PyTorch and TensorFlow Performance Evaluation in Big data Recommendation System.

Over the past decade, there has been a renewed interest in Artificial Intelligence and its diverse frameworks. This research proposed a big data recommendation system that leverages two prominent deep learning frameworks which are PyTorch and TensorFlow to enhance collaborative filtering (CF) outcomes. The goal is to recommend more relevant products to users by incorporating textual review comments and combining various attributes. The study explores two recommendation system approaches within each framework. In the PyTorch framework, two collaborative filtering models were developed: one without incorporating user review text and the other with the inclusion of text from user reviews. Similarly, for the TensorFlow framework, two recommendation models were created: one without utilizing user review text, and the other with the integration of text reviews. The dataset, sourced from the Amazon website, comprised over 600,000 ratings and reviews. The outcomes showed significant enhancements by employing the proposed text-based method in both PyTorch and TensorFlow frameworks and addressed challenges such as dynamic preferences and data sparsity. Finally, the study provides a comparative analysis of the advantages and drawbacks of each utilized framework. [ABSTRACT FROM AUTHOR]