Applied research of the algorithm combined of PCA and SVMS on stock features

For the problem of feature selection of stock, this paper presents a new algorithm which is the optimal combination of Principle Components Analysis (PCA) with Support Vector Machines (SVMs). The new algorithm is based on weight measure. Because of specialty of this problem, a weight measure is learned by PCA and SVMs with linear kernel function. Good stock and bad stock with many features belong to two classifications. Experiments prove the effective of our method compared with traditional feature selection.