Performance of Regression Models as a Function of Experiment Noise
Journal article, 2021
Background: A challenge in developing machine learning regression models is that it is difficult to know whether maximal performance has been reached on the test dataset, or whether further model improvement is possible. In biology, this problem is particularly pronounced as sample labels (response variables) are typically obtained through experiments and therefore have experiment noise associated with them. Such label noise puts a fundamental limit to the metrics of performance attainable by regression models on the test dataset. Results: We address this challenge by deriving an expected upper bound for the coefficient of determination (R2) for regression models when tested on the holdout dataset. This upper bound depends only on the noise associated with the response variable in a dataset as well as its variance. The upper bound estimate was validated via Monte Carlo simulations and then used as a tool to bootstrap performance of regression models trained on biological datasets, including protein sequence data, transcriptomic data, and genomic data. Conclusions: The new method for estimating upper bounds for model performance on test data should aid researchers in developing ML regression models that reach their maximum potential. Although we study biological datasets in this work, the new upper bound estimates will hold true for regression models from any research field or application area where response variables have associated noise.
machine learning
label noise
regression models
upper bound
experiment noise
Author
Gang Li
Chalmers, Biology and Biological Engineering, Systems and Synthetic Biology
Jan Zrimec
Chalmers, Biology and Biological Engineering, Systems and Synthetic Biology
Boyang Ji
Technical University of Denmark (DTU)
Chalmers, Biology and Biological Engineering, Systems and Synthetic Biology
Jun Geng
Chalmers, Biology and Biological Engineering, Systems and Synthetic Biology
Johan Larsbrink
Chalmers, Biology and Biological Engineering, Industrial Biotechnology
Aleksej Zelezniak
Science for Life Laboratory (SciLifeLab)
Chalmers, Biology and Biological Engineering, Systems and Synthetic Biology
Jens B Nielsen
BioInnovation Institute
Chalmers, Biology and Biological Engineering, Systems and Synthetic Biology
Technical University of Denmark (DTU)
Martin Engqvist
Chalmers, Biology and Biological Engineering, Systems and Synthetic Biology
Bioinformatics and Biology Insights
11779322 (eISSN)
Vol. 15Predictive and Accelerated Metabolic Engineering Network (PAcMEN)
European Commission (EC) (EC/H2020/722287), 2016-09-01 -- 2020-08-30.
Subject Categories
Bioinformatics (Computational Biology)
Probability Theory and Statistics
Control Engineering
DOI
10.1177/11779322211020315
PubMed
34262264