Dissertation > Excellent graduate degree dissertation topics show

Research on the key Issues in Personalized Recommendation Based on SNS

Author: SunZuo
Tutor: YangChangChun
School: Changzhou University
Course: Applied Computer Technology
Keywords: SNS Personalized recommendation Collaborative filtering Random walks Authoritative users Muti-interest trustlevel recommendation
CLC: TP391.3
Type: Master's thesis
Year: 2011
Downloads: 79
Quote: 0
Read: Download Dissertation

Abstract


Since entering21st century, the websites have been socializedobviously. More and more websites integrate with shopping，makingfriends，chatting，forum and bolg, which could provide users with all-roundservice. As a way of personalized service, personalized recommendationsystem can recommend items which users interested and can help usersmake decisions. This system significantly reduces the cost of people gettingthe information.At present, the actual recommendation systems mostly adoptcollaborative filtering recommendation algorithms. The collaborativefiltering recommendation System exists the problems of data sparseness,cold starting, low expandability and so on. In the environment of socialnetworking, some researchers brought the social trust mechanism into thetraditional collaborative filtering recommendation System, so that theserecommendation Systems based on trust improved the above problemspartly.However, the trust between users can not be accurately described. Inmany cases, the useful effect of user trust in the recommendation is evenworse than the traditional methods used by the user similarity. Someresearchers have proposed the method of combining traditional trust-basedcollaborative filtering and trust-based recommendation, but these studiesneglect the effect of users groups characteristics on the recommendation.Focusing on how to improve the recommendation satisfaction of usersabout recommendation system, this paper explores and researches some keyissues of the personalized recommendation system in the environment ofsocial networking.Firstly, aiming at the shortage of present recommendation algorithms,this paper presents a new social recommendation model-Cliqueswalk using the analytical approach of social networking. Experiment resultsdemonstrate that the new method can greatly narrowed the scope ofsearching target rating information and the effect of recommendation isbetter than the existing methods of Collaborative filtering, trust-based andtheir combined.Secondly, taking into account of the different effect of different usersin social networking, this paper adopts the analytical approach of socialnetworking and Web Ming method, designs the search algorithm ofauthoritative users(opinion leader) and provides the correspondingrecommendation algorithm. Separating the authoritative users from theordinary users is not only improves the effectiveness of the recommendation,but also brought a better user experience.Finally, as a service in social networking service, blog system holds animportant position. This paper provides the user multiple interests of thetrust-based personalized recommendation algorithm. The algorithm dividesthe whole trust score into multi-trust scoring so that the trust relationshipcan better roundly measure the users’ interest, and then bring the combinedapproach of Collaborative filtering recommendation and trust-basedrecommendation into multi-interest recommendation. The experimentalresults show that the new algorithm not only has better recommendationeffect for the ordinary users, but also improves the recommendation effectfor the cold start users.

Related Dissertations

  1. Establishment and Update of Similar Users’ Cluster in Personalized Information Retrieval,TP391.3
  2. Feature based on complex network propagation characteristics of SNS social networking sites,G206
  3. Upload impact on interpersonal social networking sites,C912.1
  4. Web Usage Mining and the Research of Personalized Recommendation,TP311.13
  5. Research on the Application of Social Networking Services in Primary and Secondary School Teachers’ Informal Learning,G635.1
  6. Design and Implementation based the Google platform promotional modules commodity module,TP311.52
  7. The Research and Analysis of Recommendation System Based on Petri-net,TP391.3
  8. The Reaserch and Reliazation of Mblog Based on ANDROID Platform,TN929.5
  9. Microblogging Recommended system based on cloud computing,TP393.092
  10. Structural similarity and its application in recommendation system,TP391.3
  11. Collaborative Filtering Method Using in Recommender System of E-commerce,TP391.3
  12. Study on Brand Communication Strategy Based on Leisure SNS in China,F273.2
  13. Research on Collaborative Recommendation System Based on User Clustering,TP391.3
  14. Research on the Application of Data Mining in the Personalized Service of College Libraries,TP311.13
  15. Research and Implementation of Employment Information Management System,TP315
  16. Collaborative Filtering Recommendation Algorithm Based on Users’ Feature and Preference,TP301.6
  17. Web log mining and collaborative filtering algorithm,TP393.092
  18. Desing and Implementation of Family Social-Networking-Services Web Site Integrateed Sina Micro-Blog,TP393.092
  19. Research on the Web Spokesman of Government under SNS Prospect,G206
  20. The Research of SNS Website’s Development Situation, Problems and Optimized Strategy in China,G206
  21. Research and Implementation for Collaborative Filtering Optimization Algorithm,TP301.6

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Retrieval machine
© 2012 www.DissertationTopic.Net  Mobile