Efficient gradient boosting for prognostic biomarker discovery

Abstract

Motivation: Gradient boosting decision tree (GBDT) is a powerful ensemble machine learning method that has the potential to accelerate biomarker discovery from high-dimensional molecular data. Recent algorithmic advances, such as Extreme Gradient Boosting (XGB) and Light Gradient Boosting (LGB), have rendered the GBDT training more efficient, scalable and accurate. These modern techniques, however, have not yet been widely adopted in biomarkers discovery based on patient survival data, which are key clinical outcomes or endpoints in cancer studies. Results: In this paper, we present a new R package Xsurv as an integrated solution which ap-plies two modern GBDT training framework namely, XGB and LGB, for the modeling of censored survival outcomes. Based on a comprehensive set of simulations, we benchmark the new approaches against traditional methods including the stepwise Cox regression model and the original gradient boosting function implemented in the package gbm. We also demonstrate the application of Xsurv in analyzing a melanoma methylation dataset. Together, these results suggest that Xsurv is a useful and computationally viable tool for screening a large number of prognostic candidate biomarkers, which may facilitate cancer translational and clinical research. Availability: Xsurv is freely available as an R package at: https://github.com/topycyao/Xsurv

Machine-Learning-Based Uplink Throughput Prediction from Physical Layer Measurements

“…After acquiring the residual, it combines them with the input of the existing model. This iterative process pushes the model in the direction of the actual value and improves the overall model prediction [31]. The GBR is represented in Algorithm 2.…”