Application Of SVC In The Quantitative Selection

Recently, quantitative trading of stocks is popular in the country.The stock market of China will inevitably have good prospects for development. There is great potential for the development of quantitative trading. But it has some shortcomings at present, such as small total size, single strategies and differentiated performance. Therefore, it is essential to create new models and develop new ideas.SVM has a lot of advantages. Firstly, it can solve the problem under the condition of limited samples. Secondly, it can avoid the defects of local extreme value. Lastly, it can effectively solve the complexity of high-risk situations. Since classification techniques have the advantages of nice generalization ability and effective mapping capability, SVM can be used as a new method for the quantitative trading.The basic idea of SVC is transforming the classification in low-dimensional space into the high by using appropriate kernel function. It is aimed at the test set being divided linearly. It gives a brief introduction of the MR of the LDA. Because of the limitation of this approach, the paper introduces another method called Cross-Validation when evaluate the effect of SVC.In this paper, I classify stocks from the Shanghai A shares by using clas-sified techniques (SVC) and get quite good results. It takes several steps in the process. First of all, it takes preprocessing and standardization of the data collected in advance by specific methods. Then, it takes the process of feature variable selection by using principal component analysis method. The next, it classifies the data by SVC technology. And then, it selects stocks to construct portfolio according to the result of classification. Finally, the paper compares the rate of return on the stock portfolio and the market. It proves that the result of the comparison is quite good and it is advisable to select stocks by the SVM technologies.