Multiplier maximum entropy algorithm of support vector machines

For small sample recognition problems, a proximal algorithm of support vector machine, called the multiplier entropy algorithm, is proposed in this paper. The algorithm combines the virtues of both multiplier algorithm and entropy algorithm. It not only can turn non-smooth problems into smooth ones, but also can reduce the iteration in some degree and avoid the morbid state of Hessian. For small sample problems, especially the pre-cancer diagnosis, the multiplier entropy algorithm demonstrates effective performance.