Sequential recommendation with long-short-term preference and side information

Key Points

• The model significantly improves sequential recommendation accuracy by incorporating user behavior over time, enhancing personalization.
• Key performance metrics show a 20% increase in recommendation relevance when using side information during the recommendation process.
• This observational analysis utilizes collaborative filtering techniques alongside side information to evaluate user preferences over time.
• The findings suggest that integrating long-short-term preferences may optimize recommendation systems for diverse applications.