Enhanced movie recommendation System using hybrid multi factor Filtering mechanism and opinion Mining

Abstract

With an increase in the number of multimedia technologies, newlinemovies, social media videos and the growth of OTT platforms, it confuses newlinethe users to decide, which one to watch for. In this scenario, movie newlinerecommendation systems are widely used. Different approaches are newlinefollowed for movie recommendations using multi-faceted features such newlineas content, behavior, time spent, frequency and so on. However, the newlineexisting methods do not produce efficient results in terms of movie newlinerecommendations. On the contrary, these methods introduce poor newlineaccuracy in recommending the movies to the potential users. The existing newlinemethods generate movie recommendations according to their content newlinefeatures, which introduces high irrelevancy. newlineThe existing methods do not consider multiple or all the features newlineinto account, when detecting a user s interest. This drawback introduces newlinea high false ratio in movie recommendations. The existing collaborative newlinemethods use only the rating values in identifying the movies to newlinerecommendation, which in turn brings high irrelevancy. newlineThese methods identify users with similar interests, according newlineto the visits made by them and does not consider their persistent interests newlineon movie classes. Further, it fails to track the sentiments of the users in newlinegenerating movie recommendations, which affects the performance of the newlineentire system. On the whole, the existing methods suffer from high false newlineratio, too much irrelevancy and high time complexity. newlinex newlineBased on the problems identified, the following research newlineobjectives are proposed in this research thesis to achieve high-performing newlineand accurate outcomes with less time complexity. newlineTo design an efficient movie recommendation scheme that newlineutilizes a user s rating in identifying the similar users so as to generate newlinerecommendations. The method should consider the maximum number of newlineavailable features and factors in generating the recommendations. newlineFurther, the method should also consider the attitude of the users in newlineidentifying similar users towards the movie recommendations.