Incorprating Sentiment Analysis Into Recommendation In Social Tagging System

In the social tagging system,the user freely annotate the interest resources through the tags,which constitutes ternary relationship user-tags-resources in social tagging system.Tag as a connection of users and resources,which can describe the user's interest and the content of resources.More full use of social tag is the core and foundation of the reasonable social tagging recommendation,while the existing research most focus on processing the label,and considerate the relationship among users resources and tags on the tag semantics.Moreover the actual situation is that the tag is an explicit expression of emotion by users,because of the user's freely tagging,which leads to problems of tag semantic and complexity.Therefore,combined with the Natural Science Fund Project “Research on social-driven context-aware personalized information service in ubiquitous computing environment”(project NO: 71471165),this paper presents incorprating sentiment analysis into recommendation in social tagging system,which elaborated the tag sentiment analysis and the emotion into the process of the different recommendation algorithm systematically,and then carry out the contrast experiment analysis.The main contributions of this thesis are listed as follows:(1)The tag emotion analysis is proposed.Through the pretreatment of the tag,then extract the concept,and the concept of tags as input,using SenticNet emotional knowledges to get tag emotions vector.Which can change the original semantic tags into five dimensions of emotional vector form,thus overcome the tag semantic and fuzziness.(2)The semantic profile model and sentiment profile model are constructed.From the ternary relation user-tag-resources in the social tagging system,then build semantic label model according to the frequency of the tag usage.At the same time change the ternary relationship into user-sentiment-resources through the tag sentiment analysis,then build setiment profile model,which can realize the affection of user interests and resources content description though changing the tag semantics into tag setiment reasonablely.(3)The method of incorprating sentiment analysis into recommendation in social tagging system are proposed.Combined with tags semantic information and setiment information,this article embarks from the content-based recommendation and collaborative filtering recommendation algorithm,at the same time considering the user preference and emotional semantic preference for resources,thus obtain social tagging of recommended results more in line with the user's interest.(4)A comparative experimental study is carried out from Movielens,using the accuracy and recall rate as well as the comprehensive evaluation index,the analysis of the proposed incorporating sentiment into recommendation in social tagging validity verification in this paper.The results show that the proposed incorporating sentiment into recommendation in social tagging has better performance in content-based recommendation and collaborative filtering recommendation algorithm,and which is superior to pure social tagging recommend using the tag.This research results has a higher practical application value on social tagging recommendation.