Incremental Perspective for Feature Selection Based on Fuzzy Rough Sets

Abstract

Feature selection based on fuzzy rough sets is an effective approach to select a compact feature subset that optimally predicts a given decision label. Despite being studied extensively, most existing methods of fuzzy rough set based feature selection are restricted to computing the whole data set in batch, which is often costly or even intractable for large data sets. To improve the time efficiency, we investigate the incremental perspective for fuzzy rough set based feature selection assuming data can be presented in sample subsets one after another. The key challenge for the incremental perspective is how to add and delete features with the subsequent arrival of sample subsets. We tackle this challenge with strategies of adding and deleting features based on the relative discernibility relations that are updated as subsets arrive sequentially. Two incremental algorithms for fuzzy rough set based feature selection are designed based on the strategies. One updates the selected features as each sample subset arrives, and outputs the final feature subset where no sample subset is left. The other updates the relative discernibility relations but only performs feature selection where there is no further subset arriving. Experimental comparisons suggest our incremental algorithms expedite fuzzy rough set based feature selection without compromising performance.

Fuzzy rough sets and fuzzy rough neural networks for feature selection: A review

“…These incremental methods can update knowledge (such as decision rules and approximations used for feature selection) in dynamic data sets, whose elements, features, or feature values will change over time. For example, Yang, Chen, Wang, and Wang (2017) developed a fuzzy rough set based feature selection method by investigating an incremental perspective. In their method, features can be added or deleted according to the updated relative discernibility relations between each feature and the feature set.…”

“…Since feature selection aims to obtain an attribute subset, redundant attributes can be removed from the original attribute set without reducing the discerning capability, so this discernibility matrix can be used to achieve this goal. In fuzzy rough set based feature selection methods (Chen et al, 2020; Chen & Yang, 2013; Chen, Zhang, et al, 2011; Dai et al, 2017; Jensen & Shen, 2007; Jensen & Shen, 2008; Qu et al, 2017; Selvakumar et al, 2019; Tsang et al, 2008; Varma et al, 2016; Yang, Chen, Wang, Tsang, & Zhang, 2017; Yang, Chen, Wang, & Wang, 2017), this matrix is usually customized depending on the selected fuzzy rough set model.…”

“…Jensen and Shen (2008) also proposed a fuzzy extension of the crisp discernibility matrix for feature subset selection. However, such heuristic algorithms may result in an over‐reduct or under‐reduct instead of a proper reduct because of their stop criteria (Chen, Hu, et al, 2011; Yang, Chen, Wang, & Wang, 2017). To find a proper reduct, the formal concepts of attributes reduction with fuzzy rough sets was introduced by Tsang et al (2008).…”

Feature Subset Selection Based on Variable Precision Neighborhood Rough Sets

“…Selecting effective features from original features to reduce the dimension of a dataset is an important way to improve the performance of learning algorithms and is also the key data preprocessing step in pattern recognition. RST is a good feature selection tool [46][47][48], which goal is using rough set positive region to construct classification accuracy. Classical variable precision RST is suitable to feature selection of noisy data.…”

Kernel Neighborhood Rough Sets Model and Its Application

“…Currently, feature selection plays an important role in machine learning and data mining. Neighborhood rough sets and kernel rough sets are widely used in feature selection [23][24][25][26]. We also can deal with the feature selection problem by using the new rough sets model.…”