Movie Rating Prediction System Based on Opinion Mining and Artificial Neural Networks

Abstract

The paper presents an artificial intelligence based approach for movie rating prediction which can be useful for producers, distributors and viewer. In the proposed system, six different parameters and along with their individual ratings are used as features to train an artificial neural network. The algorithm used to train the neural network is the Levenberg-Marquardt back propagation algorithm. It can be observed that the proposed algorithm attains steep descent in prediction errors due to the mechanism of back propagation. As a standard convention, 70% of the data has been used for training and 30% data has been used for testing. The data set contains of 150 movies and the data is collected from IMDB. The proposed system achieves an accuracy of 97.33% and a sensitivity of 98.63%. An overall regression of 0.9182 has been attained exhibiting the fact that actual and predicted data bear high resemblance. It has been shown that the proposed technique achieves higher accuracy compared to contemporary techniques.