SVMs

Support Vector Machines are a supervised learning method developed by Vladimir Vapnik and his colleagues, with the modern formulation published in September 1995. They work by finding the boundary that best separates two classes of data while keeping the widest possible margin between them. The handful of data points that sit closest to that boundary, the support vectors, give the method its name. Because the boundary depends only on those points, SVMs generalized well even with limited data, and they delivered the best accuracy of their time on high dimensional problems such as handwriting…