Feature Subset Research Articles

Mutual information-based feature selection for inverse mapping parameter updating of dynamical systems

AbstractA digital twin should be and remain an accurate model representation of a physical system throughout its operational life. To this end, we aim to update (physically interpretable) parameters of such a model in an online fashion. Hereto, we employ the inverse mapping parameter updating (IMPU) method that uses an artificial neural network (ANN) to map features, extracted from measurement data, to parameter estimates. This is achieved by training the ANN offline on simulated data, i.e., pairs of known parameter value sets and sets of features extracted from corresponding simulations. Since a plethora of features (and feature types) can be extracted from simulated time domain data, feature selection (FS) strategies are investigated. These strategies employ the mutual information between features and parameters to select an informative subset of features. Hereby, accuracy of the parameters estimated by the ANN is increased and, at the same time, ANN training and inference computation times are decreased. Additionally, Bayesian search-based hyperparameter tuning is employed to enhance performance of the ANNs and to optimize the ANN structure for various FS strategies. Finally, the IMPU method is applied to a high-tech industrial use case of a semi-conductor machine, for which measurements are performed in closed-loop on the controlled physical system. This system is modeled as a nonlinear multibody model in the Simscape multibody environment. It is shown that the model updated using the IMPU method simulates the measured system more accurately than a reference model of which the parameter values have been determined manually.

Radiomic study of antenatal prediction of severe placenta accreta spectrum from MRI.

We previously demonstrated the potential of radiomics for the prediction of severe histological placenta accreta spectrum (PAS) subtypes using T2-weighted MRI. We aim to validate our model using an additional dataset. Secondly, we explore whether the performance is improved using a new approach to develop a new multivariate radiomics model. Multi-centre retrospective analysis was conducted between 2018 and 2023. Inclusion criteria: MRI performed for suspicion of PAS from ultrasound, clinical findings of PAS at laparotomy and/or histopathological confirmation. Radiomic features were extracted from T2-weighted MRI. The previous multivariate model was validated. Secondly, a 5-radiomic feature random forest classifier was selected from a randomized feature selection scheme to predict invasive placenta increta PAS cases. Prediction performance was assessed based on several metrics including area under the curve (AUC) of the receiver operating characteristic curve (ROC), sensitivity, and specificity. We present 100 women [mean age 34.6 (±3.9) with PAS], 64 of whom had placenta increta. Firstly, we validated the previous multivariate model and found that a support vector machine classifier had a sensitivity of 0.620 (95% CI: 0.068; 1.0), specificity of 0.619 (95% CI: 0.059; 1.0), an AUC of 0.671 (95% CI: 0.440; 0.922), and accuracy of 0.602 (95% CI: 0.353; 0.817) for predicting placenta increta. From the new multivariate model, the best 5-feature subset was selected via the random subset feature selection scheme comprised of 4 radiomic features and 1 clinical variable (number of previous caesareans). This clinical-radiomic model achieved an AUC of 0.713 (95% CI: 0.551; 0.854), accuracy of 0.695 (95% CI 0.563; 0.793), sensitivity of 0.843 (95% CI 0.682; 0.990), and specificity of 0.447 (95% CI 0.167; 0.667). We validated our previous model and present a new multivariate radiomic model for the prediction of severe placenta increta from a well-defined, cohort of PAS cases. Radiomic features demonstrate good predictive potential for identifying placenta increta. This suggests radiomics may be a useful adjunct to clinicians caring for women with this high-risk pregnancy condition.

Unsupervised feature selection based on bipartite graph and low-redundant regularization

Unsupervised feature selection (UFS) has attracted increasing attention because of the difficulty and high cost of obtaining data labels. Since the ignorance of redundancy between features and the use of a fixed similarity matrix, current UFS algorithms do not perform well. Moreover, because of the high computational complexity, existing UFS algorithms usually cannot handle large-scale data. A novel UFS method is proposed to solve these problems through bipartite graph learning with low-redundant regularization (BGLR). Firstly, BGLR uses a variance-based de-correlation anchor selection method to ensure that selected anchors can be evenly distributed across classes. Secondly, BGLR builds an adaptive bipartite graph by using selected anchors and original data in the projection space. L2,0-norm constraint is applied to the projection matrix. Thus, the feature subset can be obtained directly. Finally, a regularization term is applied to control the redundancy of selected features. Moreover, BGLR is solved by an iterative method, which can converge quickly. Experiments are performed on public datasets in comparison with some relevant methods. The experimental results demonstrate the BGLR’s effectiveness.

