High Rating Recent Preferences Based Recommendation System

Abstract

The availability of huge amount of information on Web makes it difficult for users to dissect relevant information from the unnecessary and irrelevant information. This paper highly facilitates the filtering of irrelevant over-abundant stuff in automotive manner. It efficiently plummet the complexity of search space for users and hence attract more users on web and ultimately increases the earn benefits of site holders. In this paper, an attempt has been made to design a high rating recent preferences based recommendation system by using Item-to-Item collaborative filtering. The movie data set is used to provide users’ recommendations based on ratings, and classified data. Classification is done in WEKA data mining tool using J48 pruned tree based classifier. The recommendations contain only high rated movies and according to users’ recent interest. Similarity, index is measured by using Pearson correlation, Cosine based similarity and Euclidean distance based similarity. The design challenges and choice guidelines are also discussed.