ebook img

recommendation system for online social network PDF

103 Pages·2006·2.34 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview recommendation system for online social network

Master Thesis Software Engineering Thesis no: MSE-2006:11 July 2006 RECOMMENDATION SYSTEM FOR ONLINE SOCIAL NETWORK Katarzyna Musiał School of Engineering Blekinge Institute of Technology Box 520 SE – 372 25 Ronneby Sweden i This thesis is submitted to the School of Engineering at Blekinge Institute of Technology in partial fulfilment of the requirements for the degree of Master of Science in Software Engineering. The thesis is equivalent to 20 weeks of full time studies. Contact Information Author: Katarzyna Musiał E–mail: ABSTRACT Although there has been much work done in the industry and academia on developing the theory and application of social networks as well as recommender systems, the relation between these research areas is still unclear. An innovative idea, which enables to integrate these areas, and applies recommendation systems to the online social network systems, is proposed in this thesis. Recommendation systems for social networks differ from the typical kinds of recommendation solutions, since they suggest human beings to other ones rather than inanimate goods. Thus, conventional recommendation methods should be enhanced by social features of the networks and their members. This thesis presents the result of the study on the recommendation framework for virtual communities. It also contains an overview of recent approaches to recommendation systems and social networks, as well as description of the online social network systems. Keywords: Social Networks, Online Social Network Systems, Personalized Recommender System iii CONTENTS 1 INTRODUCTION........................................................................................................................... 1 1.1 BACKGROUND AND MOTIVATION................................................................................................ 1 1.2 RESEARCH QUESTIONS AND OBJECTIVES................................................................................... 1 1.3 OUTLINE OF THESIS ..................................................................................................................... 2 2 RECOMMENDER SYSTEMS...................................................................................................... 3 2.1 BACKGROUND ............................................................................................................................... 3 2.2 DEFINITION OF RECOMMENDER SYSTEMS ................................................................................. 4 2.3 GOALS OF RECOMMENDER SYSTEMS ......................................................................................... 5 2.4 CATEGORIES OF RECOMMENDER SYSTEMS ............................................................................... 6 2.4.1 TAXONOMIES OF RECOMMENDER SYSTEMS............................................................................... 6 2.4.2 SIMPLIFIED APPROACHES.......................................................................................................... 10 2.4.3 DEMOGRAPHIC FILTERING ........................................................................................................ 11 2.4.4 COLLABORATIVE FILTERING..................................................................................................... 11 2.4.5 CONTENT–BASED FILTERING.................................................................................................... 13 2.4.6 HYBRID METHOD ...................................................................................................................... 14 3 ONLINE SOCIAL NETWORKS................................................................................................ 15 3.1 BACKGROUND ............................................................................................................................. 15 3.2 SOCIAL NETWORKS .................................................................................................................... 16 3.3 ONLINE SOCIAL NETWORKS...................................................................................................... 20 3.4 FUNCTIONS OF ONLINE SOCIAL NETWORK SYSTEMS ............................................................. 21 3.5 FEATURES OF SOCIAL NETWORKS ............................................................................................ 25 3.6 SOCIAL CAPITAL......................................................................................................................... 26 3.7 SOCIAL CAPITAL IN ONLINE SOCIAL NETWORK SYSTEMS..................................................... 28 3.8 SOCIAL NETWORK ANALYSIS .................................................................................................... 29 4 RECOMMENDER SYSTEM FOR ONLINE SOCIAL NETWORK ..................................... 31 4.1 PROBLEM DESCRIPTION............................................................................................................. 31 4.2 REQUIREMENTS AND PREREQUISITES OF SYSTEM................................................................... 32 4.3 THE CONCEPT OF RECOMMENDATION FOR SOCIAL NETWORKS ........................................... 33 4.4 USER PROFILE............................................................................................................................. 33 4.4.1 DEMOGRAPHIC COMPONENT .................................................................................................... 34 4.4.2 INTEREST COMPONENT ............................................................................................................. 38 4.4.3 PREFERRED COMPONENT .......................................................................................................... 