Capturing Popularity Trends: A Simplistic Non-Personalized Approach for Enhanced Item Recommendation

By Jiazheng Jing et al

Read the original document by opening this link in a new tab.

Table of Contents

1. Introduction
2. Related Works
3. Popularity-Aware Recommender
3.1 Problem Definition
3.2 Model Architecture
3.2.1 Popularity History Module
3.2.2 Temporal Impact Module
3.2.3 Periodic Impact Module
3.2.4 Side Information Module
3.2.5 Fusion Module
4. Experiments
5. Conclusion
6. References

Summary

The document discusses the importance of capturing item popularity trends in recommendation systems. It introduces a Popularity-Aware Recommender (PARE) model that focuses on non-personalized recommendations by predicting items with the highest popularity. PARE consists of four modules - Popularity History, Temporal Impact, Periodic Impact, and Side Information - each contributing to predicting item popularity. The model utilizes an attention layer in the Fusion Module to combine the predictions from the four modules. Experimental results show the effectiveness of PARE in enhancing recommendation accuracy, outperforming state-of-the-art methods. The simplicity of PARE makes it a valuable baseline for future research and practical applications in the industry.