Recommendation Systemfor Daily Consumer Purchases List Usingspecial Hybrid Approach

Abstract

Recommendation systems have attained widespread prevalence in the current digital world, providing consumers with specific recommendations for a diverse range of products, services, and information. These systems have a significant role in shaping customer behavior, particularly in the realm of online shopping. This study aims to evaluate the performance of a hybrid recommendation system in suggesting daily purchase items to consumers by examining three different recommendation system methods: collaborative filtering (CF), Content-based filtering (C-B), and Hybrid. Then we assess their efficiency in delivering suggested items to consumers through the utilization of recall, precision, and F1 metrics. The study reveals that each strategy exhibited distinct strengths. However, the hybrid approach is considered the most effective method for recommending items for new users who do not have a history profile.