned by a low-rank constraint to preserve the subspace structure. Meanwhile, we further leverage the sparse penalty to remove the redundancy features, and thus obtain the discriminative features with intrinsic structures. In this way, the sample distribution can be preserved by low-rank constraint more precisely via using discriminative features. In turn, the refined sample structure boosts the selection of more representative features. The effectiveness of our method is supported by both theoretical and experimental results. [ABSTRACT FROM AUTHOR]

In this paper, we focus on unsupervised feature selection. As we have known, the combination of several feature units into a whole feature vector is broadly adopted for effective object representation, which may inevitably includes some irrelevant/redundant feature units or feature dimensions. Most of the traditional feature selection models can only select the feature dimensions without concerning the intrinsic relationship among different feature units. By taking into consideration the group sparsity of feature dimensions and feature units based on an ℓ 2 , 1 minimization, we propose a new unsupervised feature selection model, unsupervised deep sparse feature selection (UDSFS) in this paper. In comparison with the state-of-the-arts, our UDSFS model can not only select the most discriminative feature units but also assign proper weight to the useful feature dimensions concurrently; moreover, the efficiency and robustness of our UDSFS can be also improved without extracting the discarded irrelevant feature units. For model optimization, we introduce an efficient iterative algorithm to solve the non-smooth, convex model and obtain a global optimization with the convergence rate as O ( 1 / K 2 ) ( K is the iteration number). For the experiments, a new medical endoscopic image dataset, Abnormal Endoscopic Image Detection dataset (AEID), is built for evaluation; we also test our model using two public UCI datasets. Various experiments and comparisons with other state-of-the-arts justified the effectiveness and efficiency of our UDSFS model. [ABSTRACT FROM AUTHOR]

The goal of Multiple Kernel Learning (MKL) is to combine kernels derived from multiple sources in a data-driven way with the aim to enhance the accuracy of a target kernel machine. State-of-the-art methods of MKL have the drawback that the time required to solve the associated optimization problem grows (typically more than linearly) with the number of kernels to combine. Moreover, it has been empirically observed that even sophisticated methods often do not significantly outperform the simple average of kernels. In this paper, we propose a time and space efficient MKL algorithm that can easily cope with hundreds of thousands of kernels and more. The proposed method has been compared with other baselines (random, average, etc.) and three state-of-the-art MKL methods showing that our approach is often superior. We show empirically that the advantage of using the method proposed in this paper is even clearer when noise features are added. Finally, we have analyzed how our algorithm changes its performance with respect to the number of examples in the training set and the number of kernels combined. [ABSTRACT FROM AUTHOR]

Currently traffic noise has become an important factor that affects human health, and thus, an application able to classify vehicles on the basis of the sound they produce becomes important in the effort of fulfilling recommendations that aim at reducing traffic noise and improving intelligent transportation systems. This paper focuses on the problem of selecting those sound-describing features that make the vehicle classifier work properly. In particular, the goal of this paper is to evaluate the feasibility of a novel feature selection method based on a special class of Genetic Algorithm (with restricted search) hybridized with a Extreme Learning Machine. Because of its great generalization performance at a very fast learning speed, the Extreme Learning Machine plays the key role of providing the fitness of candidate solutions in each generation of the Genetic Algorithm. After a number of experiments comparing its performance to that of other fast learning algorithms, our approach has been found to be the most feasible for the application at hand. The proposed method helps the Extreme Learning Machine-based classifier to increase its performance from a mean probability of correct classification of 74.83% (with no feature selection) up to 93.74% (when using the optimum subset of selected features). [ABSTRACT FROM AUTHOR]

Research on feature relevance and feature selection problems goes back several decades, but the importance of these areas continues to grow as more and more data becomes available, and machine learning methods are used to gain insight and interpret, rather than solely to solve classification or regression problems. Despite the fact that feature relevance is often discussed, it is frequently poorly defined, and the feature selection problems studied are subtly different. Furthermore, the problem of finding all features relevant for a classification problem has only recently started to gain traction, despite its importance for interpretability and integrating expert knowledge. In this paper, we attempt to unify commonly used concepts and to give an overview of the main questions and results. We formalize two interpretations of the all-relevant problem and propose a polynomial method to approximate one of them for the important hypothesis class of linear classifiers, which also enables a distinction between strongly and weakly relevant features. [ABSTRACT FROM AUTHOR]

