论文部分内容阅读
支持向量机是一种新型的学习方法,该方法以结构风险最小化原则取代传统机器学习中的经验风险最小化原则,在小样本的机器学习中显示出了优异的性能。传统的支持向量机是解凸二次规划问题,而最小二乘支持向量机是解等式线性方程,显得尤为方便。针对最小二乘支持向量机的特点,通过Bootstrap建立适当的性能指标,用遗传算法(GA)优化最小二乘支持向量机的有关参数,并在非线性经济系统中应用。用最小二乘支持向量机对非线性经济系统进行预测的结果与神经网络预测的结果比较证明,该模型的预测精确度是令人满意的,文中提出的方法是可行的。
SVM is a new learning method. It replaces the principle of minimizing empirical risk in traditional machine learning with the principle of minimizing structural risk, and shows excellent performance in small sample machine learning. The traditional support vector machine is to solve the quadratic programming problem, and least squares support vector machine is to solve the equation linear equation, it is particularly convenient. According to the characteristics of least square support vector machine, Bootstrap is used to establish the appropriate performance index, and the genetic algorithm (GA) is used to optimize the parameters of least square support vector machine, which is applied to nonlinear economy system. The results of the prediction of nonlinear economic system by least squares support vector machine and the results of neural network prediction prove that the prediction accuracy of this model is satisfactory. The proposed method is feasible.