Complex-Valued Neuro-Fuzzy Inference System Based Classifier

被引:0
作者
Subramanian, Kartick [1 ]
Savitha, Ramaswamy [1 ]
Suresh, Sundaram [1 ]
Mahanand, B. S. [2 ]
机构
[1] Nanyang Technol Univ, Sch Comp Engn, Singapore 639798, Singapore
[2] Sri Jayachamarajendra Coll Engn, Dept Informat Sci & Engn, Mysore, Karnataka, India
来源
SWARM, EVOLUTIONARY, AND MEMETIC COMPUTING, (SEMCCO 2012) | 2012年 / 7677卷
关键词
Complex-valued neuro-fuzzy system; wirtinger calculus; classification; SEQUENTIAL LEARNING ALGORITHM; NETWORK; PREDICTION;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In this paper, we propose a complex-valued Takagi-Sugeno-Kang type-0 neuro-fuzzy inference system (CNFIS) and develop for it, a gradient-descent based learning algorithm to solve classification problems. The gradient-descent based learning algorithm is derived based on Wirtinger calculus: which preserves the amplitude-phase correlation. The performance of the developed algorithm is evaluated on a set of four binary classification problems and three multi-category classification problems. Comparison with various real-valued and complex-valued classifiers show the improved performance of CNFIS.
引用
收藏
页码:348 / 355
页数:8
相关论文
共 32 条