The latent factor model (LFM) is a classic model in the field of personalized recommendation. However, there are two inevitable challenges in LFM-based recommendation methods. First, a sparse rating matrix restricts the recommendation performance of the LFM. Second, the LFM lacks explanatory ability. To address these deficiencies, we propose an item attribute-aware latent factor model (ILFM) for personalized recommendation, which assumes that explainable attributes of an item are key factors of the user’s rating of that item. The ILFM models the user’s interests using explainable attributes of items, which not only relieves the sparse rating matrix limitation but also gives the LFM intuitive explanatory ability. We validate the performance of the ILFM using two public datasets; the experimental results show that the ILFM not only outperforms state-of-art recommendation methods in terms of recommendation performance but also provides effective explanatory ability for the recommendations.
This work was partly supported by the National Natural Science Foundation of China (nos. 71871019, 71471016).
Kwon, O Chol; Gan, Mingxin; and Zhang, Xiongtao, "ILFM: Item Attribute-Aware Latent Factor Model for Personalized Recommendation" (2021). PACIS 2021 Proceedings. 199.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.