Abstract:

Yioop is an open source search engine, wiki system and user discussion group system. A recommendation system analyzes user behaviour on a website to make suggestions about what a user should do in the future on the website. For this project we will develop a recommendation system for discussion groups for Yioop.

CS297 Results

• Implemented the max flow algorithm - Ford Fulkerson Algorithm and Push Relabel Algorithm.
• Implemented the Recommendation Engine Algorithm using baseline predictors (Collaborative Filtering) for the Yelp dataset
• Improved the manage group activity page in Yioop - added a feature to delete users by their join date and implemented client side validation of register account page.
• Implemented the Recommendation Engine Algorithm using Latent Matrix Factorization (Collaborative filtering) for the Yelp dataset

Proposed Schedule

Key Deliverables:

• Software
  • A baseline predictor (collaborative filtering) class for Yioop recommendation system
  • Use the baseline predictor class to recommend threads to users
  • Use the baseline predictor class to recommend groups to users
  • A user interface for recommending threads and groups to users
  • Test the deliverables so that it integrates with the Yioop codebase. Perform unit testing on the baseline predictor class
• Report
  • CS298 Report
  • CS298 Presentation

Innovations and Challenges

• Processing a corpus of historical data for building the recommendation system. Will optimize the code to reduce the processing time.
• Creating a new class which can be used for recommendations.
• Adding a new feature to Yioop by recommending threads and groups of their interests.

References:

[1] Langville, A. N., & Meyer, C. D. (2012). Who's #1?: The science of rating and ranking. Princeton: Princeton University Press.

[2] Pollett, C. "Open Source Search Engine Software!" Open Source Search Engine Software  Seekquarry. Retrieved on 13 Dec. 2016.