A Fast Global Center Fuzzy Clustering Method

In terms of the problems that the fuzzy C-means algorithm is sensitive to the initial center, easy to fall into the local optimal solution, and the algorithm iteration speed is slow, a rapid global center fuzzy clustering system model is established according to the global center theory of fuzzy...

Full description

Saved in:
Bibliographic Details
Main Authors: SUN Dong-pu, TAN Jie-qiong
Format: Article
Language:zho
Published: Harbin University of Science and Technology Publications 2019-08-01
Series:Journal of Harbin University of Science and Technology
Subjects:
Online Access:https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=1716
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In terms of the problems that the fuzzy C-means algorithm is sensitive to the initial center, easy to fall into the local optimal solution, and the algorithm iteration speed is slow, a rapid global center fuzzy clustering system model is established according to the global center theory of fuzzy clustering, and the relevant theoretical analysis and algorithm process is given. In the model, the initial centroid is determined by the DKC value scheme, and the self-defined optimization function is proposed based on the AM metric. According to this function, the cluster centers are dynamically added one by one to every stage of algorithm operation until the algorithm converges. Through experimental comparison and verification, the process reduces the influence of random selection of cluster centers on clustering results, and jumps out of local optimal solution, reduces computation, and has higher clustering accuracy and faster convergence speed.
ISSN:1007-2683