search for: modelrob

Hey there. I am running a multiple regression and want to see whether a robust regression would provide different results, since there is some heteroscedasticity and other minor violations of regression assumptions. When I run the model using the robustbase package, > modelrob=lmrob(m1) I receive the following error: > Error in qr.default(x * sqrt(ret$weights)) : > NA/NaN/Inf in foreign function call (arg 1) Google wasn't helpful, maybe you can help me out. If this problem is not obvious to someone who uses R regularly, please let me know what further inform...