COllaborative Filtering Engine

The Intelligent Information Systems (IIS) research group of Oregon State University is pleased to announce Version 0.4 of CoFE: a free Java recommendation engine for collaborative filtering. CoFE is short for "COllaborative Filtering Engine". CoFE was formerly known as CFEngine, but has been renamed due to a popular unrelated software package with the same name.

Collaborative filtering is the process of computing personalized recommendations by finding users with similar taste. Such techniques have been successfully used to recommend newsgroup articles, movies, music, and e-commerce products. Collaborative filtering is most known for its use on popular e-commerce sites such as Amazon.com or NetFlix.com. Today, most recommender systems are proprietary.

This free server (including source code) allows anyone to easily set up a recommendation system. It should run on any platform that supports Java 1.4. We hope it will enhance academic research and commercial application of such systems by removing the need for developers to create such an engine themselves.

This release of CoFE will run as a server to generate recommendations for individual items, top-N recommendations over all items, or top-N recommendations limited to one item type. Recommendations are computed using a popular, well-tested nearest-neighbor algorithm (Pearson's algorithm). This release of the CoFE can be integrated with any system that supports Java. User data is stored in an SQL database, of which MySQL is currently supported.

This release includes a new interface and documentation that makes it easy for developers to create and experiment with algorithms in CoFE. We have also fixed several bugs since the previous version, and have cleaned up some issues with user caching.

The main goals for future releases are greater developer control of the system via interfaces for other components (such as DataManager and caching), cleaner code, and instructions for how to do common tasks with CoFE. Please see the file RELEASE_NOTES in the distribution for more information. Any feedback on the development interfaces or overall usability and usefulness would be appreciated.

Included with the release are manuals describing more about collaborative filtering, the architecture of the CoFE, and how to integrate existing software with the CoFE. We have also included a general setup guide for installing and configuring CoFE for new users, as well as a detailed document describing how to use the new algorithm interface.

Visit http://www.cs.orst.edu/iis/CoFE to download CoFE or obtain more information.

To prepare for 0.4 release, CoFE homepage gets a make-over, now it is powered by Drupal.
Drupal is a dynamic web site platform which allows an individual or community of users to publish, manage and organize a variety of content, Drupal integrates many popular features of content management systems, weblogs, collaborative tools and discussion-based community software into one easy-to-use package.