论文部分内容阅读
要想正确预测公司财务业绩,首先必须选择合适的预测方法。现有文献所采用的财务业绩预测模型普遍存在着泛化能力不强的问题。本文提出用支持向量机方法来预测我国上市公司的财务业绩。为了提高预测准确率,本文还用AdaBoost算法对支持向量机进行了改进(集成支持向量机)。在支持向量机核函数的选择上,我们采用了实验法,即对每个核函数及其相关参数的预测效果都进行了测算,以期找出最适用的预测模型。实证结果表明,径向基核函数(rbf)的效果最好,支持向量机方法预测准确率远远高于其它方法。
To correctly predict the financial performance of the company, we must first select the appropriate forecasting method. The existing models of financial performance prediction used in the literature generally have the problem of poor generalization ability. This paper proposes to use the support vector machine method to predict the financial performance of China’s listed companies. In order to improve the prediction accuracy, we also improve the support vector machine by using AdaBoost algorithm (integrated support vector machine). In the choice of support vector machine kernel function, we use the experimental method, that is, the prediction of each kernel function and its related parameters have been measured in order to find the most suitable prediction model. The empirical results show that the radial basis function (rbf) is the best, and the SVM prediction accuracy is much higher than other methods.