Feature selection has playing an important role in many pattern recognition and machine learning applications, where meaningful features are desired to be extracted from high dimensional raw data and noisy ones are expected to be eliminated. l 2,1 -norm regularization based Robust Feature Selection (RFS) has extracted a lot of attention due to its efficiency and high performance of joint sparsity. In this paper, we propose a more general framework for robust and discriminative multi-class feature selection. Four types of weighting, which are based on correlation information between features and labels, are adopted to strengthen the discriminative performance of l 2,1 -norm joint sparsity. F-norm regularization, which is extended from multi-class Elastic Net, is added to improve the stability of the method. An efficient algorithm and its corresponding convergence proof are provided. Experimental results on several two-class and multi-class datasets are performed to verify the effectiveness of the proposed feature selection method. [ABSTRACT FROM AUTHOR]

Feature selection is an important issue in pattern recognition and machine learning, which aims at selecting relevant features from a set of candidates. Obviously, the establishment of proper criteria to evaluate the relevance of features is pivotal to the selection. In this paper, a criterion is proposed to assess the relevance of individual input features based on Radial Basis Function Neural Networks (RBFNNs) involved in solving classification and regression problems. The criterion takes a quantified output sensitivity of RBFNNs to input variation as a measure, which is defined as a mathematical expectation and can in a statistic sense reflect the effect of an RBFNN's input variation on its output. The basic idea is that a well-trained RBFNN can capture relevant features of the problem it deals with and thus become more sensitive to the variation of those input features that make more contributions to the RBFNN's behavior. Since the sensitivity is difficult to exactly compute, a numerical integral technique is employed to approximately compute the sensitivity. Experimental results on several artificial and real datasets show that our proposed feature selection approach works well. [ABSTRACT FROM AUTHOR]

Assessing liver fibrosis with chronic hepatitis B (CHB) in patients is quite important. Some non-invasive approaches for evaluating liver fibrosis include blood tests and ultrasound elastography. How to effectively combine multiple methods to improve the diagnostic performance remains a challenging problem. The main goal of this paper is to assess and stage liver fibrosis in CHB using feature selection and machine learning methods based on multimodal data. Popular machine learning approaches (e.g., support vector machine (SVM)) and feature selection (FS) were explored to stage the CHB. 16 volunteers and 92 patients with CHB were investigated for liver fibrosis staging based on transient elastography (TE) and acoustical radiation force impulse imaging (ARFI) data. The accuracy of the staging result using FS and a SVM classifier was an accuracy of 90.68% for significant fibrosis (≥F2) and an accuracy of 93.52% for cirrhosis (F4), respectively. The proposed method also increased the sensitivity, specificity, and area under curve (AUC) values for both significant fibrosis and cirrhosis diagnosis, which is very promising for staging liver fibrosis from multimodal information. It outperforms any single method and their linear combination and also achieves a state-of-the-art performance. [ABSTRACT FROM AUTHOR]

Naïve Bayes, k-nearest neighbors, Adaboost, support vector machines and neural networks are five among others commonly used text classifiers. Evaluation of these classifiers involves a variety of factors to be considered including benchmark used, feature selections, parameter settings of algorithms, and the measurement criteria employed. Researchers have demonstrated that some algorithms outperform others on some corpus, however, inconsistency of human labeling and high dimensionality of feature spaces are two issues to be addressed in text categorization. This paper focuses on evaluating the five commonly used text classifiers by using an automatically generated text document collection which is labeled by a group of experts to alleviate subjectivity of human category assignments, and at the same time to examine the influence of the number of features on the performance of the algorithms. [ABSTRACT FROM AUTHOR]

Abstract: One of the most important issues in the PQ assessment is diagnosis of Power Quality Disturbances (PQDs) using an effectual low computational burden strategy. We recommend a new approach for PQ analysis that addresses several major problems of prior works, including algorithm execution time, computational complexity, and accuracy. This paper suggests an effective and comprehensive method, so-called “integrated approach”, for extracting features using integration of discrete wavelet transform and hyperbolic S transform. Moreover, a comparative assessment of PQDs recognition using various combinations of different Feature Selection (FS) and classification methods is presented. FS can reduce the dimension of feature space which leads to better performance of detection system. Four well-known FS techniques namely modified relief, mutual information, sequential forward selection, sequential backward selection and three benchmark classifiers, are considered. The particle swarm optimization is used to obtain optimal parameters of these classifiers. The key attribute of this paper is that it yields good time–frequency resolution with low computation burden for optimal PQ monitoring structure. Empirical results show that the proposed structures can yield an automatic online/offline monitoring of PQ with sparser structures and less computational execution time, both in the training and recognition phases, without sacrificing generality of performance. [Copyright &y& Elsevier]

