Multilayer Evolving Fuzzy Neural Networks With Self-Adaptive Dimensionality Compression for High-Dimensional Data Classification
Xiaowei Gu, Qiang Ni, Qiang Shen
Abstract
High-dimensional data classification is widely considered as a challenging task in machine learning due to the so-called “curse of dimensionality.” In this article, a novel multilayer jointly evolving and compressing fuzzy neural network (MECFNN) is proposed to learn highly compact multilevel latent representations from high-dimensional data. As a metalevel stacking ensemble system, each layer of MECFNN is based on a single jointly evolving and compressing neural fuzzy inference system (ECNFIS) that self-organizes a set of human-interpretable fuzzy rules from input data in a samplewise manner to perform approximate reasoning. ECNFISs associate a unique compressive projection matrix to each individual fuzzy rule to compress the consequent part into a tighter form, removing redundant information while boosting the diversity within the stacking ensemble. The compressive projection matrices of the cascading ECNFISs are self-updating to minimize the prediction errors via error backpropagation together with the consequent parameters, empowering MECFNN to learn more meaningful, discriminative representations from data at multiple levels of abstraction. An adaptive activation control scheme is further introduced in MECFNN to dynamically exclude less activated fuzzy rules, effectively reducing the computational complexity and fostering generalization. Numerical examples on popular high-dimensional classification problems demonstrate the efficacy of MECFNN.