Hi, Is there anyone have some experience in the boost package. I used the logitboost function to do the two-class classification. This is the format I used. logitboost(xtrain, ytrain, xtest, presel = 2, mfinal = 100) # with feature selection logitboost(xtrain, ytrain, xtest, presel = 0, mfinal = 100) # without feature selection xtrain is n*p, n is sample size, p is dimension The error is always like this Error in model.frame.default(formula = y ~ bx, weights = w/mean(w), na.action = function (x) : invalid type (list) for variable 'bx' I tried to implement by copying the original code of this function to track down but still I don't know what's the reason. Could somebody help me with that? I greatly appreciate your help! sincerely, Linda _________________________________________________________________ Refresh_family_safety_052008 [[alternative HTML version deleted]]