As an important pre-processing stage in many machine learning and pattern recognition domains, feature selection deems to identify the most discriminate features for a compact data representation. As typical feature selection methods, Lasso and its variants using the l 1 -norm based regularization have received much attention in recent years. However, most of existing l 1 -norm based sparse feature selection methods ignore the structure information of data or only consider the pairwise relationships among samples. In this paper, we propose a hypergraph regularized sparse feature learning method, where the high-order relationships among samples are modeled and incorporated into the learning process. Specifically, we first construct a hypergraph with multiple hyperedges to capture the high-order relationships among samples, followed by the computation of a hypergraph Laplacian matrix. Then, we propose a hypergraph regularization term, and a hypergraph regularized Lasso model. We conduct a series of experiments on a number of data sets from UCI machine learning repository, and two real-world neuroimaging based classification tasks. Experimental results demonstrate that the proposed method achieves promising classification results, compared with several well known feature selection approaches. [ABSTRACT FROM AUTHOR]

one of the most important tasks of Mars rover, a robot which explores the Mars surface, is the process of automatic segmentation of images taken by front-line Panoramic Camera (Pancam). This procedure is highly significant since the transformation cost of images from Mars to earth is extremely high. Also, image analysis may help Mars rover for its navigation and localization. In this paper, a new feature vector including wavelet and color features for Mars images is proposed. Then, this feature vector is presented for extreme learning machine (ELM) classifier which leads to a high accuracy pixel classifier. It is shown that this system statistically outperforms support vector machine (SVM) and k-nearest neighbours (KNNs) classifiers with respect to both accuracy and run time. After that, dimension reduction in feature space is done by two proposed feature section algorithms based on ant colony optimization (ACO) to decrease the time complexity which is very important in Mars on-board applications. In the first proposed feature selection algorithm, the same feature subset is selected among the feature vector for all pixel classes, while in the second proposed algorithm, the most significant features are selected for each pixel class, separately. Proposed pixel classifier with complete feature set outperforms prior methods by 6.44% and 5.84% with respect to average Fmeasure and accuracy, respectively. Finally, proposed feature selection methods decrease the feature vector size up to 76% and achieves Fmeasure and accuracy of 91.72% and 91.05%, respectively, which outperforms prior methods with 87.22% and 86.64%. [ABSTRACT FROM AUTHOR]

Feature selection as one of the key problems of data preprocessing is a hot research topic in pattern recognition, machine learning, and data mining. Evaluating the relevance between features based on information theory is a popular and effective method. However, very little research pays attention to the distinguishing ability of feature, i.e., the degree of a feature distinguishes a given sample with other samples. In this paper, we propose a new feature selection method based on the distinguishing ability of feature. First, we define the concept of maximum-nearest-neighbor, and use this concept to discriminate the nearest neighbors of samples. Then, we present a new measure method for evaluating the quality of feature. Finally, the proposed algorithm is tested on benchmark datasets. Experimental results show that the proposed algorithm can effectively select a discriminative feature subset, and performs as well as or better than other popular feature selection algorithms. [ABSTRACT FROM AUTHOR]

The Sparse Bayesian Extreme Learning Machine (SBELM) has been recently proposed to reduce the number of units activated on the hidden layer. To deal with high-dimensional data, a novel sparse Bayesian Double-Layer ELM (DL-ELM) is proposed in this paper. The first layer of the proposed DL-ELM is based on a set of SBELM subnetworks which are separately applied to the features on the input layer. The second layer consists in a set of weight parameters to determine the contribution of each feature in the output. Adopting a Bayesian approach and using Gaussian priors, the proposed method is sparse in both the hidden layer (of SBELM subnetworks) and the input layer. Sparseness in the input layer (i.e. pruning of irrelevant features) is achieved by decaying weights on the second layer to zero, such that the contribution of the corresponding input feature is deactivated. The proposed framework then enables simultaneous feature selection and classifier design at the training time. Experimental comparisons on real benchmark data sets show that the proposed method benefits from efficient feature selection ability while providing a compact classification model of good accuracy and generalization properties. [ABSTRACT FROM AUTHOR]

