Publication: User Emotion Direction for Recommendation Systems-A Decade Review
Issued Date
2021-01-01
Resource Type
Other identifier(s)
2-s2.0-85123222252
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
2021 26th International Conference on Automation and Computing: System Intelligence through Automation and Computing, ICAC 2021. (2021)
Suggested Citation
Suphitcha Chanrueang, Sotarat Thammaboosadee, Keng Goh, Hongnian Yu User Emotion Direction for Recommendation Systems-A Decade Review. 2021 26th International Conference on Automation and Computing: System Intelligence through Automation and Computing, ICAC 2021. (2021). doi:10.23919/ICAC50006.2021.9594209 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/76719
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
User Emotion Direction for Recommendation Systems-A Decade Review
Other Contributor(s)
Abstract
Recommendation systems are rapidly gaining popularity in software development, including e-commerce, news, advertising, social networking, and entertainment. It filters appropriate information for user decisions. The most popular approaches of recommendation systems are content-based and collaborative filtering-based, which are created as a model by using user preferences and providing recommendations. Additionally, the hybrid approach is proposed to improve recommendations typically by combining the advantages of two techniques to increase efficiency and prediction performance. However, general recommendation systems typically abandon users' contextual preferences such as culture, emotions, and other details in different situations. Researchers are attempting to apply knowledge from other scientific fields to improve the performance of their recommendation systems. Psychology is one of approaches that can be applied to understand and explain humanity and shows that emotions influence decision-driven, efficient, and predictable. This paper reviews relevant research analyzes state-of-arts, gaps, and further recommendation system research based on emotion. We find that most of the selected research use sentiment data extracted from open data sources and social networks. As for the data extraction and data analysis depend on data sciences and statistics theory, and Cold start is still a challenge for researchers. However, we find that the data from social media reaction can compare with the emotional wheel in psychology and present emotion as more complex than sentiment. Future research on an individual recommendation system will bring the complexity of psychological emotion into improving the system.