Social Interactome Recommender Project


Our team is working with the Social Interactome team to assist in coding the recommender functionality for the Social Interactome network. That is supported by a website and system (modified version of Friendica) designed by the Social Interactome team to help recovering addicts. The team used Python to parse participant’s answers to survey questions, and applied an algorithm to that data to show each participant's most favorable friend matches. The team is working in concert with Prashant Chandrasekar, a Graduate Research Assistant (GRA). He provided us access to the participant’s answers to survey questions. As more and more surveys are filled out by users the team will continue to refine their algorithm to accommodate that extra data. As a further step we will work towards a hybrid recommender which will incorporate not only content, but also collaborative-based recommending.

Contains a Word document and PDF of the final report, and the Powerpoint and PDF for our final presentation.
Friendica, SIRecommender, Similarity Matrix, Homophily, Social Interactome, recommender, Addiction Recovery Research Center