As more features are introduced in pattern recognition and machine learning applications, feature selection remains a critically important task to find the most compact representation of data, especially in unsupervised learning scenarios without enough class labels. Although there are a number of unsupervised feature selection methods in the literature, most of existing methods only focus on using conventional distances (e.g., Euclidean distance) to measure the similarity between two samples, which could not capture the dynamic structure of data due to the static characteristics of conventional distances. To reflect the dynamic structure of data, in this paper, we propose a set of effective distance-based feature selection methods, where a probabilistically motivated effective distance is used to measure the similarity of samples. Specifically, we first develop a sparse representation-based algorithm to compute the effective distance. Then, we propose three new filter-type unsupervised feature selection methods using effective distance, including an effective distance-based Laplacian Score (EDLS), and two effective distance-based Sparsity Scores (i.e., EDSS-1, and EDSS-2). Experimental results of clustering and classification tasks on a series of benchmark data sets show that our effective distance-based feature selection methods can achieve better performance than conventional methods using Euclidean distance. [ABSTRACT FROM AUTHOR]

Nowadays, data always have multiple representations, and a good feature representation usually leads to a good clustering performance. Existing multi-view clustering works generally integrate multiple complementary information to gain better clustering performance rather than relying on a single view. However, these works usually focus on the combination of information rather than improving the feature representation capability of each view. As a new method, extreme learning machine (ELM) has excellent feature representation capability, easy parameter selection, and promising performance in various clustering tasks. This paper proposes a novel multi-view clustering framework with ELM to further improve clustering performance, and implements three algorithms based on this framework. In this framework, the normalized features of each individual view are mapped onto a higher dimensional feature space by the ELM random mapping. Afterwards, the unsupervised multi-view clustering is performed in this feature space. Thus far, this is the first work on multi-view clustering with ELM. Numerous baseline methods on five real-world datasets are empirically compared to show the effectiveness of the proposed algorithms. As indicated, the proposed algorithms yield superior clustering performance when compared with several state-of-art multi-view clustering methods in recent literatures. [ABSTRACT FROM AUTHOR]

The features used in many social media analysis-based applications are usually of very high dimension. Feature selection offers several advantages in highly dimensional cases. Recently, multi-task feature selection has attracted much attention, and has been shown to often outperform the traditional single-task feature selection. Current multi-task feature selection methods are either supervised or unsupervised. In this paper, we address the semi-supervised multi-task feature selection problem. We firstly introduce manifold regularization in multi-task feature selection to utilize the limited number of labeled samples and the relatively large amount of unlabeled samples. However, the graph constructed in manifold regularization from a single feature representation (view) may be unreliable. We thus propose to construct the graph using the heterogeneous feature representations from multiple views. The proposed method is called manifold regularized multi-view feature selection (MRMVFS), which can exploit the label information, label relationship, data distribution, as well as correlation among different kinds of features simultaneously to boost the feature selection performance. All these information are integrated into a unified learning framework to estimate feature selection matrix, as well as the adaptive view weights. Experimental results on three real-world image datasets, NUS-WIDE, Flickr and Animal, demonstrate the effectiveness and superiority of the proposed MRMVFS over other state-of-the-art feature selection methods. [ABSTRACT FROM AUTHOR]

Many organizations recognize the necessities of utilizing sophisticated tools and systems to protect their computer networks and reduce the risk of compromising their information. Although many machine-learning-based data classification algorithm has been proposed in network intrusion detection problem, each of them has its own strengths and weaknesses. In this paper, we propose an effective intrusion detection framework by using a new adaptive, robust, precise optimization method, namely, time-varying chaos particle swarm optimization (TVCPSO) to simultaneously do parameter setting and feature selection for multiple criteria linear programming (MCLP) and support vector machine (SVM). In the proposed methods, a weighted objective function is provided, which takes into account trade-off between the maximizing the detection rate and minimizing the false alarm rate, along with considering the number of features. Furthermore, to make the particle swarm optimization algorithm faster in searching the optimum and avoid the search being trapped in local optimum, chaotic concept is adopted in PSO and time varying inertia weight and time varying acceleration coefficient is introduced. The performance of proposed methods has been evaluated by conducting experiments with the NSL-KDD dataset, which is derived and modified from well-known KDD cup 99 data sets. The empirical results show that the proposed method performs better in terms of having a high detection rate and a low false alarm rate when compared with the obtained results using all features. [ABSTRACT FROM AUTHOR]

