Research On Recommendation Method Of Enterprise Word-of-mouth Information Based On BP Neural Network

The popularity of the Internet, our society entered the era of information open and intelligent behavior, represented by sina weibo platform for the Internet gradually become the mainstream of information sharing platform. And there is a rapid growth in the number of internet users for we knew more about the internet. However, the abundance of the interactive information on the internet brings questions like the credibility crisis of web users and the information, information overload and information isolated island. At the same time, individuals and businesses to obtain useful information from network platform is becoming more and more difficult, especially to rely on the development of the enterprise network, network users are no longer naive but hope to take the initiative to search more passive to accept a network platform is recommended to obtain information. In such a social background, from the enterprise perspective, the study of word of mouth information recommendation, in improving the quality of products, improve enterprise's reputation has important research significance and application prospects.Or recommend against traditional recommendation algorithm of the model in the aspect of information updated in real time and special needs of enterprises to recommend the limitations of this paper proposes a recommendation method based on BP neural network for word-of-mouth information model.Starting from the quality of Internet users, the model based on the LDA (Latent Dirichlet Allocation) theme high-quality user model mining, excavated the user refactoring, optimization, "small world" through the network, and optimize the network using the improved K-means algorithm for clustering process, divided into high-quality user community as the hidden layer into the BP-Neural-Network to build model for referrals.Sina weibo data acquisition enterprise to carry on experiment, the results show the method has obvious effects on word of reputatio recommendation.