IntentRec: An Advanced Recommender System Leveraging User-Item Intent

1. Introduction

The proposed IntnetRec is a neural recommendation algorithm. The proposed model consists of 1) the Intent layer and 2) the Kernel layer. Achieving state-of-the-art performance on the Movielens-100K and Douban datasets, beating all the existing models.

2. Requirements

• NumPy
• PyTorch

3. Code execution commands

1. ./run.sh file_dir_path_of(dataset) config_file(config.json)

4. RMSE Results

• Douban: 0.7208
• ML-100K: 0.8868
• ML-1M: 0.823

4. Data References

1. Harper, F. M., & Konstan, J. A. (2015). The movie lens datasets: History and context. Acm transactions on interactive intelligent systems (tiis), 5(4), 1-19.