Other Information
Operations Research and Data Analysis
These pages present the results of our research of a technique for assessing the sensitivity of the optimal solution to a widely encountered problem in data analysis. The problem of estimating the parameters of the single equation linear regression model under the minimum sum of absolute errors (MSAE) criterion has a well known linear programming (LP) formulation. We have identified a methodology that allows the data analyst to assess the impact of unforeseen changes in the data used for formulating the LP problem (fitting the model). It is a sensitivity analysis of the technical or left-hand side (LHS) coefficients of the non-binding constraints of the LP formulation of the MSAE problem.
- Mathematical proof of the analysis
