Towards trust-aware recommender systems
1 like807 views
The document describes a thesis that aims to develop a trust-aware recommender system for Twitter. It introduces recommender systems and discusses computing trust through direct interactions and propagating trust through random walks. It then describes building a Twitter crawler and prototype recommender system that scores tweets based on trust and content. Experiments analyze trust properties in Twitter and benchmark the recommender system. The thesis contributes a trust metric, crawler, and initial recommender system while finding that trust models may be useful but do not exhibit transitivity in Twitter.
![Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions
Recommender
Scoring: Trust-aware + Content-based
Features:
• has_url [True, False]
• bag-of-words or tf ratio
• trust [0,1]
Label:
• retweet [True, False]](https://image.slidesharecdn.com/masterthesispresentationv2-120706135546-phpapp02/85/Towards-trust-aware-recommender-systems-22-320.jpg)
![Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions
Transitivity tests
Normalize trust rankings [0-10]
∆: Disagreement about ranking of common neighbors.
If transitivity: the higher trust on a user, the smaller ∆ between their ratings](https://image.slidesharecdn.com/masterthesispresentationv2-120706135546-phpapp02/85/Towards-trust-aware-recommender-systems-25-320.jpg)
![Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions
Transitivity tests
Interactions
Question: are interactions of Twitter users transitive?
Table: Relation of interactions rank and delta
Ranking [0-10] ∆
[0 − 1) 1.36
[1 − 2) 0.75
[2 − 3) 1.14
[3 − 4) 0.83
[4 − 5) 0.78
[5 − 6) 0.52
[6 − 7) 1.06
[7 − 8) 0.53
[8 − 9) 0.57
[9 − 10) 0.17
Agreement about common neighbors (no matter the ∆)
... but we do not see transitivity.](https://image.slidesharecdn.com/masterthesispresentationv2-120706135546-phpapp02/85/Towards-trust-aware-recommender-systems-26-320.jpg)
Towards trust-aware recommender systems
- 1. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Towards trust-aware recommender systems Alberto Lumbreras Ricard Gavaldà (Advisor) July 2012
- 2. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions 1. Introduction 2. Recommender systems 3. Related work 4. A recommender system for Twitter 5. Experiments 6. Conclusions
- 3. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions 1. Introduction 2. Recommender systems 3. Related work 4. A recommender system for Twitter 5. Experiments 6. Conclusions
- 4. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Introduction The paradox of choice: "more is less" Recommender systems to reduce information overload Social networks information to enhance recommendations
- 5. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Aims of the Thesis Recommend tweets to users based on their social network information Studying the concept of trust in Twitter Can trust improve tweet recommendations? What other techniques are useful?
- 6. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions 1. Introduction 2. Recommender systems 3. Related work 4. A recommender system for Twitter 5. Experiments 6. Conclusions
- 7. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Recommender systems Collaborative Filtering: • Look at "similar users" (similar ratings) • Average their ratings weighted by closeness. • Sparsity • Cold-start problem Content Based: • Based on items similarity / machine learning • Features extracted automatically or by experts • Cold-start problem • What features are relevant? Trust-aware: • Use trust to enhance recommendation methods • What is trust? • How to compute and propagate trust?
- 8. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Computing Trust Direct trust computation: • Explicit: Users annotations • Implicit: Inferred from users’ behavior and/or interactions Trust propagation: • Algorithms fit network properties (decay, trust horizon,...) • Network as Markov Chain: PageRank, EigenTrust... • ... Trust-aware recommendations: • Trust + Collaborative Filtering • Trust + Content Based Filtering • ...
