You are interested in fitting a logistic regression model for a binary classification application by maximizing the likelihood. You used the function in R and receive warning messages suggesting the fitting algorithm has crashed. What could be a reason why the method crashed? The classes could be linearly separable You forgot to specify the initial guess to the likelihood noise You should have used linear model for a binary classification problem The logistic regression likelihood surface is multimodal