Statistics in practical decision making

Abstract

This issue is devoted to statistics applied to decision making problems. The issue begins from consideration of methodological possibilities to construct regressions with meaningful and interpretable coefficients for individual predictors and shares of their importance. Ewa Nowakowska explores properties of the so-called Shapley value regression developed for adjusting regression coefficients with multicollinearity among the predictors and estimating their importance. D. Liakhovitski, Y. Bryukhov, and M. Conklin compare the approaches of random forests and relative weights of the predictors in regressions under various scenarios. Together with M. Conklin we investigate further features of Shapley value regression and its predicting abilities. The next several works present statistical methods for solving different practical decision making problems. Dr. B. Vilge considers possibilities of regression predictions for chemically active materials by physical and chemical characteristics measured by non-destructive testing. Prof. J. Subramani describes how proper and meaningful decisions should be taken in modern manufacturing process control. Profs. S. Saxena, H.P. Singh, O.K. Gupta, and K.S. Rao share their experience on Bayesian assessment of failure and repair times for systems and components helping to reliability engineers to take correct decisions. And Prof. E. Demidenko shows how to choose the optimal portfolio to reach the investors’ financial goals. And continuing already established tradition of giving topic quotations, here are some more of them on decision making.