Programming Edit Recommendation Framework Based on View Histories and Big-Data Framework

Abstract

Nowadays the recommender system has grown in social media, mobile devices, personal use, and etc. on the internet every site has used the recommender system to attract the users or improve the site uses. But the existing recommender system has less accuracy, less recommendation speed and insufficient support to the current environment. To solve this problem, we propose the Hybrid MI technique which is based on existing MI technique. The existing MI technique is insufficient in speed, accuracy, and support, to solve this issue we proposed Hybrid MI technique. The existing methods failed to achieve accuracy, flexibility and early recommendations. These problems are overcome by a recently presented method called MI which is recommendation system extending ROSE. But the limitation of MI is that no end user satisfaction is taken into the considerations, and hence there is always scope for improvement in accuracy. In this project we are presenting HMI (Hybrid MI) technique in which we are improving the accuracy by relevance feedback method, in which log of feedbacks should be maintained and based on end users feedbacks, the proposed system can refine and regenerate more accurate recommendations next time for the same query with less time.