A Novel Unascertained C-Means Clustering with Application

Using the theory and method of unascertained measure, a novel unascertained C-means clustering model and the clustering weight are established. The basic knowledge of the unascertained sets and concept of unascertained clustering was introduced briefly. Then, the unascertained measure was defined and clustering weight were set up. Experimental results show that the presented algorithm performs more robust to noise than the fuzzy C-means clustering (FCM) algorithm do. Furthermore, the results of stock market board analysis using proposed method that indicates the unascertained C-means clustering model provides a quantitative objective and efficient method of stock market board analysis, and hence is suitable to stock market board analysis.