Interactive and Complementary Feature Selection via Fuzzy Multigranularity Uncertainty Measures

Novel Ensemble Approach with Incremental Information Level and Improved Evidence Theory for Attribute Reduction

With the development of intelligent technology, data in practical applications show exponential growth in quantity and scale. Extracting the most distinguished attributes from complex datasets becomes a crucial problem. The existing attribute reduction approaches focus on the correlation between attributes and labels without considering the redundancy. To address the above problem, we propose an ensemble approach based on an incremental information level and improved evidence theory for attribute reduction (IILE). Firstly, the incremental information level reduction measure comprehensively assesses attributes based on reduction capability and redundancy level. Then, an improved evidence theory and approximate reduction methods are employed to fuse multiple reduction results, thereby obtaining an approximately globally optimal and a most representative subset of attributes. Eventually, using different metrics, experimental comparisons are performed on eight datasets to confirm that our proposal achieved better than other methods. The results show that our proposal can obtain more relevant attribute sets by using the incremental information level and improved evidence theory.

Unsupervised attribute reduction based on variable precision weighted neighborhood dependency

Neighborhood rough set (NRS) have been successfully applied to attribute reduction (AR). However, most current methods of AR based on NRS are supervised or semi-supervised. This limits their ability to process data without decision information. When granulating data samples, NRS considers only the number of samples within the neighborhood radius. It does not consider distribution information between samples, which can result in the loss of original data information. To address the aforementioned issue, we propose an unsupervised attribute reduction (UAR) strategy based on variable precision weighted neighborhood dependency (VPWND) (UAR_VPWND). We compare algorithm UAR_VPWND to existing classical UAR algorithms using public datasets. The experimental results show that algorithm UAR_VPWND can select fewer attributes to maintain or improve the performance of clustering learning algorithms.

Multigranularity Data Analysis With Zentropy Uncertainty Measure for Efficient and Robust Feature Selection

Multigranularity data analysis has recently become an active research topic in the intelligent computing and data mining fields. Feature selection via multigranularity data analysis is an effective tool for characterizing hierarchical data and enhancing the accuracy of the results. Although the multigranularity data analysis method has been widely adopted for feature selection, existing studies still present one prevalent disadvantage: multigranularity data analysis mostly focuses on information presented at a single granularity while ignoring the hierarchical structure of multigranularity data, which is contrary to the nature of multigranularity. Hence, this article proposes a multigranularity data analysis with a zentropy uncertainty measure for efficient and robust feature selection. Specifically, a consistent degree is first introduced to obtain optimal granularity combinations and establish an efficient neighborhood model for multigranularity information processing. Then, a novel and robust uncertainty measure is developed by integrating the multigranularity information, namely the zentropy-based measure. Considering its accuracy among uncertainty measures, two important measures are further designed and applied to feature selection. Extensive experiments demonstrate that the proposed method can achieve better robustness and classification performance than other state-of-the-art methods.

Selection of healthcare waste management treatment using fuzzy rough numbers and Aczel-Alsina Function

The COVID-19 pandemic led to an increase in healthcare waste (HCW). HCW management treatment needs to be re-taken into focus to deal with this challenge. In practice, there are several treatments of HCW with their advantages and disadvantages. This study is conducted to select the appropriate treatment for HCW in the Brčko District of Bosnia and Herzegovina. Six HCW management treatments are analyzed and observed through twelve criteria. Ten-level linguistic values were used to bring this evaluation closer to human thinking. A fuzzy rough approach is used to solve the problem of inaccuracy in determining these values. The OPA method from the Bonferroni operator is used to determine the weights of the criteria. The results of the application of this method showed that the criterion Environmental Impact ( C 4 ) received the highest weight, while the criterion Automation Level ( C 8 ) received the lowest value. The ranking of HCW management treatments was performed using MARCOS methods based on the Aczel-Alsina function. The results of this analysis showed that the best-ranked HCW management treatment is microwave (A6) while landfill treatment (A5) is ranked worst. This study has provided a new approach based on fuzzy rough numbers where the Bonferroni function is used to determine the lower and upper limits, while the application of the Aczel-Alsina function reduced the influence of decision-makers on the final decision because this function stabilizes the decision-making process.

