Research On Recommendation System Based On Corpus Of Harmony

Harmony,as a global language based on symbols,currently involves 5707 characters.The number of characters and users are expected to grow rapidly and sustainably with the development of the instant messaging tool and the corpus of Harmony.Aside with this expansion in characters,information searching is made much more difficult for users to find exactly what they are interested in.This thesis studies on the way to provide personalized user service using of the characters of the Harmony and the users' social relations.With the focus on Harmony,the thesis analyzes the current situation and achievements of recommendation system,recommendation algorithm and the global village language.A recommendation system based on Harmony(RSoH)is then designed and developed.The main contributions of this thesis are as follows:1.Proposes Harmony-Content-based Recommendation Algorithm(HCB)based on the analysis of Harmony and content-based recommendation.HCB analyzes the content of Harmony,improves the recommendation upon contents,and applies it to the recommendation of input box of RSoH.2.Proposes Comprehensive-Trust-based Recommendation(CTB)through summarizing and improving the related technology of social recommendation algorithm.CTB uses the social relations and applies it to independent recommendation of RSoH.3.Analyzes the information of the instant messaging tool and corpus of Harmony in order to model the data of Harmony.4.Designs and implements RSoH based on HCB and CTB algorithm,and tests the system to assure the availability of RSoH.In conclusion,RSoH provides personalized recommendation services for users and sets up different recommendation algorithms according to users' requirements.The recommendation from the input box provides users with fast and accurate list for expressions,while independent recommendations make recommended list with higher coverage and innovation.RSoH not only accomplishes the proactive recommendation with the explanatory information,but also improved the users' adoption rate of Harmony.