7.3.2 Presenting Ordered Reviews 133 7.4 User study 135 7.5 Conclusions and Discussion 137 8 TrustMail: Trust Networks for Email Filtering 137 8.1 Background and introduction. 140 8.2 The TrustMail Application 142 8.3 Case Study: The Enron Email Corpus 144 8. 4 Conclusions 146 9 Conclusions 10 Future Work 154 10.1 Validation of Current results 155 10.2 Extensions to Current Work 155 10.2.1 Network Structure and Trust Inferences 157 10.2.2 Recommendations with filmtrust 159 10.2.3 Privacy 10.3 Filtering Semantic Web statements with Trust 162 10.3.1 From Trust Network Inferences to Accepting Claims. 163 10.3.2 Using Claim Ratings in Semantic Web systems 10.3.3 Filtering Inferences in Knowledge Bases with Trust values 167 10.4 Meal of a Meal: Inferring Trophic Relationships in Food Webs 10.5 Conclusions and Vision 171 ererences
viii 7.3.2 Presenting Ordered Reviews. . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 User Study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 Conclusions and Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 TrustMail: Trust Networks for Email Filtering 8.1 Background and Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 The TrustMail Application. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Case Study: The Enron Email Corpus. . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Conclusions 10 Future Work 10.1 Validation of Current Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 Extensions to Current Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.1 Network Structure and Trust Inferences . . . . . . . . . . . . . . . . . 10.2.2 Recommendations with FilmTrust . . . . . . . . . . . . . . . . . . . . . . 10.2.3 Privacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 Filtering Semantic Web Statements with Trust. . . . . . . . . . . . . . . . . . 10.3.1 From Trust Network Inferences to Accepting Claims. . . . . . . . 10.3.2 Using Claim Ratings in Semantic Web Systems. . . . . . . . . . . . 10.3.3 Filtering Inferences in Knowledge Bases with Trust Values. . . 10.4 Meal of a Meal: Inferring Trophic Relationships in Food Webs. . . . . 10.5 Conclusions and Vision. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References 133 135 137 137 140 142 144 146 154 154 155 155 157 159 160 162 163 164 167 170 171
LIST OF TABLES 2.1 Million member Web-based Social Networks 16 2.2 Cate gories of Web-based Social Networks 18 2.3 A sampling of websites with options they offer for describing relationship 2.4 A sampling of websites that allow relationship features to be rated on a numeric scale 2.5 FOAF Vocabulary Summary 2.6 WBSNS that provide foaf profiles of users' social networks 3.1 The range of values available for rating trust in web-based social networks 4.1 Payoff Matrix for the Prisoner's Dilemma. 6. 1 Data from experiments run to determine accuracy 6.2 ANOVA Analysis of the results in Table 6.1 6.3 ANOVA Analysis of Randomized Network Results 888 6. 4 Expected A in original and randomized networks 6.5 Average A values for paths from source to sink of length 2, 3, and 4 6.6 Average A values for paths from source to sink of length 5 6.7 6.8 Minimum average A for paths of various lengths 6.9 verage a for the methods of inferring trust 114 Average A for TidalTrust, Beth-Borcherding-Klein(BBK), and simple
ix LIST OF TABLES 2.1 2.2 2.3 2.4 2.5 2.6 3.1 4.1 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 Million Member Web-based Social Networks . . . . . . . . . . . . . . . . . . . . . . . Categories of Web-based Social Networks . . . . . . . . . . . . . . . . . . . . . . . . . . A sampling of websites with options they offer for describing relationships A sampling of websites that allow relationship features to be rated on a numeric scale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . FOAF Vocabulary Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . WBSNs that provide FOAF profiles of users' social networks. . . . . . . . . . . The range of values available for rating trust in web-based social networks Payoff Matrix for the Prisoner's Dilemma. . . . . . . . . . . . . . . . . . . . . . . . . . . Data from experiments run to determine accuracy. . . . . . . . . . . . . . . . . . . . . ANOVA Analysis of the results in Table 6.1. . . . . . . . . . . . . . . . . . . . . . . . . ANOVA Analysis of Randomized Network Results . . . . . . . . . . . . . . . . . . . Expected Δ in original and randomized networks . . . . . . . . . . . . . . . . . . . . . Average Δ values for paths from source to sink of length 2, 3, and 4. . . . . . Average Δ values for paths from source to sink of length 5. . . . . . . . . . . . . . Minimum average Δ for paths of various lengths. . . . . . . . . . . . . . . . . . . . . . Average Δ for the methods of inferring trust. . . . . . . . . . . . . . . . . . . . . . . . . Average Δ for TidalTrust, Beth-Borcherding-Klein (BBK), and simple 16 18 21 22 26 41 80 81 83 86 92 93 94 101 114
avera X
x average. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
