Multi-codebook Fuzzy Neural Network Using Incremental Learning for Multimodal Data Classification

Abstract

One of the challenge in classification is classification in multimodal data. This paper proposed multi-codebook fuzzy neural network by using incremental learning for multimodal data classification. There are 2 variations of the proposed method, one uses a static threshold, and the other uses a dynamic threshold. Based on the experiment result, the multi-codebook FNGLVQ using dynamic incremental learning has the highest improvement compared to the original FNGLVQ. It achieves 15.65% margin in synthetic dataset, 5.02 % margin in benchmark dataset, and 11.30% on average all dataset.