Archived Past Project Ideas

Recommender Systems’ survey and usability study [Archived]

Title: Recommender Systems’ survey and usability study
Duration: One Semester
Responsible TA: Nicolas Jones
Goals: The goal of this project is to study a number of recommender technologies and to find ways to improve them.
Keywords: qualitative user study design, recommender systems, usability study
Abstract:

Wikipedia defines recommender system as “programs which attempt to predict items (movies, music, books, news, web pages) that a user may be interested in, given some information about the user’s profile”. These systems are currently growing more and more popular. Everyday, new websites incorporate features and tools based on recommender technologies. However systems seem to be tightly linked to the field of activity. So can we see a pattern emerging? Could it just be that the usability of an implementation sells, respectively kills, a solution in respect to another and has lead us to believe that solutions were domain-specific?

This project is divided into two main steps and a small last one (this can evolve, depending on the student’s interrests). The first step would be to perform a survey of different recommender systems currently popular. If possible, the student should try to find the main kind of systems that are used, and explain why they are popular. The second step would focus on making a usability study of a selection of the systems found in step one in order to “measure” how much the usability concerns influence the efficiency of a recommender system. Finally, as a third step, the student may implement a prototype module of one of the systems selected in step 2, into the open source ecommerce system: osCommerce. 

Project Results:
  • a survey report of 10+ recommender systems
  • a usability study report of 3-5 recommender systems
  • a system prototype of one recommender system on a specific domain based on oscommerce
Required Skills
  • HTML
  • PHP/Mysql

Modelisation of users’ adaptive decision behaviour [Archived]

Title: Modelisation of users’ adaptive decision behaviour
Duration: One Semester
Responsible TA: Nicolas Jones
Goals: The goal of this project is to try and model user’s adaptive decision behaviour.
Keywords: statistical analysis, user modelling 
Abstract: Many methods have been used to try and model user’s adaptive decision behaviors. Unfortunately, many of them suffer from problems. Artificial Intelligence has proposed the concept of DBN: dynamic bayesian networks which can potentially be used to do the job. The goal of the project will be to define a framework, and then perform an implementation.
Project Results:
  • implementation of DBN for modeling user’s adaptive decision behaviors
Required Skills
  • algorithmic abstraction qualities
  • HTML/PHP/Mysql (or other programming language – will be determined when chosing the framework)

Recommender System in the EPFL framework [Archived]

Title: Recommender System in the EPFL framework
Duration: One Semester
Responsible TA: Nicolas Jones
Goals: The goal of this project is to design and implement a simple recommender system for EPFL users.
Keywords: recommender systems
Abstract: This project combines a little research and a real implementation. The first part focuses on finding a framework that students use regularly and where they would be eager to rate. First ideas are rating teachers or courses. The challenge of this first step is to have as many people as possible involved in this procedure on a regular basis. The second part of the project is to implement a recommender system, according to results of step one. The solution must be simple, and can be an item-to-item or user-to-user recommender. The goal for the lab is to have a first-hand experience on how the existing methods work and provide some insights on how to improve it.
Project Results:
  • research for an adequate framwork at EPFL for students
  • implementation of a recommender system for rating
Required Skills
  • previous experience of web interfaces
  • HTML/PHP/Mysql (or other programming language – will be determined when chosing the framework)