Trust-Based Rating Prediction for Recommendation in Web 2.0 Collaborative Learning Social Software_Na Li

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Trust-Based Rating Prediction for Recommendation in Web 2.0 Collaborative Learning Social Software

Na Li, Sandy El Helou, Denis Gillet

Real-Time Coordination and Distributed Interaction Systems (ReAct) Automatic Control Lab, Swiss Federal Institute of Technology in Lausanne

ITHET 29th April – 1st May 2010, Cappadocia, Turkey

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Outline

•  Introduction • Collaborative Learning Domain •  3A Interaction Model •  Trust-Based Rating Prediction Approach • Evaluation and Results • Conclusion and Future Work

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Introduction • Web 2.0 social software ▫  A large amount of user generated content ▫  New challenge: selection of useful resources

RSS Feeds

Pictures

Documents

Videos

Wiki Pages

Pictures

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Introduction

• Rating systems ▫  Evaluate quality of content in open environment ▫  Provide recommendation for different users

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Introduction • Rating systems – application level

• Rating systems – academic research level ▫  TidalTrust (J. Golbeck), MoleTrust(P. Massa) ▫  User explicitly specifies a trust value towards another user ▫  Build trust network, Random walk in trust network ▫  Personalized rating prediction

Epinions   1 to 5 stars   A set of aspects for shops and products (ordering, delivery, service)   Status for members (Advisor, Top reviewer, Category Lead)

ePractice.eu   Use “Kudos” to measure the activity of members   Award a number of “Kudos” according to each user action

Everything2   “Positive” and “Negative” votes for articles   Users’ ranking according to their contribution

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Collaborative Learning Domain

• Collaborative learning environment ▫  Unlike e-commerce and review sites ▫  Gift economy

• Rating systems ▫  Evaluate user generated content ▫  Filter helpful learning resources, peers and group

activities ▫  Personalized rating prediction for recommendation

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

3A Interaction Model

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Trust-Based Rating Prediction Approach

• Objective ▫  Build users’ trust network using 3A graph structure ▫  Personalize the rating prediction ▫  Infer trust value in an implicit way

• Basic idea ▫  What influences rating opinion: similarity and

familiarity ▫  People tend to trust the opinions of acquaintance and

those having similar interests and tastes.

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Trust-Based Rating Prediction Approach

•  Trust measurement ▫  Multi-relational trust metric ▫  Build a “Web of Trust” for a particular user using

heterogeneous types of relationships •  Trust Inference ▫  Direct trust ▫  Indirect trust

Trust

How Much?

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Trust-Based Rating Prediction Approach

• Direct trust (DT): derived from a particular type of relationship

W (Membership): weight of “membership” relationship N (Alice, Membership): number of group activities Alice joins

Alice Advanced

Algorithms Group Activity

Is Member of

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Trust-Based Rating Prediction Approach •  Trust propagation •  Propagation distance (PD)

Alice

French Learning Activity

Is Member

Article Create

Video

Propagate

Luis Has Member

Rated by Sara

Rated by Ben

Bob

Commented by

Jack Propagate Propagate

PD

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Trust-Based Rating Prediction Approach

•  Indirect Trust (IT) Inference

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Trust-Based Rating Prediction Approach

• Rating prediction from a user to an item ▫  Using user’s “Web of Trust” ▫  People in “Web of Trust” are seen as trustable ▫  Average of all the rating scores given by trustable

people, weighted by their trust value

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Evaluation and Results • Using Remashed data set ▫  50 users, 6000 items, 3000 tags and 450 ratings ▫  “Leave-one-out” method ▫  Compare “predicted score – actual score” deviation of

trust-based prediction and simple average

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Evaluation and Results • Change parameters ▫  Weights for relationships doesn’t make a significant

difference in rating prediction ▫  Increasing size of trust network might add noise, lead

to bigger prediction error

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Conclusion and Future Work

•  Propose a trust-based rating prediction approach, inferring trust in an implicit way

•  Provide personalized rating prediction so as to evaluate user-generated content in collaborative learning environment

•  Future deploy and evaluation will be conducted in a collaborative learning platform, namely Graaasp(graaasp.epfl.ch)

Swiss Federal Institute of Technology in Lausanne EPFL, CH-1015 Lausanne, Switzerland

Questions?