作者: Mukesh Prasad , Chin-Teng Lin , Dong-Lin Li , Chao-Tien Hong , Wei-Ping Ding
DOI: 10.1109/TSMC.2015.2507139
关键词:
摘要: This correspondence paper proposes an improved version of the self-constructing neural fuzzy inference network (SONFIN), called soft-boosted SONFIN (SB-SONFIN). The design softly boosts learning process in order to decrease error rate and enhance speed. SB-SONFIN power by taking into account numbers rules initial weights which are two important parameters SONFIN, advances by: 1) initializing with width sets rather than just random values 2) improving parameter rates number learned rules. effectiveness proposed soft boosting scheme is validated on several real world benchmark datasets. experimental results show that possesses capability outperform other known methods various