Target: Implement the retrieval system by Vector Space Model with Rocchio Relevance Feedback (pseudo version)
https://www.kaggle.com/c/wm-2020-vsm-model/
| No Feedback | Feedback once | |
|---|---|---|
| cosine similarity | 0.73228 | 0.77450 |
| bm25 rsv | 0.78876 | 0.79027 |
Target： Implement the personalized recommendation with implicit feedback
https://www.kaggle.com/c/wm-2020-personalized-recommendation
Run BPR
bash run.sh [path of your output(csv) ranking list]