Personalized E-commerce Recommendation System using Marine Predator Algorithm with Gated Recurrent Unit

Through the constant advancement of Internet expertise and the crucial development of the network environments, e-commerce has grown vastly. With respect to resolve the issue of data overload and loss in choosing of e-commerce online shopping area. A personalized Recommendation System (RS) utilizing data filtering advancement has become most efficient. Hence, this research proposes the Marine Predator Algorithm with Gated Recurrent Unit (MPA-GRU) approach for the personalized e-commerce RS. The standard datasets like Amazon Product review and Retail Rocket are used to estimate effectiveness of the MPA-GRU approach. The pre-processing steps like stemming, lemmatization and tokenization are used to enhance the quality of the data. The Term Frequency-Inverse Document Frequency (TF-IDF) technique is used to extract relevant features and then these data are provided to the recommendation process. The experimental result represents that the proposed MPA-GRU approach attains the better accuracy of 96.21% and 95.29% on both Amazon Product review and Retail Rocket datasets as compared to the GRU approach.