A group method of symmetric triangular fuzzy numbers handling and its application to analyze monitoring data of small and micro enterprises in Sichuan Province

Abstract

Group method of data handling has been proven an effective knowledge mining tool to emerge the influencing factors of enterprises growth in the past researches, but when employed to analyze the small and micro enterprises in Sichuan Province, its effectiveness is reduced by the noise in the data obtained from the enterprises' produce and business operations monitor platform website. In order to increase the noise immunity of the method, the monitoring data is firstly processed as some symmetric triangular fuzzy numbers in this paper, and the parameter estimation technique for all self-organized models named partial functions in the method is changed from the former regression analysis with determined training data set to the fuzzy programming with fuzzy set. Basing on this transform, a group method of the symmetric triangular fuzzy numbers handling is presented and the result of its empirical study in Sichuan Province proves the method is able to disclose some key influencing factors of the small and micro enterprises' growth with unstable and noisy monitoring data.