If a dataset exhibits non$-$linear dependencies, the linear or logistic regression model, without any variable transformation, may produce lower accuracy, in comparison to other mining methods. is it true