Logo image
Deep Reconciled & Self-Paced TSK Fuzzy System Ensemble for Imbalanced Data Classification: Architecture, Interpretability and Theory
Journal article   Peer reviewed

Deep Reconciled & Self-Paced TSK Fuzzy System Ensemble for Imbalanced Data Classification: Architecture, Interpretability and Theory

Yuanpeng Zhang, Guanjin Wang, Ta Zhou, Ge Ren, Saikit Lam, Weiping Ding and Jing Cai
IEEE transactions on fuzzy systems, Early Access
2024

Abstract

ensemble learning Fuzzy systems interpretability Nearest neighbor methods Radio frequency reconciled & self-paced sampling Sensitivity Stacking Task analysis testing-compatible sample sensitivity Training TSK fuzzy system
Stacking-based TSK fuzzy system ensemble has been successfully applied in imbalanced data classification. However, there still exists many challenges need to be further addressed. For example, during stacking, augmenting output variables into the input feature space reduces interpretability of antecedents of fuzzy rules. During sampling for balancing, discovering informative samples usually only replies on training samples which may reduce generalizability. More importantly, there is no theory to support reliability of stacking. To address aforementioned challenges, in this study, we propose a deep reconciled & self-paced TSK fuzzy system ensemble framework termed as D-RSP-TSKE for imbalanced data classification. Compared with existing ensemble frameworks, its superiorities can be exhibited from the following three aspects. (i) In the first layer, we use random under-sampling (RUS) to generate a class-balanced training set to train an initial zero-order TSK fuzzy classifier. Based on the TSK fuzzy classifier, then we define classifier-specific & testing-compatible sample sensitivity to discover informative (high-sensitive) samples and design a reconciled & self-paced sampling approach to balance the minority class for the training of following layers. (ii) To improve interpretability of antecedents of fuzzy rules, we propose to transfer the output variables from antecedents to consequents through equivalent mathematical transformations, while keeping the final output unchanged. These transferred output variables are interpreted as the dynamic fuzzy rule confidence. (iii) Furthermore, we engage in a comprehensive theoretical examination of our stacking-based ensemble to elucidate the underlying mechanisms that enable the stacking strategy to consistently deliver superior performance. We conduct tests and comparisons on 7 artificial datasets and 30 real-world datasets to evaluate D-RSP-TSKE. The experimental results demonstrate the effectiveness and interpretability of D-RSP-TSKE for imbalanced data classification.

Details

Metrics

25 Record Views
Logo image