41 4.4.4 SEARCH COMPONENT................................................................................................................ 41 4.4.5 ACTIVITY COMPONENT ............................................................................................................. 43 4.4.6 RELATIONSHIP COMPONENT ..................................................................................................... 44 4.5 RECOMMENDATION PROCESS.................................................................................................... 44 4.6 DATA PREPARATION ................................................................................................................... 46 4.7 PRELIMINARY FILTERING .......................................................................................................... 48 4.8 FINAL SIMILARITY BETWEEN USERS ........................................................................................ 48 4.8.1 DIRECT SIMILARITY .................................................................................................................. 49 4.8.1.1 Straight Comparison of Pair of Attributes ............................................................................. 50 4.8.1.2 Hierarchies ............................................................................................................................. 52 4.8.2 SEARCH MATCHING................................................................................................................... 54 4.8.3 COMPLEMENTARY OF RELATIONSHIP INITIATION .................................................................... 56 4.8.3.1 Assign users to specific class of initiators.............................................................................. 57 iv 4.8.3.2 Assign users to specific class of responders .......................................................................... 60 4.8.4 ACTIVITY OF THE USER............................................................................................................. 61 4.8.5 STRENGTH OF THE RELATIONSHIP ............................................................................................ 63 4.8.5.1 Simple Strength of Relationships........................................................................................... 64 4.8.5.2 Recurrent Strength of Relationships ...................................................................................... 67 4.9 SOCIAL FILTERING ..................................................................................................................... 67 4.10 SUMMARY.................................................................................................................................. 70 5 CHALLENGES............................................................................................................................. 71 5.1 ADJUSTMENT OF WEIGHTS........................................................................................................ 71 5.2 RULES .......................................................................................................................................... 71 6 CASE STUDY ............................................................................................................................... 74 6.1 DATA ABOUT USERS ................................................................................................................... 74 6.2 PRELIMINARY FILTERING.......................................................................................................... 77 6.3 DIRECT SIMILARITY ................................................................................................................... 77 6.4 SEARCH MATCHING .................................................................................................................... 78 6.5 COMPLEMENTARY OF RELATIONSHIP....................................................................................... 78 6.6 ACTIVITY OF THE USER .............................................................................................................. 78 6.7 STRENGTH OF RELATIONSHIP.................................................................................................... 79 6.8 FINAL SIMILARITY BETWEEN USERS ......................................................................................... 80 7 FUTURE WORK.......................................................................................................................... 82 8 CONCLUSIONS ........................................................................................................................... 83 9 REFERENCES.............................................................................................................................. 85 APPENDIX A – ABBREVIATIONS................................................................................................. 89 APPENDIX B – NOTATION............................................................................................................. 90 APPENDIX C – POLISH SUMMARY ............................................................................................. 92 v LIST OF FIGURES FIGURE 2.1 LEVELS OF THE PERSONALIZATION IN THE RECOMMENDER SYSTEMS........................ 4 FIGURE 2.2 THE EXAMPLE OF TAXONOMY OF THE RECOMMENDER SYSTEMS [ADTU05] ............. 7 FIGURE 2.3 THE EXAMPLE OF TAXONOMY OF THE RECOMMENDER SYSTEMS [BUR02] ................ 7 FIGURE 2.4 THE EXAMPLE OF TAXONOMY OF THE RECOMMENDER SYSTEMS [MOLO03]............. 9 FIGURE 2.5 THE EXAMPLE OF TAXONOMY OF THE RECOMMENDER SYSTEMS [SAKO01] ............. 9 FIGURE 2.6 THE CLASSIFICATION OF THE RECOMMENDER SYSTEMS PRESENTED IN THE MASTER THESIS ............................................................................................................................... 