Time Aggregation Features for XGBoost Models
Mykola Pinchuk, PhD · 2026
Read Time Aggregation Features for XGBoost Models by Mykola Pinchuk, PhD (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/arxiv-2601-10019.txt.
Description
Mykola Pinchuk adds time sums as inputs for a boost tree model. The model type is XGBoost. The task is to guess if a user will click an ad. The data set is the Avazu click set. Time sums pool past clicks by hour spans.
A strict rule guards the build. Sums for one hour can use only prior hours. No data from the same hour or later may leak in. That rule keeps the test fair, as in real use. The paper asks which spans help most. It also checks if the gains hold on a later time split, not just in the build span.
Catalog record
| CREATOR | Mykola Pinchuk, PhD |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 27 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-10019