基於特徵組合的使用者偏好預測

學習使用者偏好有助於了解使用者和做出適當推薦。過去的研究利用使用者的個人資料或使用者的公開評論作為偏好分數預測的依據，但使用個人資料有隱私疑慮，也並非所有場景下都有用戶評論資料；相較之下，被推薦物品的資訊和使用者歷史行為更容易取得。此外，在許多場景下，被推薦物品的資訊是標籤類型的特徵。我們的研究目標是針對前述情境，利用更細維度的特徵組合和使用者的歷史行為做出偏好預測。我們提出了建模非結構性特徵組合的模型，對於未曾在歷史資料出過的特徵組合也能有效的加以預測，並提供實驗來驗證模型的合理性。

關鍵字

推薦系統；稀疏特徵；神經網路；深度學習；偏好學習

並列摘要

Learning user preferences is critical in understanding users and recommending items. Previous works utilized user profiles or public reviews for user preference modeling. However, utilizing user profiles might cause privacy issues, and there are scenarios where items do not come with user reviews. In contrast, item metadata and user historical behaviors are easier to obtain. Moreover, item metadata are tag-formed features in many real-world scenarios. We aim to focus on the described scenario and utilize finer-grained features of the item and user historical behaviors to predict user preference. We proposed a model for unstructured feature combinations, which can handle unknown items, and provide experiments to justify the rationality of our model.

並列關鍵字

recommender system ； sparse feature ； neural networks ； deep learning ； preference learning

參考文獻

[1] Yelp open dataset. Available at https://www.yelp.com/dataset.

Google Scholar

[2] J. L. Ba, J. R. Kiros, and G. E. Hinton. Layer normalization, 2016.

Google Scholar

[3] K. Bi, Q. Ai, and W. B. Croft. Learning a finegrained reviewbased transformer model for personalized product search. In Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 123–132, 2021.

Google Scholar

[4] T. Bogers. Tagbased recommendation. In Social Information Access, pages 441–479. Springer, 2018.

Google Scholar

[5] Y.W. Chang, C.J. Hsieh, K.W. Chang, M. Ringgaard, and C.J. Lin. Training and testing low degree polynomial data mappings via linear svm. The Journal of Machine Learning Research, 11:1471–1490, 04 2010.

Google Scholar

主題瀏覽