Agreed. Deep learning receives all the hype these days (look at the machine learning subreddit), but in reality it provides very little practical value for the

Agreed. Deep learning receives all the hype these days (look at the machine learning subreddit), but in reality it provides very little practical value for the vast majority of data mining tasks that most folks will see. For that, standard methods like you mentioned (and don't forget SVM of course!) are still the go-to methods in my toolbox.