Maximal cliques-based hybrid high-dimensional feature selection with interaction screening for regression

Studies on feature selection have been extensively conducted in the literature, as it plays a significant role in both supervised and unsupervised machine learning tasks. Since the bulk of features in high-dimensional data sets might not be significant, feature selection plays a key role in removing unimportant variables and improving prediction and data analysis performance. Many of the current feature selection methods, meanwhile, become ineffective when used on contemporary datasets, which exhibit an escalating number of features in relation to sample size. This paper introduces a novel supervised feature selection method for regression problems. The proposed algorithm is called maximal Clique with Interaction Screening (ISClique). The ISClique algorithm’s overall structure can be described in two steps. Initially, a filter approach is used to select relevant features from an initial feature space and examine the different interactions between them. This is done using an innovative coefficient based on Kendall’s tau and partial Kendall’s tau. Secondly, the maximal clique strategy is applied as a wrapper to the selected set from the previous step to construct subsets of features. The most optimal subset that minimizes prediction error is selected. The proposed method integrates the advantages of graph theory with feature screening. Additionally, because the criteria employed in developing the ISClique method accommodate variable heterogeneity, this method is equally suitable for classification tasks. The proposed hybrid approach has been evaluated through applications involving various simulation scenarios and real datasets. Experimental findings demonstrate the advantages of ISClique over comparable methods.

Multi-Objective Unsupervised Feature Selection and Cluster Based on Symbiotic Organism Search

Unsupervised learning is a type of machine learning that learns from data without human supervision. Unsupervised feature selection (UFS) is crucial in data analytics, which plays a vital role in enhancing the quality of results and reducing computational complexity in huge feature spaces. The UFS problem has been addressed in several research efforts. Recent studies have witnessed a surge in innovative techniques like nature-inspired algorithms for clustering and UFS problems. However, very few studies consider the UFS problem as a multi-objective problem to find the optimal trade-off between the number of selected features and model accuracy. This paper proposes a multi-objective symbiotic organism search algorithm for unsupervised feature selection (SOSUFS) and a symbiotic organism search-based clustering (SOSC) algorithm to generate the optimal feature subset for more accurate clustering. The efficiency and robustness of the proposed algorithm are investigated on benchmark datasets. The SOSUFS method, combined with SOSC, demonstrated the highest f-measure, whereas the KHCluster method resulted in the lowest f-measure. SOSFS effectively reduced the number of features by more than half. The proposed symbiotic organisms search-based optimal unsupervised feature-selection (SOSUFS) method, along with search-based optimal clustering (SOSC), was identified as the top-performing clustering approach. Following this, the SOSUFS method demonstrated strong performance. In summary, this empirical study indicates that the proposed algorithm significantly surpasses state-of-the-art algorithms in both efficiency and effectiveness. Unsupervised learning in artificial intelligence involves machine-learning techniques that learn from data without human supervision. Unlike supervised learning, unsupervised machine-learning models work with unlabeled data to uncover patterns and insights independently, without explicit guidance or instruction.

An improved equilibrium optimization algorithm for feature selection problem in network intrusion detection

In this paper, an enhanced equilibrium optimization (EO) version named Levy-opposition-equilibrium optimization (LOEO) is proposed to select effective features in network intrusion detection systems (IDSs). The opposition-based learning (OBL) approach is applied by this algorithm to improve the diversity of the population. Also, the Levy flight method is utilized to escape local optima. Then, the binary rendition of the algorithm called BLOEO is employed to feature selection in IDSs. One of the main challenges in IDSs is the high-dimensional feature space, with many irrelevant or redundant features. The BLOEO algorithm is designed to intelligently select the most informative subset of features. The empirical findings on NSL-KDD, UNSW-NB15, and CIC-IDS2017 datasets demonstrate the effectiveness of the BLOEO algorithm. This algorithm has an acceptable ability to effectively reduce the number of data features, maintaining a high intrusion detection accuracy of over 95%. Specifically, on the UNSW-NB15 dataset, BLOEO selected only 10.8 features on average, achieving an accuracy of 97.6% and a precision of 100%.