A novel approach based on rough set theory for analyzing information disorder

The paper presents and evaluates an approach based on Rough Set Theory, and some variants and extensions of this theory, to analyze phenomena related to Information Disorder. The main concepts and constructs of Rough Set Theory, such as lower and upper approximations of a target set, indiscernibility and neighborhood binary relations, are used to model and reason on groups of social media users and sets of information that circulate in the social media. Information theoretic measures, such as roughness and entropy, are used to evaluate two concepts, Complexity and Milestone, that have been borrowed by system theory and contextualized for Information Disorder. The novelty of the results presented in this paper relates to the adoption of Rough Set Theory constructs and operators in this new and unexplored field of investigation and, specifically, to model key elements of Information Disorder, such as the message and the interpreters, and reason on the evolutionary dynamics of these elements. The added value of using these measures is an increase in the ability to interpret the effects of Information Disorder, due to the circulation of news, as the ratio between the cardinality of lower and upper approximations of a Rough Set, cardinality variations of parts, increase in their fragmentation or cohesion. Such improved interpretative ability can be beneficial to social media analysts and providers. Four algorithms based on Rough Set Theory and some variants or extensions are used to evaluate the results in a case study built with real data used to contrast disinformation for COVID-19. The achieved results allow to understand the superiority of the approaches based on Fuzzy Rough Sets for the interpretation of our phenomenon.

Boosting Whale Optimizer with Quasi-Oppositional Learning and Gaussian Barebone for Feature Selection and COVID-19 Image Segmentation

Whale optimization algorithm (WOA) tends to fall into the local optimum and fails to converge quickly in solving complex problems. To address the shortcomings, an improved WOA (QGBWOA) is proposed in this work. First, quasi-opposition-based learning is introduced to enhance the ability of WOA to search for optimal solutions. Second, a Gaussian barebone mechanism is embedded to promote diversity and expand the scope of the solution space in WOA. To verify the advantages of QGBWOA, comparison experiments between QGBWOA and its comparison peers were carried out on CEC 2014 with dimensions 10, 30, 50, and 100 and on CEC 2020 test with dimension 30. Furthermore, the performance results were tested using Wilcoxon signed-rank (WS), Friedman test, and post hoc statistical tests for statistical analysis. Convergence accuracy and speed are remarkably improved, as shown by experimental results. Finally, feature selection and multi-threshold image segmentation applications are demonstrated to validate the ability of QGBWOA to solve complex real-world problems. QGBWOA proves its superiority over compared algorithms in feature selection and multi-threshold image segmentation by performing several evaluation metrics.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s42235-022-00297-8.

Online group streaming feature selection using entropy-based uncertainty measures for fuzzy neighborhood rough sets

Online group streaming feature selection, as an essential online processing method, can deal with dynamic feature selection tasks by considering the original group structure information of the features. Due to the fuzziness and uncertainty of the feature stream, some existing methods are unstable and yield low predictive accuracy. To address these issues, this paper presents a novel online group streaming feature selection method (FNE-OGSFS) using fuzzy neighborhood entropy-based uncertainty measures. First, a separability measure integrating the dependency degree with the coincidence degree is proposed and introduced into the fuzzy neighborhood rough sets model to define a new fuzzy neighborhood entropy. Second, inspired by both algebra and information views, some fuzzy neighborhood entropy-based uncertainty measures are investigated and some properties are derived. Furthermore, the optimal features in the group are selected to flow into the feature space according to the significance of features, and the features with interactions are left. Then, all selected features are re-evaluated by the Lasso model to discard the redundant features. Finally, an online group streaming feature selection algorithm is designed. Experimental results compared with eight representative methods on thirteen datasets show that FNE-OGSFS can achieve better comprehensive performance.