Guide: Dr. B. Ravindran.
About: I love mathematics. All types of sports are my hobbies.
Research Interest: I am interested in Probabilty, Machine Learning, Data Mining and Optimization. Current focus is on Bayesian Probabilistic Matrix Factorization.
Current Project: Factorization Machine (FM) a generic matrix factorization framework which combines advantages of SVM and factorization model. Like SVM, FM is a general predictor working with any real valued feature vector. However in contrast to SVM, FM models all interactions between variables using latent factors. Thus FM can estimate interactions even in problems with huge sparsity where SVM fails. There are many different factorization models like matrix factorization, parallel factor analysis or specialized models like SVD++, PITF or FPMC which can be mimiced by FM by feature engineering which makes FM very promising model for Recommendation. Drawback with FM is that it assumes the data comes from Gaussian which is a restrictive assumption. We are relaxing this assumption to make it more flexible.
Publication: A. Saha, J. Rajendran, S. Shekhar, and B. Ravindran, “How popular are your tweets?,” in Proceedings of the 2014 Recommender Systems Challenge, RecSysChallenge ’14, (New York, NY, USA), pp. 66:66–66:69, ACM, 2014. File:Paper.pdf File:Poster.pdf
Challenge: We stood rank 8 in ACM RecSys compitition 2014.