Application of Compositive Clustering Analysis Methods in Data Preprocessing
Bo Zhong · Journalof Southwest China Normal University · 2002
The problem of defective data often arises during the course of mining data. Thus data preprocessingis necessary. On the basis of traditional cluster analysis methods, this paper attempts to present a compositive clustering analysis method which is employed to remedy lost data in data source. The result of calculationson mass data shows that this method is more reasonableand believablecompared with traditional datapreprocessing methods.