A Probabilistic Approach for Item Based Collaborative Filtering

Abstract

In this era, it is essential to know the customer’s necessity before they know it themselves. The Recommendation system is a sub-class of machine learning which deals with the user data to offer relevant content or product to the user based on their taste. This paper aims to develop an integrated recommendation system using statistical theory and methods. Therefore, the conventional Item Based Collaborative filtering integrated the probabilistic approach and the pseudo-probabilistic approach is proposed to update the k-NN approach. Here we synthesize the data using the Monte-Carlo approach with the binomial and the multinomial distribution. Then we examine the performance of the proposed methodologies on the synthetic data using the RMSE calculation.