10 FIGURE 3.1 MEMBERS OF THE SEMANTIC WEB COMMUNITY AND THEIR ASSOCIATION WITH RESEARCH INTERESTS [STA05] .......................................................................................... 17 FIGURE 3.2 THE DIVISION OF SOCIAL NETWORKS BASED ON THE TYPE OF THE RELATIONSHIP ... 18 FIGURE 3.3 THE EXAMPLE OF RELATIONSHIPS BETWEEN PEOPLE WORKING IN A COMPANY....... 19 FIGURE 3.4 THE DIVISION OF SOCIAL NETWORKS BASED ON THE TYPE OF THE COMMUNICATION CHANNEL........................................................................................................................... 19 FIGURE 3.5 THE TAXONOMY OF CSSN...................................................................................... 20 FIGURE 3.6 SOCIAL FEATURES OF RELATIONSHIPS IN A SOCIAL NETWORK................................. 21 FIGURE 3.7 AN EXAMPLE OF THE ONLINE SOCIAL NETWORK SYSTEM ........................................ 22 FIGURE 3.8 MAIN FUNCTIONS OF THE ONLINE SOCIAL NETWORKS SYSTEMS.............................. 23 FIGURE 3.9 THE PROCESS OF STARTING NEW RELATIONSHIP ..................................................... 24 FIGURE 3.10 THE PROCESS OF MAINTENANCE OF THE RELATIONSHIP ........................................ 24 FIGURE 3.11 THE WAYS OF STIMULATING THE SOCIAL CAPITAL ................................................ 27 FIGURE 3.12 SOCIAL CAPITAL OF USER X GROUPED IN COMPONENTS......................................... 28 FIGURE 4.1 ONLINE AND OFFLINE ELEMENTS OF RECOMMENDATION PROCESS.......................... 33 FIGURE 4.2 USER DATA GROUPED IN PROFILE COMPONENTS ..................................................... 34 FIGURE 4.3 NAME ELEMENT AVAILABLE IN MYSPACE NETWORK ............................................... 35 FIGURE 4.4 BASIC INFO ELEMENT AVAILABLE IN MYSPACE NETWORK...................................... 36 FIGURE 4.5 BACKGROUND & LIFESTYLE ELEMENT AVAILABLE IN MYSPACE NETWORK.............. 37 FIGURE 4.6 DEMOGRAPHIC ATTRIBUTES USEFUL FROM RECOMMENDATION POINT OF VIEW ...... 38 FIGURE 4.7 INTEREST & PERSONALITY ELEMENT AVAILABLE IN MYSPACE NETWORK................. 39 FIGURE 4.8 INTEREST ELEMENT AVAILABLE IN FRIENDSTER NETWORK ..................................... 39 FIGURE 4.9 INTEREST ATTRIBUTES USEFUL FROM RECOMMENDATION POINT OF VIEW ............... 40 FIGURE 4.10 THE HIERARCHY OF THE INTEREST ATTRIBUTES.................................................... 40 FIGURE 4.11 PREFERRED ATTRIBUTES USEFUL FROM RECOMMENDATION POINT OF VIEW ......... 41 FIGURE 4.12 SIMPLE SEARCH AVAILABLE IN FRIENDSTER NETWORK......................................... 42 FIGURE 4.13 ADVANCED SEARCH AVAILABLE IN FRIENDSTER NETWORK .................................. 42 FIGURE 4.14 SEARCH ATTRIBUTES USEFUL FROM RECOMMENDATION POINT OF VIEW ............... 43 FIGURE 4.15 ACTIVITY ATTRIBUTES USEFUL FROM RECOMMENDATION POINT OF VIEW.............. 43 FIGURE 4.16 RELATIONSHIP ATTRIBUTES USEFUL FROM RECOMMENDATION POINT OF VIEW ..... 44 FIGURE 4.17 THE PROCESS OF RECOMMENDATION FOR USER X IN A SOCIAL NETWORK ............. 46 FIGURE 4.18 CLASSES OF VARIABLES USED IN CALCULATION OF DIRECT SIMILARITY FUNCTION .......................................................................................................................................... 50 FIGURE 4.19 STRAIGHT COMPARISON OF PAIRS OF ATTRIBUTES ................................................ 51 FIGURE 4.20 EXAMPLE OF HIERARCHY OF THE SPORTS FOR DATA FROM TABLE 4.3 .................. 53 FIGURE 4.21 SEARCHES MADE BY USER X.................................................................................. 55 FIGURE 4.22 THE EXAMPLE OF VIRTUAL COMMUNITY............................................................... 56 FIGURE 4.23 SENT INVITATIONS AND RESPONSES TO THESE INVITATIONS IN THE COMMUNITY FROM FIGURE 4.22 ............................................................................................................ 57 FIGURE 4.24 THE CREATION OF THE REQUIRED INTERVALS ....................................................... 58 FIGURE 4.25 THE CREATION OF LEVELS OF INITIATORS ............................................................. 58 vi FIGURE 4.26 THE LEVELS OF INITIATORS – EXAMPLE ................................................................ 58 FIGURE 4.27 THE CREATION OF LEVELS OF INITIATORS BY UTILIZING STANDARD DEVIATION ... 59 FIGURE 4.28 THE LEVELS OF INITIATORS – EXAMPLE STANDARD DEVIATION ............................ 60 FIGURE 4.29 THE ACTIVITIES WITHIN SOCIAL NETWORK ........................................................... 61 FIGURE 4.30 THE THEORETICAL RELATIONSHIPS OF USER Y ...................................................... 63 FIGURE 4.31 THE EXAMPLE OF VIRTUAL COMMUNITY............................................................... 65 FIGURE 4.32 EXAMPLES OF SOCIAL NETWORKS WITH CALCULATED SP(X) AND C(Y�X) FOR EACH USER, e=1................................................................................................................. 68 FIGURE 4.33 THE ELEMENTS OF SOCIAL FILTERING ................................................................... 69 FIGURE 5.1 USER DATA GROUPED IN PROFILE COMPONENTS ..................................................... 72 FIGURE 6.1 NEW USER JOINS TO THE EXISTING COMMUNITY ..................................................... 74 FIGURE 6.2 USER PROFILE USED IN THE EXAMPLE ..................................................................... 74 vii LIST OF TABLES TABLE 2.1 THE DIVISION OF RECOMMENDATION TECHNIQUES [BUR02]...................................... 8 TABLE 4.1 THE HIERARCHY AND EXAMPLES OF INTERESTS ....................................................... 