An improved LSI dimension reduction algorithm based on vector space modal

There is not a good method to find k value when using LSI to reduce dimension of Vector Space Modal. to solve the problem above, an improved LSI dimension reduction algorithm called Dimension Reduction Parameter Produced algorithm (DRPP algorithm), was proposed. it adopts machine learning to find k value,instead of traditional manual set parameter. the k value can be obtained by using fit methods of statistical learning theory to fit the data. Theory analysis and experimental results show that the DRPP algorithm can enhance efficiency of dimension reduction.