Abstract: Customer portrait is a hot topic in commercial banks. One core issue of customer portrait is to select effective attributes from high-dimensional and complex customer data. The high-dimensional data makes it challenging to conduct customer portrait accurately. Based on the clustering analysis of customer data, we combine the rough set theory and information entropy to study the attribute reduction of customer portraits in commercial banks, and propose the attribute reduction algorithm. The experimental results show that the proposed method can obtain the optimal attribute reduction of customer portrait and can provide a basis for attribute selection.

Key words: customer portrait, cluster, information entropy, attribute reduction