Recent studies have witnessed the effectiveness of L 2 , 1 -norm based methods on AD/MCI diagnosis. Nonetheless, most of them suffer from the following three main problems: (1) L 2 , 1 -norm based loss function does not take into account different distances between target labels and prediction values; (2) L 2 , 1 -norm based feature selection does not possess sufficient flexibility to adapt to different types of data sources and select more informative features; (3) intrinsic correlation between the processes of feature selection and classification (or regression) are inevitably ignored. In this paper, we propose a novel method which incorporates additional flexibility and adaptability by employing the more generalized L 2 , p -norm based prediction loss function and L 2 , q -norm based feature selection, as well as utilizes a joint model to perform feature selection and classification simultaneously. Besides, we introduce a regularizer to preserve local structure information between samples in the original feature space and prediction values in the projected space. In order to validate the effectiveness of the proposed method, we conducted extensive experiments on the ADNI dataset, and showed that the proposed method enhanced the performance of disease status classification, compared to the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

In this paper, we explore the potential of extreme learning machine (ELM) and kernel ELM (KELM) for early diagnosis of Parkinson’s disease (PD). In the proposed method, the key parameters including the number of hidden neuron and type of activation function in ELM, and the constant parameter C and kernel parameter γ in KELM are investigated in detail. With the obtained optimal parameters, ELM and KELM manage to train the optimal predictive models for PD diagnosis. In order to further improve the performance of ELM and KELM models, feature selection techniques are implemented prior to the construction of the classification models. The effectiveness of the proposed method has been rigorously evaluated against the PD data set in terms of classification accuracy, sensitivity, specificity and the area under the ROC (receiver operating characteristic) curve (AUC). Compared to the existing methods in previous studies, the proposed method has achieved very promising classification accuracy via 10-fold cross-validation (CV) analysis, with the highest accuracy of 96.47% and average accuracy of 95.97% over 10 runs of 10-fold CV. [ABSTRACT FROM AUTHOR]

Each example in a multi-label dataset is associated with multiple labels, which are often correlated. Learning from this data can be improved when dimensionality reduction tasks, such as feature selection, are applied. The standard approach for multi-label feature selection transforms the multi-label dataset into single-label datasets before using traditional feature selection algorithms. However, this approach often ignores label dependence. In this work, we propose an alternative method, LCFS, that constructs new labels based on relations between the original labels. By doing so, the label set from the data is augmented with second-order information before applying the standard approach. To assess LCFS, an experimental evaluation using Information Gain as a measure to estimate the importance of features was carried out on 10 benchmark multi-label datasets. This evaluation compared four LCFS settings with the standard approach, using random feature selection as a reference. For each dataset, the performance of a feature selection method is estimated by the quality of the classifiers built from the data described by the features selected by the method. The results show that a simple LCFS setting gave rise to classifiers similar to, or better than, the ones built using the standard approach. Furthermore, this work also pioneers the use of the systematic review method to survey the related work on multi-label feature selection. The summary of the 99 papers found promotes the idea that exploring label dependence during feature selection can lead to good results. [ABSTRACT FROM AUTHOR]

Sparse representation, a locality-based data representation method, leads to promising results in many scientific and engineering fields. Meanwhile in the study of feature selection, locality preserving is widely recognized as an effective measurement criterion. In this paper, we introduce l 1 -norm driven sparse representation into feature selection, and propose a novel joint feature weights learning algorithm, named sparse discriminative feature weights (SDFW). SDFW assigns the highest score to the feature that has the smallest difference between within-class reconstruction residual and between-class reconstruction residual in the space of selected features. It possesses the following advantages: (1) compared with feature selection methods based on k nearest neighbors, SDFW automatically (vs. manually) determines neighborhood for individual sample; (2) compared with conventional heuristic feature search which selects features individually, SDFW selects feature subset in batch mode. Extensive experiments on different data types demonstrate the effectiveness of SDFW. [ABSTRACT FROM AUTHOR]

