A LightGBM-XGBoost quantile regression model predicts CI build memory needs, delivering 36 GB average savings per build with under-allocation below 0.3% in production.
https://github.com/SAP/task-execution-data-set
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Intelligent resource prediction for SAP HANA continuous integration build workloads
A LightGBM-XGBoost quantile regression model predicts CI build memory needs, delivering 36 GB average savings per build with under-allocation below 0.3% in production.