Service Recommendation Based On User's Information In Social Networking

With the rapid development of the Internet,social networks have attracted much attention.There is a lot of data in social networks,where users are basic elements.Social networking is used to accomplish many purposes,all of which serve the users.User-created text messages and social relationships are crucial in social networks,the text message contains text published by user and text of Conversation,but texts of conversation usually also contain text-published during conversation.itj,make full use of user information to achieve personalized recommendations,has become a major challenge in current research.This article aims to use the users'information to recommend servers in socia.l networks,The main achievements of research works in this thesis are summarized as follows:(1)In text,we can get two types of information that users' interest and users'relationship.In order to get the interest and relationship of users,this paper proposes a topic detection and tracking model based on the conversation content.This paper treats the entire conversation as a large topic graph,through extracting and cluster-ing topic,we track users'interests and relationships.The model largely considers the features of conversation texts and improves the utilization of information.(2)In order to achieve personalized recommendation,this paper constructs the structure of user-user level and user-service level respectively.In social network,people think the services,recommended by others who are related to you,are more acceptable.And we consider user-service level to recommend services which users needs.The method makes full use of user information to achieve more accurate and effective service recommendation.