Ensembling a Linear Eegression model with an error Mitigation Component

Abstract

This �a�er �resents a �ro�osal o� a model error mi�ga-�on te�hni��e based on the error distrib��on anal�sis o� the original model and �rea�ng the addi�onal model that tem�ers the error im�a�t in �ar���lar domain areas iden-��ed as the most sensi�ve. �oth models are then �ombined into single ensemble model. The idea is demonstrated on the trivial two-dimensional linear regression model.