I am defining some conditional probability tables based on simple regression equations, but I would like to reflect the relative fit of the regression line in the resulting tables. For example, a regression with an r2=0.4 should ideally generate a "noisier" table than a regression with an r2=0.8, for the same equation. Is there a function I can use in the equation editor to accomplish this?I know this is an odd approach, but my regressions are based on only 5-10 points and don't cover the range of possible outcomes I need to model, so my data are too sparse to use parameter estimation to generate tables directly.thanks
Quote 0 0