Rough sets and principal components analysis: A comparative study on customer database attributes selection
Fábio Henrique Pereira · AFRICAN JOURNAL OF BUSINESS MANAGEMENT · 2012
The databases of real world contain a huge volume of information; however, part of these data is not interesting for the knowledge extraction. So, data are preprocessed for reducing the amount of information and selecting more relevant attributes. This paper addresses a contrastive study between rough sets and principal components analysis on customer database attributes selection. The experiments were carried out using the insurance company database to evaluate a k-means clustering. The objective is to investigate the capacity of these techniques to improve the identification of customers segments in databases, what appears as an important tool to increase the effectiveness of business communication. Key words: Rough sets, principal components analysis, attributes selection, customer database.