41 TABLE 4.2 OCCUPATIONS OF PERSONS' X AND Y........................................................................ 47 TABLE 4.3 COMPARISON OF USERS' X AND Y INTERESTS............................................................. 52 TABLE 4.4 POINTS ASSIGNED FOR LEVELS FOR SPECIFIC HIERARCHICAL INTERESTS.................. 53 TABLE 4.5 POINTS ASSIGNED FOR LEVELS FOR SPECIFIC HIERARCHICAL INTERESTS.................. 54 TABLE 4.6 COMPARISON OF USERS' PLACES OF ORIGIN.............................................................. 54 TABLE 4.7 DEMOGRAPHIC COMPONENT OF USER Y .................................................................... 55 TABLE 4.8 ATTRIBUTES AND THEIR FREQUENCY OF OCCURRENCE IN SEARCH COMPONENT OF USER X............................................................................................................................... 55 TABLE 4.9 SENT, RECEIVED INVITATIONS AND RESPONSES TO INVITATIONS WITHIN COMMUNITY FROM FIGURE 4.22 ............................................................................................................ 57 TABLE 4.10 INITIATIONS LEVELS ASSIGNED TO USERS BASED ON UTILIZING DIFFERENT METHODS .......................................................................................................................................... 60 TABLE 4.11 RESPONSES LEVELS ASSIGNED TO USERS................................................................ 60 TABLE 4.12 FREQUENCY OF CHANGING PHOTO ALBUM IN THE THEORETICAL COMMUNITY....... 62 TABLE 4.13 THE WAYS AND FREQUENCY OF EXCHANGING INFORMATION BETWEEN USER X AND THEIR ALTERS.................................................................................................................... 65 TABLE 4.14 THE WAYS AND FREQUENCY OF EXCHANGING INFORMATION BETWEEN USER Y AND THEIR ALTERS.................................................................................................................... 65 TABLE 4.15 THE WAYS AND FREQUENCY OF EXCHANGING INFORMATION BETWEEN USER V AND THEIR ALTERS.................................................................................................................... 66 TABLE 4.16 THE WAYS AND FREQUENCY OF EXCHANGING INFORMATION BETWEEN USER U AND THEIR ALTERS.................................................................................................................... 66 TABLE 4.17 THE WAYS AND FREQUENCY OF EXCHANGING INFORMATION BETWEEN USER Z AND THEIR ALTERS.................................................................................................................... 66 TABLE 6.1 DATA ABOUT USER X................................................................................................ 75 TABLE 6.2 DATA ABOUT MEMBERS OF THE NETWORK............................................................... 76 viii Relationships matter – ICk4U I dedicate this thesis to my parents and sister ix Recommendation system for online social network 1 Introduction This is an introductory chapter of the thesis. First section enumerates the reasons why the research on recommender systems (RS) in online social networks (OSN) should be done. Second section presents not only the aim and the objectives of the thesis but also research questions for which the answers are presented in the further part of this thesis. Finally, last section outlines the remainder of the thesis. 1.1 Background and Motivation The idea of both online social networks and recommender systems was developed by many researches. All existing recommendation methods suggest products or services to people [AdTu05, MoLo03, SaKo01]. Whereas method presented in this thesis recommends one human being to another. The whole idea of applying recommender systems in social network is innovative, thus many new challenges have occurred. Building such a framework requires integrating two separate branches of knowledge: computer science and sociology. This makes work even more interesting because enables to create interdisciplinary groups in which people with different backgrounds can cooperate with each others. On one hand, when we recommend one person to another, we should have deep knowledge about algorithms that can be utilized and pick the appropriate one. Usually, the chosen method must be further tailored to the specific needs of the system. Additionally, it is common practice that the creators of RS integrate many different recommendation methods because some techniques complement other ones. In other words, one technique usually copes with the shortcoming of other one and this relationship is bidirectional. Nevertheless, not only the knowledge about recommender systems is required. On the other hand, we should know a social behaviour of people who create the network and be aware of their preferences. In the recommender system for online social network not only the needs and preferences of a person who expects some suggestions should be considered, but also the expectations of a person who will be suggested. This factor makes the whole problem even more complicated but also more fascinating. 1.2 Research Questions and Objectives The main aim of this thesis is to create the recommendation method that will support the evolution of online social networks known as virtual communities. The objectives of the thesis are as follows: • Gain knowledge about the subject (recommender systems and social networks) through research in literature • Create the classification of existing social networks • Build the user profile in which different kind of data will be gathered • Present the way in which different sources of data can be integrated • Define the recommendation process for social network that concerns the preferences and human limitations of people who will be recommended The thesis addressing following research questions: • What kinds of recommendation methods do exist? 1

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.