Machine Learning is against my better nature. Traditional statistical modeling has it’s own problem that it does not infer causality, as in which is the dependent and which is the independent variable? But to me the difficulty in processing machine learning is that it provides no straightforward formulaic relationship at all, it’s a black box. Don’t know why it works, you’re just happy that it does.
And yet ML’s lack of ‘discipline’ might be it’s most interesting quality. A second question is which do you best use for ML, R or Python? Might have to be both.
