Research on Optimize Technology in Latent Semantic Indexing Based on Semantic Block

Latent semantic indexing (LSI) is an effective method in the way of dimensionality reduction, which has been applied to many text learning mission, such as text categorization, information retrieval. This paper sufficiently analyses influence of text window toward mapping of latent semantic indexing and bring forward a latent semantic analysis method based on the semantic block which strengthen the term co-occurrences´ precise in the text window so as to rationalize mapping of latent semantic indexing. The experimental result shows that this method which is applied to text categorization improves categorization precise in effect.