Collaborative Filtering (CF) is a Information Retrieval (IR) technique that uses of a set of explicitly or implicitly gathered/derived user preferences as its information quality/relevance measure. For example MovieLens is a typical CF system that collects movie preferences from users and then groups users with similar tastes. Based on the movie ratings expressed by all the users in a group it attempts to predict for each individual their opinion on movies they have not yet seen. Most of the projects in our research group focus on collaborative filtering (CF) and its integration into various applications. Our movie recommeder, MovieLens allows us to perform experiments in a "real-world" environment. Our projects can be divided into the following categories:
Algorithm design, implementation and evaluation.
The goal of these projects is to find new methods to perform collaborative filtering. While there is a broad literature discussing various CF methods the applicability of these for different user tasks is less studied.
UI design and evaluation
Since we have access to a large userbase we are able to deploy and evaluate interfaces obeying various UI design paradigms.
CF and Emerging Technologies
These project study collaborative filtering in the context of emerging technologies such as handhelds, internet enabled devices and smart appliances.
CF on different domains.
Traditionally Collaborative Filteringis applied in E-Commerce. But there are indications that it is well suited for other domains too. One such domain that we are exploring is considering CF in a research paper recommendation service.
Social Interaction an CF
This is a domain that explores the social aspects of collaborative filtering.
Online Communities
Recommender systems, social psychology theory and online communities. How can these three things play nicely together to create interesting research and interfaces?