A Multitask Learning framework for E‐commerce time series analytics: Operational risk, demand forecasting, and anomaly detection
Junzhong He, Xiaorui An · 2026
Read A Multitask Learning framework for E‐commerce time series analytics: Operational risk, demand forecasting, and anomaly detection by Junzhong He, Xiaorui An (2026) 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/plos-pone-0352809.txt.
Description
He and An built a multi-task learning system for online trade data in a 2026 study. Cross-border sellers must forecast demand, judge running risk, and spot anomalies at once. Their model shares one feature reader across tasks, with focused attention shaped in a pyramid, then splits into task-specific heads.
Tests on a global trade data set and a network traffic set split data into training, checking, and test parts. The system cut forecasting error and raised anomaly detection scores versus a strong video-style baseline. Sharing one reader across tasks mirrors how trade signals overlap, and it lowers computing cost for sellers watching many markets.
Catalog record
| CREATOR | Junzhong He, Xiaorui An |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49735 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |