Researches and Comparisons of the Methods that Simplify Plenty of Attributes in Machine Learning
Decisions are often made under limited information in the real world. In our previous study, a method named mega-fuzzification that is using continuous data and domain external expansion
methods was proposed to solve the so called small data set learning problem in FMS. However, when the number of input attributes is too large, it is very hard to perform. This paper presents a novel method to solve the problem that knowledge has plenty of attributes in learning. Large number of input attributes not only causes hard computation but also breaks software limits. In addition, the proposed method will compared with other machine learning.
Keywords: small data set learning, mega-fuzzification, fuzzy neural network, machine learning