A Comparative Analysis of Modeling Paradigms and Techniques in Sequential Recommendation

DOI

10.2991/978-94-6463-538-6_39How to use a DOI?

Keywords

Sequential Recommendation; Transformers; State Space Models (SSM); Attention Mechanisms; Recurrent Neural Networks (RNN)

Abstract

The evolution of sequential recommendation systems is marked by significant advancements, particularly in capturing user preferences from sequential data. Transformer-based models, despite their success, grapple with efficiency in processing long sequences. This study conducts a thorough comparative analysis of these models, highlighting the tradeoffs between efficiency and effectiveness across various modeling approaches and techniques. Research is categorized into Transformer-based, RNN-based, and the nascent SSM paradigm, examining their capacity for modeling complex sequences and their computational demands. Selective State Space Models receive special focus due to their potential in achieving a balance between performance and speed. The discussion progresses to specific sequential modeling techniques like attention mechanisms, memory architectures, and methods for managing long-term dependencies. The application of these techniques within different frameworks and their effects on performance and efficiency are analyzed. Additionally, the study evaluates how user data types (explicit and implicit) and recommendation tasks shape model development. It also considers dataset attributes like sequence length and sparsity, and their influence on the complexity and efficiency of models.

Copyright

© 2024 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Yuxiang Jia
PY  - 2024
DA  - 2024/10/01
TI  - A Comparative Analysis of Modeling Paradigms and Techniques in Sequential Recommendation
BT  - Proceedings of the 4th International Conference on Economic Development and Business Culture (ICEDBC 2024)
PB  - Atlantis Press
SP  - 329
EP  - 335
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6463-538-6_39
DO  - 10.2991/978-94-6463-538-6_39
ID  - Jia2024
ER  -