Internet brand user management method and system based on big data

The invention discloses an internet brand user management method and system based on big data. The method comprises the steps of data collection, pre-definition, calculation of similarity between users, personal preference prediction and generation of an optimal scheme. The invention belongs to the technical field of computers, and particularly relates to an internet brand user management method and system based on big data, and the method comprises the steps: combining SVNNs of different opinion words in a convex combination manner, and comprehensively considering the weights of the different opinion words; user demands and preferences are comprehensively understood by adopting a comprehensive sentiment analysis method, and similarity among users is obtained based on user evaluation behavior similarity and cloud similarity, so that personalized recommendation and more diversified recommendation content are provided; multi-party negotiation is adopted, individual preferences are comprehensively considered, an