Robust Weighted Random Forest Regression with Alternative Bootstrap
Aylin Alın · 2024
Random Forest Method (RFM) introduced by Breiman (2001) is a highly data adaptive machine learning tool. It allows to consider both regression and classification problems, and is capable of handling large datasets, missing values in predictors, multi-collinearity problem or datasets where the sample size is much smaller than the number of predictors. It is applied in wide range of areas such as for predicting energy consumption (Ahmad et al. 2017), for automatic selection of molecular descriptors (Cano et al. 2017), for real time radar derived rainfall forecasting (Yu et al. 2017), for image classification (Mahapatra 2014, Xu et al. 2012), for modeling statistical arbitrage (Krauss et al. 2017), for diagnosing aviation turbulence (Williams 2014), for outcome prediction in antibody incompatible kidney transplantation (Shaikhina et al. 2019), for genetic data analysis (Chen and Ishwaran 2012, Díaz-Uriarte and Alvarez de Andres 2006), for the predictive validitý of the items from 2 mindfulness instruments (Sauer et al. 2015), and for ecological prediction (Prasad et al. 2006).