对于小样本而言，SVM的仿真效果要比神经网络好，但是SVM的性能依赖于它的两个训练参数，本算法是用GA自动选择SVM的两个参数。-For small sample case, SVM simulation results than the neural network is good, but the performance of SVM depends on its two training parameters, the algorithm is automatically selected

这是组合一个优化算法GABP，很好的提高BP预测精度-It is a combination of an optimization algorithm GABP, good way to improve the prediction accuracy of BP