Learning persona-driven personalized sentimental representation for review-based recommendation

A large amount of information exists in many e-commerce and review websites as a valuable source for recommender systems. Recent solutions focus on exploring the correlation between sentiment and textual reviews in the review-based recommendation. However, these studies usually pay less attention to the differences of different users in sentimental expression styles or language usage habits when a user writes reviews. In this work, we argue that the individual reviewing behavior is closely related to personality, and sentimental expression is a manifestation of personality. Therefore, we propose a novel Persona-driven Sentimental Attentive Recommendation model (named PSAR) via personalized sentimental interactive representation learning for the review-based recommendation. The proposed model is devised to learn fragment-level and sequence-level personalized sentimental representation simultaneously from reviews. Besides, an attentive persona-driven interaction module is designed to capture word-level usage habits and sentence-level analogous tones. Comprehensive experimental results on four real-world datasets demonstrate that our model outperforms the state-of-the-art methods.

Automated summary

This study proposes a review-based recommendation model based on personalized sentimental interactive representation learning. The model simultaneously learns fragment-level and sequence-level personalized sentimental representations to capture differences in users' sentimental expression styles and language usage habits.