Fig. 5From: Discovering optimal weights in weighted-scoring stock-picking models: a mixture design approachThe coefficient of this quadratic term is four times the difference between the central-point regression estimate of the side and the average value of regression estimates from the two apexes of this sideBack to article page