Li Jin

Personalized recommender systems play an essential role in enhancing user experience by accurately predicting user preferences. Previous approaches mainly focus on modeling long-term preferences or capturing short-term dynamics through sequential patterns, while few achieve an effective balance between the two. This study proposes Rec-SSP, a novel review-aware recommendation model that integrates long-term and short-term preferences through a gated fusion mechanism. Long-term preferences are extracted from aggregated user reviews, whereas short-term preferences are modeled by identifying sequential patterns from recent interactions at both the review and category levels. This multilevel design captures fine-grained opinions across items, ensuring a more accurate understanding of the evolving user intent. This study conducted various experiments on real-world datasets, showing that Rec-SSP outperforms baseline models. These findings demonstrate that balancing long-term and short-term preferences with multilevel sequence modeling can significantly improve recommendation accuracy across diverse domains.

Online reviews are crucial when building a recommendation model because they contain the specific and rich preferences of users related to different aspects of a particular item. Incorporating these online reviews into the recommendation model mitigates the data sparsity issue to some extent and contributes to better recommendation performance. Despite this success, review-based recommender systems have the limitation that they do not fully consider the relevance of the review text to the target item. Specifically, the review text should reflect the user’s detailed opinion about the target item to extract detailed preference information. Meanwhile, the review content must be directly related to the target item to extract the customer’s specific preferences for the item. However, previous studies have overlooked both of these aspects. Therefore, it is necessary to build a recommendation model that considers the relevance of the review content to the target item. To address this issue, this study proposes a novel recommendation model that accurately estimates users’ preferences by carefully considering the relevance of the review content to the items. The proposed model effectively extracts feature representations from the text using bidirectional encoder representations from a transformer and obtains fused features by considering the importance of features through the attention mechanism. To evaluate the performance of the model, we used a publicly accessible dataset of reviews from Amazon.com and compared it to various baseline models. The experimental results demonstrated better recommendation performance of the proposed model compared to other baseline models.