A Collaborative Filtering based model for recommending graduate schools
Madhav Iyengar, Ayanava Sarkar, Shikhar Singh · 2017
Recently, with the surge of students pursuing graduate studies after completing their bachelors, there is a lack of open source resources which could point out universities and programs, based on an individual's profile. In this paper, we present our novel approach of predicting universities for graduate studies based on one's whole profile. A model is built which is able to predict the list of top-`n' universities based on the user profile. In our implementation, a user profile comprises one's undergraduate grades, graduate examination scores of GRE or GMAT, other exams like TOEFL, research experience and publications, work experience, and number of relevant projects. Data is collected from a variety of sources comprising peoples' profiles, who got through to graduate programs and is fed into the model, in order to serve as benchmark for an incoming query. For the model to predict the list of universities best suiting the user, Collaborative Filtering is used in order to compare the user's profile to the existing dataset. The output is a list of universities, to which an individual could apply to, based on the profile.