Wavelet Enhanced Adaptive Frequency Filter for Sequential Recommendation
Huayang Xu, Huanhuan Yuan, Guanfeng Liu, Junhua Fang, Lei Zhao, Pengpeng Zhao · 2025
Read Wavelet Enhanced Adaptive Frequency Filter for Sequential Recommendation by Huayang Xu, Huanhuan Yuan, Guanfeng Liu, Junhua Fang, Lei Zhao, Pengpeng Zhao (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-07028.txt.
Description
Huayang Xu and five co-authors improve sequential recommendation, the task of predicting a user's next action from their history. Users' preferences mix several periodic patterns that are hard to separate in the time domain, so recent systems analyze behavior in the frequency domain. Current frequency filters are mostly static, ignoring that patterns are personal, and global Fourier analysis blurs short-term, non-stationary changes.
Their method pairs two modules. A dynamic frequency filter adjusts its filtering to each behavioral sequence, extracting personalized global structure, while a wavelet enhancement module reconstructs the sequence to restore short-term fluctuations the Fourier view smears. Together they target both comprehensive modeling and efficiency on long histories. Tests on four widely used benchmark data sets showed the combined method superior to previous methods.
Catalog record
| CREATOR | Huayang Xu, Huanhuan Yuan, Guanfeng Liu, Junhua Fang, Lei Zhao, Pengpeng Zhao |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-07028