The initial purpose of this paper was to expose various mitigations to fit models to observed data with heteroscedatic behaviour. The research target evolved in the last months, and another issue resulting in model misfit appeared. This issue is about the multiplicity of the poles of the underlaying model. This issue will be discussed and mitigations presented.
Dupuis, P., Van Overstraeten, N., Raskin, J.-P., Francis, L., & Flandre, D. (2012). Some mitigations for unequal data variance in linear regression. In Advanced Mathematical and Computational Tools in Metrology and Testing IX (p. pp. 118-125). World Scientific Publishing Vompany. https://doi.org/10.1142/9789814397957_0015