Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Markov Chain Monte Carlo for effective personalized recommendations

Papilaris Michail-Aggelos

Full record

---

Summary

Personalized recommender systems aim to help users access and retrieve relevant information or items from large collections, by automatically finding and suggesting products or services of potential interest. User preferences are difficult to infer, and doing so often requires a tedious elicitation process relying on evidence of others’ behavior. To overcome such limitations, we propose a Bayesian approach for finding personalized top recommendations, by capturing user preferences using a utility function which the system learns via a passive preference elicitation sampling-based framework. In brief, instead of asking the user to specify this function explicitly, which is unrealistic, we explicitly model the uncertainty over the utility function and learn it through feedback, in the form of clicks, provided by the user. The utility function is a linear combination of (weighted) features, and beliefs are maintained using a Markov Chain Monte Carlo algorithm. Additionally, we handle situations where not enough data about the user is available, by exploiting the information from clusters of (feature) weight vectors created by observing other users’ behavior. Finally, in order to evaluate our system’s performance, we applied it in the online hotel booking recommendations domain using a real-world dataset.

Available Files

Services