A hybrid multi-measure and improved UMAP approach for train traction motor bearing protection state assessment

Abstract Different protection states have different protection effects on bogie traction motor bearings under road transportation conditions, which directly affects the stability and safety of later vehicle operations. The motor bearing monitoring system usually adopts the time-frequency domain and other indicators to accurately describe its state, and the selection of specific indicators usually originates from the experience of domain experts. To explore the weak and effective features hidden in the data more deeply, this paper researches feature selection and fusion, and proposes a hybrid multi-measure and improved UMAP approach for train traction motor bearing protection state assessment. Firstly, a platform is built to collect multiple sensor signals of traction motor vibration during road transportation and the time-frequency domain feature set is extracted. Then, by comparing the feature extraction effects of 10 single-measure models, a hybrid multi-measure feature selection model and its evaluation indexes, which comprehensively consider the combined distance measure, correlation, and information measure, are constructed to optimize the best hybrid multimetric and the feature set corresponding to the optimal strategy. To perform multi-sensor feature fusion, the PSO-optimized UMAP algorithm is used for feature fusion of the feature set and compared with eight other algorithms. Finally, the results of feature selection and fusion are input into eight classifiers for performance comparison, and the method achieves more than 85% accuracy in distinguishing different protection states, obviously due to the original feature subset. The experimental results prove the effectiveness of the method proposed in this paper, and it provides a theoretical basis and technical reference for the protection scheme of the traction motor and the design and selection of bearings.

Hybridizing flower pollination algorithm with particle swarm optimization for enhancing the performance of IPv6 intrusion detection system

Internet Protocol version 6 (IPv6) is the most recent iteration of IP, designed to accommodate hundreds of thousands of devices with unique IP addresses. IPv6 introduces new features, such as the Neighbor Discovery Protocol (NDP) and Address Auto-configuration Scheme. For its effective operation, IPv6 relies on several protocols, including ICMPv6, which carries significant responsibilities. Similar to IPv4, IPv6 is susceptible to various attacks, including newer types like DDoS attacks executed via ICMPv6 messages, posing serious security and financial threats. Consequently, an Intrusion Detection System (IDS) is essential to safeguard against these attacks. IDS continuously evolve to incorporate features that can accurately detect such threats. However, feature selection strategies, particularly bio-inspired algorithms, often yield incorrect subsets of features. During machine learning processes, these inaccuracies impede the detection accuracy of DDoS attacks using ICMPv6. Many Optimization Search Algorithms become trapped in local minima and fail to consider multi-objective approaches, resulting in suboptimal feature selection. To address this, optimizing a bio-inspired algorithm within an IPv6 network has been proposed. Specifically, hybridizing the MFPA algorithm with the PSO algorithm is suggested to enhance detection accuracy. The selected features are used to train the dataset with a Support Vector Machine (SVM) classifier. The proposed approach is evaluated using the ICMPv6 dataset on various attacks, demonstrating superior classification accuracy of 97.99 %. It also reduced the number of features from 19 to 8, showcasing its efficiency.

Global Structure Preservation and Self-Representation-Based Supervised Feature Selection

Feature selection aims to select a subset of features from high-dimensional data, which can overcome the curse of dimensionality for the next dealing steps. However, the feature selection itself could face the curse of dimensionality. To overcome the above problem, in this paper, a new feature selection framework is designed according to a human processing in our daily life. In our daily life, to evaluate a candidate's ability to work, the related professional knowledge and the comprehensive ability of a candidate should be both evaluated. Actually, a candidate only with good professional knowledge often hardly solves new problems in the work. Based on the above analysis, in our new designed framework, the features are selected by evaluating its ability of global structure preservation and self-representation, which are respectively similar to the professional knowledge and comprehensive ability in evaluating candidate. As a result, the selected features can accommodate larger changes in test data. The conducted experiments validate the effectiveness of our feature selection.

An efficient ranking based binary salp swarm optimization for feature selection in high dimensional datasets

Feature selection is a major challenge in data mining which involves complex searching procedure to acquire relevant feature subset. The effectiveness of classification approaches is greatly susceptible to data dimensionality. The Higher dimensionality intricate numerous problems like higher computational costs and over fitting problem. The essential key factor to mitigate the problem is feature selection. The main motive is to minimize the number of features through eliminating noisy, insignificant, and redundant features from the original data. The Metaheuristic algorithm attains excellent performance for solving this kind of problems. In this paper, the grading based binary salp swarm optimization has been proposed to solve various complex problems with lesser computational time. The grading system has been used to maintain the balance among exploitation and exploration. The proposed method is examined using ten benchmark real datasets. The comparative result exhibits the promising performance of our proposed method and surpasses with other optimization interms of investigating evaluation measures.