Feature selection is a key step in many machine learning applications, such as categorization, and clustering. Especially for text data, the original document-term matrix is high-dimensional and sparse, which affects the performance of feature selection algorithms. Meanwhile, labeling training instance is time-consuming and expensive. So unsupervised feature selection algorithms have attracted more attention. In this paper, we propose an unsupervised feature selection algorithm through R ̲ andom P ̲ rojection and G ̲ ram- G ̲ chmidt O ̲ rthogonalization (RP-GSO) from the word co-occurrence matrix. The RP-GSO algorithm has three advantages: (1) it takes as input dense word co-occurrence matrix, avoiding the sparseness of original document-term matrix; (2) it selects “basis features” by Gram–Schmidt process, guaranteeing the orthogonalization of feature space; and (3) it adopts random projection to speed up GS process. Extensive experimental results show our proposed RP-GSO approach achieves better performance comparing against supervised and unsupervised feature selection methods in text classification and clustering tasks. [ABSTRACT FROM AUTHOR]

High Efficiency Video Coding (HEVC) standard adopts several new approaches to achieve higher coding efficiency (approximately 50% bit-rate reduction) compared to its predecessor H.264/AVC with same perceptual image quality. Huge computational time has also increased due to the algorithmic complexity of HEVC compared to H.264/AVC. However, it is really a demanding task to reduce the encoding time while preserving the similar quality of the video sequences. In this paper, we propose a novel efficient intermode selection technique and incorporate into HEVC framework to predict motion estimation and motion compensation modes between current and reference blocks and perform faster inter mode selection based on three dissimilar motion types in divergent video sequences. Instead of exploring and traversing all the modes exhaustively, we merely select a subset of candidate modes and the final mode from the selected subset is determined based on their lowest Lagrangian cost function. The experimental results reveal that average encoding time can be downscaled by 40% with similar rate-distortion performance compared to the exhaustive mode selection strategy in HEVC. [ABSTRACT FROM AUTHOR]

This paper presents a novel computer-aided diagnosis (CAD) system for the diagnosis of breast cancer based on extreme learning machine (ELM). In view of a mammographic image, it is first eliminated interference in the preprocessing stages. Then, the preprocessed images are segmented by the level set model we proposed. Subsequently, a model of multidimensional feature vectors is built. Since not every feature vector contributes to the improvement of performance, feature selection is done by the combination of support vector machine (SVM) and extreme learning machine (ELM). Finally, an optimal subset of feature vectors is inputted into the classifiers for distinguishing malignant masses from benign ones. We also compare our breast mass classification approach based on ELM with several state-of-the-art classification models, and the results show that the proposed CAD system not only has good performance in terms of specificity, sensitivity and accuracy, but also achieves a significant reduction in training time compared with SVM and particle swarm optimization-support vector machine (PSO-SVM). Ultimately, our system achieves the better performance with average accuracy of 96.02% which indicates that the proposed segmentation model, the utilization of selected feature vectors and the effective classifier ELM provide satisfactory system. [ABSTRACT FROM AUTHOR]

In recent years, there has been a great interest in prostate segmentation, which is an important and challenging task for CT image guided radiotherapy. In this paper, a learning-based segmentation method via joint transductive feature selection and transductive regression is presented, which incorporates the physician׳s simple manual specification (only taking a few seconds), to aid accurate segmentation, especially for the case with large irregular prostate motion. More specifically, for the current treatment image, experienced physician is first allowed to manually assign the labels for a small subset of prostate and non-prostate voxels, especially in the first and last slices of the prostate regions. Then, the proposed method follows the two step: in prostate-likelihood estimation step, two novel algorithms, tLasso and wLapRLS, will be sequentially employed for transductive feature selection and transductive regression, respectively, aiming to generate the prostate-likelihood map. In multi-atlases based label fusion step, the final segmentation result will be obtained according to the corresponding prostate-likelihood map and the previous images of the same patient. The proposed method has been substantially evaluated on a real prostate CT dataset including 24 patients with 330 CT images, and compared with several state-of-the-art methods. Experimental results show that the proposed method outperforms the state-of-the-arts in terms of higher Dice ratio, higher true positive fraction, and lower centroid distances. Also, the results demonstrate that simple manual specification can help improve the segmentation performance, which is clinically feasible in real practice. [ABSTRACT FROM AUTHOR]