Abstract: With the rapid development of e-commence under the network environment,product review has become an important data source for enterprises to improve quality and enhance service.The review comprises user’s emotional tendency in all aspects of the product.Emotional analysis can not only help business to understand the advantages and disadvantages of the product,but also provide data support for the potential consumer’s purchase decision.This paper presented a novel method to cluster commodity attribute based on combination neural network and computd sentiment of internet commodities review using word2vec.This essay computed the semantic similarity and built emotional dictionary based on word2vec,then used the emotional dictionary to obtain the emotional tendencies of the test texts.The effectiveness and accuracy of the method is validated through experiments.

Key words: Word2vec,Emotional tendency,Emotional directory,Emotional classification