Quantile Regression Comprehensive in Machine Learning: A Review

The quantile regression (QR) approach is an improved variant of the least squares method. The study explains why quantile regression outperforms linear regression. It is frequently employed in a variety of fields, including economics, investing, finance, medical analysis, and recommendation systems. Unlike traditional mean regression, QR is more resistant to outliers and poorly defined error distributions, and it provides more comprehensive statistical modeling. It can describe the entire conditional distribution of the outcome variables. When normality assumptions are violated or outliers exist, QR models can be used not just to uncover the various effects of covariates on unique outcome magnitudes, but also to produce more complete and trustworthy estimates than mean regression. These advantages make QR appealing, and they include applications for risk computation, longitudinal data, time-to-event data and among others. Finally, an high-level overview is provided for the QR and its associated models and methodologies for diverse data kinds and application domains.