I've been successful in using the relaimpo-Package for R in SPSS through STATS_RELIMP to calculate the Importances of different predictors (in cases of multicollinearity). What im wondering now is how I can use the results from Shapley Value Regression to predict the dependent variable. I'm sure that I can't just use the resulting importances in a linear combination as I would the b-values in a linear regression. Does anyone know how to predict with Shapley Value Regression results?

  • 1
    $\begingroup$ This is a little hard for me to follow. Can you provide a concrete example? What is your situation? What are your data? What are you trying to achieve? Can you paste in the output? $\endgroup$ – gung - Reinstate Monica May 17 '16 at 15:09
  • $\begingroup$ So I have for example asked in a survey the propensity for switching to a different internet provider. I have several variable which explain this propensity, such as image items or satisfaction with different aspects of the service. I would now like to predict the propensity to switch from the image items and the satisfaction items for other customers that I don't have the propensity yet. You could calculate a normal linear regression, but since there is a lot of multicollinearity I used Shapley value to calculate the importances. The Question is now how to use the results for prediction. $\endgroup$ – chuelibrueder May 18 '16 at 6:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.