On air targets recognition based on probability support vector machines

For the problem of standard support vector machines do not provide posteriori probability that needed in many uncertain classification problems, a modeling method of probability support vector machines based on cross entropy is proposed, and the method of determining model parameters is given in detail. on this base, the multi-calss support vector machines probability model is built and the probability model of sample belongs to calss in multi-class classification is given. A great deal of experiments show that the posteriori probability support vector machines model is reasonable and effective in air target recognition.