A zero altered Poisson random forest model for genomic-enabled prediction

In genomic selection choosing the statistical machine learning model is of paramount importance. In this paper, we present an application of a zero altered random forest model with two versions (ZAP_RF and ZAPC_RF) to deal with excess zeros in count response variables. The proposed model was compared with the conventional random forest (RF) model and with the conventional Generalized Poisson Ridge regression (GPR) using two real datasets, and we found that, in terms of prediction performance, the proposed zero inflated random forest model outperformed the conventional RF and GPR models.

Automated summary

Choosing the right statistical machine learning model is crucial in genomic selection. This study introduces a zero-inflated random forest model, which outperforms conventional random forest and Generalized Poisson Ridge regression models in prediction performance when dealing with excessive zeros in count response variables.