The Theory Of Support Vector Machine And Its Application In Finance

People learn from everything based on classification ,many excerllent learn-ing algorithms are all based on the foundations of classification. Recently a lot ofmethods are applied on pattern classification,including traditional methods,forexample, Bayesian method, distance-distinguish, Fishier distinguish,etc. Recentmethods such as fuzzy classification,coarse classification and net works,as well assupport vector machine.The theory of support vector machine is a professional theory of solving non-linear classification and regression problems which avoid the hard establishing ofthe net structure and the problem of over-learning, under-learning and local min-imum.It is known as the best method of solving the classification and regressionof using a small amount of sample. The support vector machine(SVM) methodwhich based on the theory of support vector machine proposed recently,providesa new thought of solving nonlinear problem.Firstly this paper goes back on the achievements which the previous peoplehave done to support vector machine.Based on the achievements, we introducethe classification algorithm and the regression algorithm based on SVM in theframe of quadric programming,as well as the classification algorithm and the re-gression algorithm in the frame of linear programming. And we also do somefurther study to the theory of support vector machine.We give the proof of heequivalence between the algorithms of classification and regression on the non-linear condition.Mean while ,we use this theory on the study of financial early-warning of List company and give the model of financial early-warning of Listcompany based on SVM.Then we present a demonstration of financial early-warning of Listed company with some data and the results we get shows thatthe SVM methods is better than other methods such as net works.