User Interest Driven Temporal Collaborative Filtering for Product Recommendation

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Recommender systems, commonly employed in e-commerce, significantly aid clients in making informed decisions.The recommender system incorporates a wide variety of algorithms but collaborative filtering algorithm is regarded as among the most efficient technologies used in customized recommendation system.However, conventional algorithms merely take user ratings into account, failing to take into account shifts in user interest or the veracity of rating data, which had a significant impact on the system's suggestion quality.An improved approach is suggested in this study to deal with this problem.Initially, a gradual time drop is employed to give the user's rating a weight.Then, the current user's neighbors are chosen from a group of users who share common interests.Finally, the average assessments of his neighbors may indicate the current user's choice for a certain item.According to experimental data, the algorithm has the potential to upgrade the recommendation system's quality and prolong the accuracy of neighbor detection.