A Neural Engine for Movie Recommendation System

There are numerous number of movies available over the world, all of those are not interesting and also impossible to watch for one user. That’s why, a recommendation system is very important for user to find out the suitable product quickly. On the other hand, a recommendation system gives the flexibility of efficient searching rather than manually. In this way, recommendation system plays a prominent role to user. In this study, we have developed a scheme for a movie recommendation system named neural engine-based recommendation system (NERS) for users. In our recommended approach (NERS), we have incorporated data contents about user’s interests via standard movie dataset, that helps us to make a neural engine called neural recommender (NR). We have used two sorts of data sets to make NR, one is general dataset associated with five different nature of data variables, and another one was based on user’s choice pattern, where some of the volunteer user contributes their efforts to create it. After combining both data sets, NR engine was applying a neural network (NN), that’s recognize user behavioral patterns and then forming a class database, where each class have constructed by using movie genres. In this way, we have initiated nine different grades of classes in the manner of various genres. Finally, two evaluation techniques were used to figure out the best solutions by selecting one or multiple class. For multiple classes, our system will combine information from selected classes and consider them as one for query purpose. At last, three estimators, mean squareerr or(MSE), mean absolute error (MAE) and mean relative error (MRE), were exploiting to demonstrates prediction accuracy of our NERS approach. And, the simulation results show that, our system achieved better performance compare to other methods.