关键词: 缺失属性值, 隶属函数, 模糊规则, 分类

Abstract: Aiming at the problem that the classification accuracy and generalization ability of model were low in missing attribute value dat
a classification algorithm, we proposed a classification algorithm of missing attribute value data based on fuzzy rule, namely “cyclereceive” model. The algorithm did not need an interpolation computation to the missing attribute value data and could directly classify the data set. The simulation experiment of UCI open data sets was carried out. The experiment results show that, compared with other algorithms, the “cyclereceive” model has higher classification accuracy and generalization ability.

Key words: missing attribute value, membership function, fuzzy rule, classification