Boosting Models for Performance Prediction in Highly Configurable Software Systems
Jailma Januario · Juliana Pereira
Abstract
In this work, we present a preliminary study using boosting models applied to highly configurable software to predict performance variables and identify the most relevant parameters. Using a dataset from the x264 video encoding system, the CatBoost model demonstrated the best performance, achieving an R² score of 0.87.
Successful Page Load