- 9. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions 1. Introduction 2. Recommender systems 3. Related work 4. A recommender system for Twitter 5. Experiments 6. Conclusions
- 10. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Related work TidalTrust: • Movie recommendation. • Customized algorithm for trust propagation • Direct trust explicitly annotated by users (0-10 range) • Pro: Do not normalize trust • Con: Requires annotated direct trust EigenTrust: • Reputation in P2P networks • Direct trust inferred from proportion of successful downloads • Trust propagation by Random Walk • Distributed computation • Pro: Simple algorithm • Con: Do not explicitly consider network properties
- 11. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions 1. Introduction 2. Recommender systems 3. Related work 4. A recommender system for Twitter 5. Experiments 6. Conclusions
- 12. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Goals and challenges GOAL: Recommend tweets to users based on their social network information CHALLENGES: In GETTING the data: • APIs limits • Implement a crawler • Crawling criteria In ANALYZING the data: • No explicit feedback mechanism • Textual items of 140 characters • Items volatility (hours or days) • Very high sparsity (items rated (retweets) by few users)
- 13. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Contributions Trust metric (for social networks) Trust-aware crawler (for social networks) Recommender system prototype Analysis of trust properties in Twitter
- 14. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Computing trust in Twitter Interactions in Twitter tweet: • @xamat: Markov Random Fields for #recsys: http://t.co/zHggOl2r mention: • @bob: Nice tutorial @xamat! • @bob: Not a very good tutorial @xamat... • trust or distrust (assumption: mostly trust) retweet (forward) : • @charles: RT "@xamat: Markov Random Fields for #recsys: http://t.co/zHggOl2r" • @charles: So bad! RT "@xamat: Markov Random Fields for #recsys: http://t.co/zHggOl2r" • trust or distrust (observation: mostly trust) favorite: • store a friend’s tweet • observation: mostly trust follow/unfollow : • users subscribe/unsubscribe to other users publications
- 15. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Computing trust in Twitter Direct trust Mentions and retweets as a sign of trust Trust tij as proportion of interactions of user i with user j. (m) (rt) wNij + (1 − w )Nij tij = (1) Ni Temporal decay of interactions: Ni (t) = λNi (t) + (1 − λ)Ni (t − 1) (2)
- 16. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Computing trust in Twitter Trust propagation through random walk Step 1: With direct trusts, build transition probabilities matrix P 0 0.2 0.8 0 0 0 0.5 0.5 P= 0 0 0 1 0.3 0.3 0.3 0
- 17. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Computing trust in Twitter Trust propagation through random walk Step 2: Propagate trust 1 T3 = (α1 P +α2 P2 +α3 P3 ) (3) 3 walk 1 step walk 2 steps walk 3 steps s 1 Ts = αn P n (4) s n=1 P: direct trust matrix (transition matrix) s: trust horizon αn : path length penalization αn Note: Assumes transitivity
- 18. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Architecture
- 19. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Crawler Algorithm 1: Crawling cycle while True do S ←− SeedUsers() ∪ TopTrustedUsers() foreach s ∈ S do statuses ←− GetLastUpdates(s) UpdateInteracctionsMatrix(s, statuses) end TruncateInteractionsMatrix() /*remove unsignificant users*/ UpdateTrustMatrix() UpdateTopTrustedUsers() end
- 20. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Recommender Optional: Tweet expansion. Optional: bag-of-words or tfPOS /tfNEG ratio. Tweet candidates from top-trusted neighborhood Learn to predict retweets.
