Attribute Fusion-based Classifier on Framework of Belief Structure
Qiying Hu, Yingying Liang, Qianli Zhou, Witold Pedrycz
TL;DR
This paper tackles uncertainty-aware multi-attribute classification within DST by identifying two main weaknesses of prior attribute-fusion classifiers: overly simple membership modeling and limited use of BPA belief structure. It proposes two key innovations: (i) selective membership modeling employing both single-Gaussian and Gaussian Mixture Models (GMMs) with cross-validated model selection, and (ii) a novel BPA-generation method that transforms normalized possibility distributions into richer BPAs, enabling broader focal-element support. The method is integrated into both the primary attribute-fusion classifier and an enhanced evidential KNN (BEKNN) by applying the BPA-generation approach to neighbor-based evidence fusion, yielding improved accuracy and robustness across eight UCI datasets and few-shot vision benchmarks. Results show consistent performance gains over state-of-the-art evidential classifiers and competitive results against conventional machine learning methods, with significant statistical support. This work advances uncertainty-aware decision making in classification by coupling flexible membership representations with belief-structure–enabled BPA fusion.
Abstract
Dempster-Shafer Theory (DST) provides a powerful framework for modeling uncertainty and has been widely applied to multi-attribute classification tasks. However, traditional DST-based attribute fusion-based classifiers suffer from oversimplified membership function modeling and limited exploitation of the belief structure brought by basic probability assignment (BPA), reducing their effectiveness in complex real-world scenarios. This paper presents an enhanced attribute fusion-based classifier that addresses these limitations through two key innovations. First, we adopt a selective modeling strategy that utilizes both single Gaussian and Gaussian Mixture Models (GMMs) for membership function construction, with model selection guided by cross-validation and a tailored evaluation metric. Second, we introduce a novel method to transform the possibility distribution into a BPA by combining simple BPAs derived from normalized possibility distributions, enabling a much richer and more flexible representation of uncertain information. Furthermore, we apply the belief structure-based BPA generation method to the evidential K-Nearest Neighbors (EKNN) classifier, enhancing its ability to incorporate uncertainty information into decision-making. Comprehensive experiments on benchmark datasets are conducted to evaluate the performance of the proposed attribute fusion-based classifier and the enhanced evidential K-Nearest Neighbors classifier in comparison with both evidential classifiers and conventional machine learning classifiers. The results demonstrate that the proposed classifier outperforms the best existing evidential classifier, achieving an average accuracy improvement of 4.86%, while maintaining low variance, thus confirming its superior effectiveness and robustness.