Iterative Learning in Support Vector Regression With Heterogeneous Variances

The presence of heterogeneous variances is the norm in practice, which makes machine learning predictions less reliable when noise are implicitly assumed to be equal. To this end, we extend support vector regression by allowing a range variance functions model training. Specifically, as function mean and other variables traditionally used statistical modeling. This leads iterative between training Extensive simulations implemented validate effectiveness proposed framework both linear nonlinear regressions. Finally, two real data sets demonstrate superiority algorithm heterogeneity.