Research on knowledge acquisition optimization of continuous attributes of product based on hybrid classification method

Considering uncertainty and indiscerniblility existed in the process of knowledge acquisition for the continuous attribute in mechanical product, a novel hybrid discrete method based on clustering theory and rough set theory mixed complementary is proposed. Fully considering the internal relationship of condition attribute and decision attribute of rough set, the traditional fuzzy-C means clustering model is improved to get the more appropriate discretized data, which avoids the blindness of classification and man-made subjective factors in a certain extent. Then, rough set theory is used to acquire the rules and forecast the fault diagnosis of the shaft of the large-scale generating equipment. Finally, a case in this paper shows that the method is superior to traditional discrete method and effective in continuous attributes of product information decisions.