Financial distress prediction using an improved particle swarm optimization wrapper feature selection method and tree boosting ensemble

Financial distress prediction (FDP) is a complex task involving both feature selection and model construction. While many studies have addressed these challenges individually, there is a lack of FDP models that integrate feature selection into the overall model building process. To address the issues of fast convergence and local optimization in particle swarm optimization (PSO), this study proposes a modified PSO algorithm with three improvement strategies. This improved PSO, known as IPSO, is embedded within state-of-the-art tree boosting ensemble models for feature selection. IPSO continuously evolves toward the optimal feature subset, preventing premature convergence and avoiding local optima. To construct the FDP model, four state-of-the-art tree boosting models are combined using an improved majority voting ensemble strategy. The integration of IPSO enhances the prediction performance of the model. Comparative experiments are conducted on samples of Chinese listed companies under short-term and long-term prediction scenarios to evaluate the proposed model. The results demonstrate that IPSO significantly improves the prediction performance, and the ensemble strategy enhances the robustness of the predictions.

Memetic multilabel feature selection using pruned refinement process

With the growing complexity of data structures, which include high-dimensional and multilabel datasets, the significance of feature selection has become more emphasized. Multilabel feature selection endeavors to identify a subset of features that concurrently exhibit relevance across multiple labels. Owing to the impracticality of performing exhaustive searches to obtain the optimal feature subset, conventional approaches in multilabel feature selection often resort to a heuristic search process. In this context, memetic multilabel feature selection has received considerable attention because of its superior search capability; the fitness of the feature subset created by the stochastic search is further enhanced through a refinement process predicated on the employed multilabel feature filter. Thus, it is imperative to employ an effective refinement process that frequently succeeds in improving the target feature subset to maximize the benefits of hybridization. However, the refinement process in conventional memetic multilabel feature selection often overlooks potential biases in feature scores and compatibility issues between the multilabel feature filter and the subsequent learner. Consequently, conventional methods may not effectively identify the optimal feature subset in complex multilabel datasets. In this study, we propose a new memetic multilabel feature selection method that addresses these limitations by incorporating the pruning of features and labels into the refinement process. The effectiveness of the proposed method was demonstrated through experiments on 14 multilabel datasets.

Utilizing radiomics and dosiomics with AI for precision prediction of radiation dermatitis in breast cancer patients

PurposeThis study explores integrating clinical features with radiomic and dosiomic characteristics into AI models to enhance the prediction accuracy of radiation dermatitis (RD) in breast cancer patients undergoing volumetric modulated arc therapy (VMAT).Materials and methodsThis study involved a retrospective analysis of 120 breast cancer patients treated with VMAT at Kaohsiung Veterans General Hospital from 2018 to 2023. Patient data included CT images, radiation doses, Dose-Volume Histogram (DVH) data, and clinical information. Using a Treatment Planning System (TPS), we segmented CT images into Regions of Interest (ROIs) to extract radiomic and dosiomic features, focusing on intensity, shape, texture, and dose distribution characteristics. Features significantly associated with the development of RD were identified using ANOVA and LASSO regression (p-value < 0.05). These features were then employed to train and evaluate Logistic Regression (LR) and Random Forest (RF) models, using tenfold cross-validation to ensure robust assessment of model efficacy.ResultsIn this study, 102 out of 120 VMAT-treated breast cancer patients were included in the detailed analysis. Thirty-two percent of these patients developed Grade 2+ RD. Age and BMI were identified as significant clinical predictors. Through feature selection, we narrowed down the vast pool of radiomic and dosiomic data to 689 features, distributed across 10 feature subsets for model construction. In the LR model, the J subset, comprising DVH, Radiomics, and Dosiomics features, demonstrated the highest predictive performance with an AUC of 0.82. The RF model showed that subset I, which includes clinical, radiomic, and dosiomic features, achieved the best predictive accuracy with an AUC of 0.83. These results emphasize that integrating radiomic and dosiomic features significantly enhances the prediction of Grade 2+ RD.ConclusionIntegrating clinical, radiomic, and dosiomic characteristics into AI models significantly improves the prediction of Grade 2+ RD risk in breast cancer patients post-VMAT. The RF model analysis demonstrates that a comprehensive feature set maximizes predictive efficacy, marking a promising step towards utilizing AI in radiation therapy risk assessment and enhancing patient care outcomes.

