ABSTRACT

With the growth of variety in every industry, Customer finds it difficult from N different options available in the market. Its the business responsibility to showcase the best suggested item depending on his/her needs, Ratings of the product, Opinions/Feedbacks of different customers. To make these recommendations very close to the customer bahaviour, we need to process huge existing data to be processed. Hence MapReduce is considered which is emerging as a parallel, distributed paradigm for processing and generating large data sets. This trend combined with the growing need to run Machine Learning (ML) algorithms on massive datasets has led to an increased interest in implementing ML algorithms on MapReduce. Hence using machine learning algorithms in conjunction with big data can bring outright value for any business transformation. This research focuses on a way of analyzing large amount of data to give better recommendations for users by ML algorithms, thereby converting e-commerce site visitors to buyers. We mainly address the challenges of building an efficient and useful recommendation system given a large dataset, and discuss different approaches on identifying like-minded neighbors by making use of similarity, nearest neighborhood, Pearson Correlation higher level ML algorithms. We report probabilities of these kind algorithms with huge amount of data on hadoop clusters