篇名 | Fuzzy Rules Reduction Based on Sparse Coding |
---|---|
卷期 | 16:3 |
作者 | Huiqin Jiang 、 Rung-Ching Chen 、 Qiao-En Liu 、 Su-Wen Huang |
頁次 | 215-227 |
關鍵字 | Data-driven FISs 、 lasso 、 LARS 、 FCM 、 sparse enconding 、 Scopus |
出刊日期 | 201911 |
DOI | 10.6703/IJASE.201911_16(3).215 |
With high-dimensional data appearing, the number of fuzzy rules increases which degrade the interpretability and increases the computation complexity of the fuzzy rule-based system. In this paper, we proposed a rule-reduced algorithm. Through the sparse encoding of the fuzzy basis functions (FBFs), rules are reduced. Least angle regression algorithm is proposed here to select the important rules. Compared with other sparse encoding algorithm, Least angle regression algorithm has the advantage of lower computation complexity and better performance. The experimental results show that our proposed algorithm has excellent performance, especially for high-dimension data.