Research on Feature Optimization Method in Personal Information Mining

Personal information mining can find out the hidden relationship and characteristics of the target people which can be used for active post operation. The original features which are included in the personal information raw data usually have high dimension and redundancy which often drags down data mining efficiency. A feature optimization method is proposed here to resolve the problem. The method with the purpose of data dimensionality deduction is based on effective association of rough set theory with PCA approach. The final classification features are derived through two steps of optimization and deduction operation. Some documents data from education field and financial field are used for the experiment. The experimental results demonstrate that the hybrid feature optimization method is effective in improving classification accuracy.