摘要: The classification of consumable media by mining relevant text for their identifying features is a subjective process. Previous attempts to perform this type feature have generally been limited in scope due having access user data. Many these studies used human domain knowledge evaluate the accuracy extracted using methods. In paper, we mine book review identify nontrivial set similar books. We make comparisons between books looking that share characteristics, ultimately performing clustering on our data set. use same process corresponding characteristics users. Finally, quality methods examining correlation similarity metric, and ratings.