An improved Genghis Khan optimizer based on enhanced solution quality strategy for global optimization and feature selection problems

Feature selection (FS) is the activity of defining the most contributing feature subset among all used features to improve the superiority of datasets with a large number of dimensions by selecting significant features and eliminating redundant and irrelevant ones. Therefore, this process can be seen as an optimization process. The primary goals of feature selection are to decrease the number of dimensions and enhance classification accuracy in many domains, such as text classification, large-scale data analysis, and pattern recognition. Several metaheuristics, such as the Genghis Khan Shark Optimizer Algorithm (GKSO), can assist in optimizing the FS issue. However, these methods tend to converge towards local solutions with a low convergence rate. In order to address this issue in GKSO, a more refined version called I-GKSO is implemented. The I-GKSO suggested introducing a new approach to modify solutions that have low fitness values. In addition, it employs the Enhanced Solution Quality (ESQ) strategy to enhance the exploration phase. It utilizes Quasi-opposite-based learning (QOBL) to enhance the best solution obtained and, consequently, the entire population. The algorithm presented aims to solve the FS problem and has been assessed using benchmark optimization problems from the CEC’2017 and CEC’2022. To assess the efficacy of the I-GKSO, it has been subjected to comparisons with multiple different algorithms. The trials conducted using FS datasets yield a quantitative consideration of the I-GKSO's capacity to attain the most optimal subset of features. Furthermore, Wilcoxon and Friedman's non-parametric tests were accomplished to support the performance of the proposed method.

Surface roughness analysis of side-polished fiber based on the importance of texture features

In order to explore the use of side-polished fibre (SPF) for microprobe-type "lab-on-fibre", this study presents an analysis of the surface roughness in side-polished fiber (SPF) using the gray level co-occurrence matrix (GLCM) texture feature analysis method. Experimental results show that the flat areas of the SPF polished surface exhibit texture characteristics with high mean values in contrast and entropy, and low mean values in the angular second moment (ASM), homogeneity, and correlation. Employing the random forest (RF) feature importance ranking method based on the Gini coefficient and out-of-bag (OOB) error estimation, this study assesses the sensitivity of various GLCM texture parameters in classifying different roughness levels of the SPF polished surfaces. A feature subset comprising variance, ASM, entropy, and contrast is identified as optimal. Utilizing this subset, the paper conducts an RF classification validation experiment on the roughness of the SPF polished surfaces, with results showing an RF classification accuracy of 95.65%.This research provides evidence for exploring the impact of rough polished surfaces in SPF optic sensors on the light coupling mechanism with environmental materials and its influence on sensor sensitivity. It lays the foundation for exploring the precise identification of high-sensitivity areas on SPF polished surfaces.

Optimizing Facial Expression Recognition with Biogeography-Based Feature Selection

Facial expression recognition is a challenging research field in computer vision due to various issues such as occlusion, lighting conditions, camera pose angles, and the selection of relevant features. Extracting and selecting pertinent features from facial images is crucial in achieving efficient expression recognition. This paper proposes a metaheuristic-based feature selection and classification methodology using the Biogeography-Based Optimization (BBO) algorithm to select the best-performing features and optimize the recognition accuracy of the classifier. The cross-validation recognition accuracy of the Support Vector Machine (SVM) is used as the evaluation criterion in the BBO algorithm to choose the optimal feature subset from the extracted features. The performance of the proposed BBO-SVM feature selection model is compared with other filter-based approaches. Experiments are conducted on three publicly available databases: JAFFE, MUG, and CK+, to validate the performance of the proposed system. The model achieves promising recognition accuracy across all datasets, with results compared to similar works presented in the literature.

