The rapid expansion of Web 2.0 has converted the Web into a multitude of user-generated applications, which in turn generates a very large and noisy dataset of logs. These datasets represent an extremely valuable source of information, strategic to support marketing, business intelligence, to enable personalization, customer profiling and segmentation, to enhance information retrieval with social search engine.

One of the main challenges remains how to effectively mine a large set of complex data affected by great level of noise, represented by non-pertinent, untrustworthy or even malicious data. Trust and Recommender systems, along with personalization techniques, appear to be essential candidates to enhance and support an effective analysis of web activity.

Trust, Recommender and Personalization mechanisms could help filter, interpret and rank web-users' behaviour to assign the relevance of web-search results and deliver the most reliable and adequate content. Similarly, the may be helpful to deploy ne user-based anti-spam techniques, support web-analytics applications that mine only trustworthy users activity, help users segmentation and decisions support tools for online marketing. Trust is a social concept that happens in society. A trustworthy person gains trust during time thanks to its behaviour in the virtual society he is part of, leaving around him footprints of its trustworthiness.

Trust can be assessed by identifying and properly understanding these footprints among logs datasets. This identification can be performed by explicit, rule-based models of trustworthiness, by pattern-matching trust concept into logs, by implicit techniques such as data mining or machine learning. This workshop focuses on discussing and identifying the most promising research directions with respect to applying trust systems in the analysis and exploitation of web-activity logs. The main focus is both on the theoretical challenge of the definition of effective trust models for large dataset of activity logs and their applications.

The workshop brings together researchers from Recommender/Trust Systems as well as Data Mining, Multi-agent Systems, Information Retrieval, Machine learning and Human Computer Interaction.

Topics

Topics of interest include, but are not limited to:

* Data-mining techniques to compute trust over Web-logs
* Explicit trust-based modelling system to process user activity
* Trust, Reputation and Recommender systems for Social Search
* Trust metrics and Recommender systems in business intelligence, online marketing, user segmentation
* Models of Trust based on logs-processing for online communities and wikis
* Machine learning techniques to compute trust Automatic content quality assessment of user generated content and Web-News
* Trust and Reputation systems based on User Behavioural Analysis
* Trust to enhance personalization
* Anonymity and privacy-aware solutions for Social computing logs mining
* Web Security, Integrity, Privacy and Trust in processing web-logs
* Applications of recommender and trust systems to web analytics, information retrieval, marketing analysis, social search
* Web Information Filtering and Retrieval in social networks and multi-agent systems
* Computational Trust techniques for Collective Intelligence and decision making processes