Robust statistical regression is a regression which is insensitive to so called excess observations, due to measurement errors or non-typical observations. Various coefficients selection criteria - such as minimization of the weighted sum of squared deviations, minimization of the sum of residuals absolute values, minimization of the median of squared residuals - are used to determine the robust regression equation for discrete data. Similarly, in the case of regression with interval data various methods are used to determine the robust regression equation, e.g. the multi criteria programming. In the paper a method of constructing a robust linear regression for interval data is proposed, which makes use of the multi criteria programming, in which the median criterion is proposed as the robustness criterion. We propose a procedure for arriving at the final interval regression (using various models) and apply this procedure to constructing an interval regression model for electricity load in a Polish city.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.