Improving E-Commerce Recommendation System with SOBPRA: A Semantic Ontology-Based Approach

The evolving world of e-commerce, there comes a vital urge for improving product recommendation systems by making their recommendations more accurate and personalized to enhance user satisfaction for better sales. In the current study, a novelty Semantic Ontology-Based Product Recommendation Algorithm, SOBPRA, is implemented that develops an advanced product recommendation system using an Amazon grocery dataset. The algorithm in use involves the design of elaborate ontologies for each product, including a wide variety of attributes, relationships, and semantic representations. Such ontologies are then processed, with the application of semantic learning techniques to gain an all-rounded understanding of products and user preference. It uses semantic similarity measures between the user profile ontology and ontologies of relevant products for highly matched product retrieval concerning both the explicit and inferred preferences of the user. The most important novelty of SOBPRA is its deep semantic processing capability, which enables the system to detect complex relationships and preferences beyond superficial data. This algorithm continuously learns from user interactions, improving ontologies to better provide more accurate recommendations by ranking organic product first. A list of personalized product recommendations sorted by an appropriateness score is generated.