A contextual information based scholary paper recommender system using big data platform

Abstract

Recommender systems for research papers have been increasingly popular.In the past 14 years more than 170 research papers,patents and webpageshave been published in this field.Scientific papers recommender systemsare trying to provide some recommendations to each user which are consistent with the users' personal interests based on performance, personal tastes and users behaviors.Since the volume of papers are growing day after day and the recommender systemshave not the ability for covering these huge volumes ofprocessing papers according to the users' preferences it is necessary to use parallel processing (mappingreducing programming) for covering and fast processing of these volumes of papers. The suggested system for this research constitutes a profile for each paper which contains context information and the scope of paper. Then, the system will advise some papers to the user according to the user work domain and the papers domain. For implementing the system it has been used hadoop bed and the parallel programming because the volume of data was a part of a big data and the time was also an important factor. The performance of the suggested system was measured by the criteria such as user satisfaction and the accuracy and the results have been satisfactory.