| Summary: | Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, February 2008. === Includes bibliographical references (p. 59). === Collaborative filtering and information filtering are tried and proven methods of utilizing aggregated data about a website's users to provide catered content. Passive filters are one subset of such algorithms that utilize data about a user's interactions with a website in viewing content, purchasing items, etc. My work develops a set of extensions for Ruby on Rails that, when inserted into an existing application, will comprehensively log information associated with different types of user interactions to provide a sound base for many passive filter implementations. The extensions will log how users interact with the application server (content accessed, forms submitted, etc) as well as how users interact with that content on their own browser (scrolling, AJAX requests, JavaScript calls, etc). Given existing open-source collaborative filtering algorithms, the ability to automatically aggregate user-interaction data in any arbitrary Rails application significantly decreases the barrier to implementing passive filtering in an already efficient agile web development framework. Further, my work utilizes the logged data to implement a web interface to view analytic information about the components of an application. === by Daniel R. Malconian. === M.Eng.
|
|---|