LIST OF FIGURES 2.1 WBSN membership for sites ranked by population 3. 1 Network paths for discovering trust 4.1 Finding Trusted Paths to the sink. 5.1 An illustration of how nodes are used in trust inferences 61 5.2 A map of how the initial accuracy in the system changes with g and p 5.3 The increasing probability of a correct trust inference......... 65 5.4 A comparison of the initial accuracy of trust ratings with the accuracy of inferred ratings using the rounding algorithm 5.5 The accuracy of inferred ratings are shown for various initial percentages of ood nodes 5.6 Accuracy of Recommendations Compared to Initial Accuracy Using Non- Rounding algorithm 5.7 Accuracy of Recommendations Using Rounding algorithm. 69 5.8 A comparison of the accuracy of trust inferences made with the rounding and non-rounding algorithms 6. 1 The structure of the Trust Project,'s network. 6.2 The structure of the filmtrust social network 6.3 The distribution of trust ratings in the Trust Project network 6. 4 Finding points of comparison in the network. 78 6.5 Distribution of trust ratings in the original network and experiments
xi LIST OF FIGURES 2.1 3.1 4.1 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 6.1 6.2 6.3 6.4 6.5 WBSN membership for sites ranked by population. . . . . . . . . . . . . . . . . . . . . Network Paths for Discovering Trust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Finding Trusted Paths to the Sink. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . An illustration of how nodes are used in trust inferences . . . . . . . . . . . . . . . . A map of how the initial accuracy in the system changes with g and pa. . . . . The increasing probability of a correct trust inference. . . . . . . . . . . . . . . . . . A comparison of the initial accuracy of trust ratings with the accuracy of inferred ratings using the rounding algorithm . . . . . . . . . . . . . . . . . . . . . . . . . The accuracy of inferred ratings are shown for various initial percentages of good nodes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Accuracy of Recommendations Compared to Initial Accuracy Using NonRounding Algorithm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Accuracy of Recommendations Using Rounding Algorithm. . . . . . . . . . . . . A comparison of the accuracy of trust inferences made with the rounding and non-rounding algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The structure of the Trust Project's network. . . . . . . . . . . . . . . . . . . . . . . . . . The structure of the FilmTrust social network. . . . . . . . . . . . . . . . . . . . . . . . . The distribution of trust ratings in the Trust Project network . . . . . . . . . . . . . Finding points of comparison in the network. . . . . . . . . . . . . . . . . . . . . . . . . Distribution of trust ratings in the original network and experiments . . . . . . 17 37 45 61 62 65 66 67 68 69 70 73 74 77 78 79
6.6 The relationship between A and Trust Rating 6.7 Averagea by trust value 6.8 Trust distributions in the original network and experiments 6.9 Distribution of trust ratings in the original network and among the pairs with common neighbors in the randomized networks 85 6.10 Paths from the source to sink of length two three. and four 6. 11 Paths of length 2. 3. 4. 5 and 7 6. 12 An illustration of how A n? values are derived for a path length of four. 91 6. 13 Minimum average a from all paths of a fixed length containing a given trust all 6. 14 The process of determining the trust threshold 6.15 A network illustrating when lines 25-28 allow for more children 105 7.1 A users' friend listing at the FilmTrust website 121 7.2 A user's movies page with titles, ratings, reviews, and options 122 7.3 The move ratings and reviews page for Jaws 7.4 Average Sa and Sr values for an increasing minimum sa threshold 127 7.5 A user's view of the page for "A Clockwork Orange" 129 7.6 The increase in s as the minimum s is increased 131 8.1 The TrustMail Interface 140 10.1 A sample social network 10.2 Inferring trophic relationships in food webs 168
xii 6.6 6.7 6.8 6.9 6.10 6.11 6.12 6.13 6.14 6.15 7.1 7.2 7.3 7.4 7.5 7.6 8.1 10.1 10.2 The relationship between Δ and Trust Rating. . . . . . . . . . . . . . . . . . . . . . . . . Average Δ by trust value. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Trust distributions in the original network and experiments. . . . . . . . . . . . . . Distribution of trust ratings in the original network and among the pairs with common neighbors in the randomized networks. . . . . . . . . . . . . . . . . . . Paths from the source to sink of length two, three, and four. . . . . . . . . . . . . . Paths of length 2, 3, 4, 5, and 7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . An illustration of how Δs,n2 values are derived for a path length of four. . . . . Minimum average Δ from all paths of a fixed length containing a given trust value. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The process of determining the trust threshold . . . . . . . . . . . . . . . . . . . . . . . . A network illustrating when lines 25-28 allow for more children . . . . . . . . . A users' friend listing at the FilmTrust website. . . . . . . . . . . . . . . . . . . . . . . . A user's movies page with titles, ratings, reviews, and options. . . . . . . . . . . . The move ratings and reviews page for Jaws. . . . . . . . . . . . . . . . . . . . . . . . . Average δa and δr values for an increasing minimum δa threshold . . . . . . . . A user's view of the page for "A Clockwork Orange" . . . . . . . . . . . . . . . . . . The increase in δ as the minimum δa is increased . . . . . . . . . . . . . . . . . . . . The TrustMail Interface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A sample social network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inferring trophic relationships in food webs . . . . . . . . . . . . . . . . . . . . . . . . . . 81 83 84 85 88 89 91 94 102 105 121 122 123 127 129 131 140 156 168