Optimal hierarchical attention network-based sentiment analysis for movie recommendation

Abstract

In recent days, recommender systems are extremely widespread applications for handling information overload issues by offering personalized recommendations, and also, the researchers began to utilize the sentiment analysis-based movie recommendation system. This paper devises a novel technique for sentimental classification, which helps to recommend the positive reviewed movies. Here, the water cycle earthworm Optimization (WCEWO) is newly devised by incorporating water cycle algorithm and earthworm optimization algorithm. The procedure is started by providing the input to the matrix calculation stage, which then builds the matrix based on the user's preferences. Here, the service grouping is performed using ROCK algorithm. Thereafter, the group matching is performed by Cosine similarity for achieving best group from all other groups. Once the group matching is done, the matching is carried out between the query and the visitor binary sequence using Kendall Tau distance measure for identifying best visitor sequence. Finally, sentimental classification is done using hierarchical attention networks (HAN), where the training process of HAN is performed using WCEWO algorithm and hence the suitable movie recommendation is achieved by offering appropriate positive reviewed movie recommendations to the users. Mean square error (MSE), root-mean-square error (RMSE), and accuracy are used as performance indicators to assess the effectiveness of the established approach. The created WCEWO-based HAN gives maximum accuracy of 89.81%, minimum MSE of 0.185, and minimum RMSE of 0.430.