Anonymous
Both the approaches are linear classification approaches. First logistic regression utilizes a sigmoid function to plot the data onto the sigmoid curve and maximizes likelihood of the input. SVM, on the other hand, utilizes support vectors or vectors which help create a margin which classifies the two classes very well. This, along with kernel trick, the trick to add non-linearity to modeling using a polynomial or rbf kernel without converting the data into the space makes it more robust.t Both the models are binary classifiers and relatively fast to train