Product Recommendation System Using Tunicate Swarm Magnetic Optimization Algorithm-Based Black Hole Renyi Entropy Fuzzy Clustering and K-Nearest Neighbour

Abstract

The recommendation system takes the information related to the user profile or interest to suggest the user with convenient materials that the user is interested in. Most of the existing implicit methods find the user preferences and automatically recommend the desired products in the interface, but failed to generate user-oriented results. Hence, an effective product recommendation method is developed in this research using the proposed Tunicate Swarm Magnetic Optimisation Algorithm-based Black Hole Renyi Entropy Fuzzy Clustering+K-Nearest Neighbour (TSMOA-based BHrEFC+KNN) to generate more user convenient result by grouping relevant products and recommends the similar products to users with great interest. The proposed TSMOA is designed by integrating the Tunicate Swarm Algorithm (TSA) and Magnetic Optimisation Algorithm (MOA), respectively. With the entropy measure and Jaro–Winkler distance, the process of group matching and the matching sequence of visitor and query are performed more effectively that enable to achieve the sentiment classification based on the binary visitor sequence. The performance obtained by the proposed TSMOA-based BHrEFC+KNN is evaluated in terms of accuracy, True Positive Rate (TPR), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE). The performance of the proposed system is compared with the existing methods, such as Deep learning+Naïve–Bayes, K-medoids clustering+Long Short-Term Memory (LSTM), BHEFC+Support Vector Machine (SVM) and TSMOA-Black Hole Entropy Fuzzy Clustering (BHEFC)[Formula: see text]Neural Network (NN), in which the TSMOA-BHEFC+NN obtained better results. The proposed TSMOA-based BHrEFC+KNN is 0.32%, 0.29%, 50%, and 6.44% is better than the existing TSMOA-BHEFC+NN in terms of accuracy, TPR, MAPE, and RMSE, respectively.