Penalised regression with multiple sources of prior effects

TL;DR

This paper introduces a penalised regression method that integrates multiple sources of prior information about feature effects, improving prediction accuracy in high-dimensional tasks, and is implemented in the R package `transreg'.

Contribution

It presents a novel approach for combining multiple prior data sources into penalised regression models, enhancing predictive performance.

Findings

Improved prediction accuracy with prior information integration

Method validated through simulations and real data applications

Available as an R package `transreg' for practical use

Abstract

In many high-dimensional prediction or classification tasks, complementary data on the features are available, e.g. prior biological knowledge on (epi)genetic markers. Here we consider tasks with numerical prior information that provide an insight into the importance (weight) and the direction (sign) of the feature effects, e.g. regression coefficients from previous studies. We propose an approach for integrating multiple sources of such prior information into penalised regression. If suitable co-data are available, this improves the predictive performance, as shown by simulation and application. The proposed method is implemented in the R package `transreg' (https://github.com/lcsb-bds/transreg).