- 21. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Recommender Query (tweet) expansion Query expansion: • Bag of words probably not enough. Too few word coincidences between tweets • If expanding the query (i.e: synonyms) more chances to get coincidences • Query expansion proved useful in some scenarios (i.e: for QA systems with search engines) Tweet expansion: • Query Bing with tweets • Get first 200 results (summaries) • Add summaries words to tweet
- 22. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Recommender Scoring: Trust-aware + Content-based Features: • has_url [True, False] • bag-of-words or tf ratio • trust [0,1] Label: • retweet [True, False]
- 23. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions 1. Introduction 2. Recommender systems 3. Related work 4. A recommender system for Twitter 5. Experiments 6. Conclusions
- 24. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Dataset 20 target users 6 month crawling Spanish, Catalan, English 314 instances/user (50% retweets, 50% non-retweets) 70% training, 30% test. Offline testing (a posteriori prediction of retweets)
- 25. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Transitivity tests Normalize trust rankings [0-10] ∆: Disagreement about ranking of common neighbors. If transitivity: the higher trust on a user, the smaller ∆ between their ratings
- 26. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Transitivity tests Interactions Question: are interactions of Twitter users transitive? Table: Relation of interactions rank and delta Ranking [0-10] ∆ [0 − 1) 1.36 [1 − 2) 0.75 [2 − 3) 1.14 [3 − 4) 0.83 [4 − 5) 0.78 [5 − 6) 0.52 [6 − 7) 1.06 [7 − 8) 0.53 [8 − 9) 0.57 [9 − 10) 0.17 Agreement about common neighbors (no matter the ∆) ... but we do not see transitivity.
- 27. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Transitivity tests Trust different path decays (αn ) Table: Relation of trust rank and delta Ranking ∆ ∆ ∆ (0-10) No decay Linear decay Exp.decay (0-1) 1.14 0.90 0.87 (1-2) 1.10 1.22 1.09 (2-3) 0.96 1.05 1.05 (3-4) 1.07 1.18 1.20 (4-5) 1.09 0.81 1.01 (5-6) 1.13 1.11 0.92 (6-7) 0.91 1.08 1.16 (7-8) 0.90 1.08 0.99 (8-9) 1.3 1.34 1.34 (9-10) 1.26 1.11 1.16 we do not see transitivity
- 28. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Trust-aware recommendations Benchmarking trust slightly improves recommendations Classif. Expand Encode Acc. Recall Precision F1 AUC bow 50.76 59.43 52.07 55.51 50.93 yes tf 52.38 58.99 52.52 55.57 52.58 NB bow 48.79 53.01 47.76 50.25 49.88 no Trust tf 51.97 50.49 50.60 50.44 51.51 bow 45.84 61.00 30.22 40.42 50.53 yes tf 47.63 51.36 46.95 49.06 47.94 SVM bow 46.25 68.42 32.47 44.04 50.00 no tf 47.79 46.38 48.44 47.39 47.32 Averages 48.93 56.13 45.13 49.08 50.09 bow 47.02 57.58 48.90 52.89 49.56 yes tf 51.13 49.56 54.81 52.05 52.22 NB bow 49.28 47.98 50.76 49.33 50.54 No trust no tf 46.12 45.18 45.84 45.51 46.28 bow 45.22 55.83 27.05 36.44 50.06 yes tf 49.23 49.36 51.47 59.39 49.80 SVM bow 43.40 34.87 17.59 23.38 50.22 no tf 47.11 41.55 48.87 44.91 47.32 Averages 47.31 47.73 43.16 45.49 49.50
- 29. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions 1. Introduction 2. Recommender systems 3. Related work 4. A recommender system for Twitter 5. Experiments 6. Conclusions
- 30. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Contributions and findings Contributions: Trust metric Trust-aware crawler for social networks Recommender system prototype Analysis of trust properties in Twitter Findings: No transitivity of trust in Twitter or bad trust model... ...but trust model might be useful
- 31. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Publication A. Lumbreras, R. Gavaldà, “Applying trust metrics based on user interactions to recommendation in social networks” in Social Knowledge Discovery and Utilization Workshop within IEEE/ACM ASONAM’2012, Istambul, August 2012.
- 32. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Future work Other query expansion techniques Further text analysis (e.g: LSA) Apply temporal decay to tweets Further study of network properties (trust, interactions, visualization...) User tests Study marginal contribution of retweets and mentions Topic-aware trust (topic detection) Open question: (how much) transitivity-based models can capture trust on a non-transitive trust network?
- 33. Introduction Recommender systems Related work A recommender system for Twitter Experiments Conclusions Questions?