Adaptive Evolutionary Computing Ensemble Learning Model for Sentiment Analysis

Standard machine learning and deep learning architectures have been widely used in the field of sentiment analysis, but their performance is unsatisfactory if the input texts are short (e.g., social media posts). Specifically, the accuracy of standard machine learning methods crucially depends on the richness and completeness of the features used to represent the texts, and in the case of short messages, it is often difficult to obtain high-quality features. Conversely, methods based on deep learning can achieve better expressiveness, but these methods are computationally demanding and often suffer from over-fitting. This paper proposes a new adaptive evolutionary computational integrated learning model (AdaECELM) to overcome the problems encountered by traditional machine learning and deep learning models in sentiment analysis for short texts. AdaECELM consists of three phases: feature selection, sub classifier training, and global integration learning. First, a grid search is used for feature extraction and selection of term frequency-inverse document frequency (TF-IDF). Second, cuckoo search (CS) is introduced to optimize the combined hyperparameters in the sub-classifier support vector machine (SVM). Finally, the training set is divided into different feature subsets for sub-classifier training, and then the trained sub-classifiers are integrated and learned using the AdaBoost integrated soft voting method. Extensive experiments were conducted on six real polar sentiment analysis data sets. The results show that the AdaECELM model outperforms the traditional ML comparison methods according to evaluation metrics such as accuracy, precision, recall, and F1-score in all cases, and we report an improvement in accuracy exceeding 4.5%, the second-best competitor.

Solving Multiobjective Feature Selection Problems in Classification via Problem Reformulation and Duplication Handling

Reducing the number of selected features and improving the classification performance are two major objectives in feature selection, which can be viewed as a multi-objective optimization problem. Multi-objective feature selection in classification has its unique characteristics, such as it has a strong preference for the classification performance over the number of selected features. Besides, solution duplication often appears in both the search and the objective spaces, which degenerates the diversity and results in the premature convergence of the population. To deal with the above issues, in this paper, during the evolutionary training process, a multi-objective feature selection problem is reformulated and solved as a constrained multi-objective optimization problem, which adds a constraint on the classification performance for each solution (e.g., feature subset) according to the distribution of nondominated solutions, with the aim of selecting promising feature subsets that contain more informative and strongly relevant features, which are beneficial to improve the classification performance. Furthermore, based on the distribution of feature subsets in the objective space and their similarity in the search space, a duplication analysis and handling method is proposed to enhance the diversity of the population. Experimental results demonstrate that the proposed method outperforms six state-of-the-art algorithms and is computationally efficient on 18 classification datasets.

An efficient intrusion detection systems in fog computing using forward selection and BiLSTM

Intrusion detection systems (IDS) play a pivotal role in network security and anomaly detection and are significantly impacted by the feature selection (FS) process. As a significant task in machine learning and data analysis, FS is directed toward pinpointing a subset of pertinent features that primarily influence the target variable. This paper proposes an innovative approach to FS, leveraging the forward selection search algorithm with hybrid objective/fitness functions such as correlation, entropy, and variance. The approach is evaluated using the BoT-IoT and TON_IoT datasets. By employing the proposed methodology, our bidirectional long-short term memory (BiLSTM) model achieved an accuracy of 98.42% on the TON_IoT dataset and 98.7% on the BoT-IoT dataset. This superior classification accuracy underscores the efficacy of the synergized BiLSTM deep learning model and the innovative FS approach. The study accentuates the potency of the proposed hybrid approach in FS for IDS and highlights its substantial contribution to achieving high classification performance in internet of things (IoT) network traffic analysis.

Learning Accurate Label-Specific Features From Partially Multilabeled Data.

Feature selection is an effective dimensionality reduction technique, which can speed up an algorithm and improve model performance such as predictive accuracy and result comprehensibility. The study of selecting label-specific features for each class label has attracted considerable attention since each class label might be determined by some inherent characteristics, where precise label information is required to guide label-specific feature selection. However, obtaining noise-free labels is quite difficult and impractical. In reality, each instance is often annotated by a candidate label set that comprises multiple ground-truth labels and other false-positive labels, termed partial multilabel (PML) learning scenario. Here, false-positive labels concealed in a candidate label set might induce the selection of false label-specific features while masking the intrinsic label correlations, which misleads the selection of relevant features and compromises the selection performance. To address this issue, a novel two-stage partial multilabel feature selection (PMLFS) approach is proposed, which elicits credible labels to guide accurate label-specific feature selection. First, the label confidence matrix is learned to help elicit ground-truth labels from the candidate label set via the label structure reconstruction strategy, each element of which indicates how likely a class label is ground truth. After that, based on distilled credible labels, a joint selection model, including label-specific feature learner and common feature learner, is designed to learn accurate label-specific features to each class label and common features for all class labels. Besides, label correlations are fused into the features selection process to facilitate the generation of an optimal feature subset. Extensive experimental results clearly validate the superiority of the proposed approach.