摘要
根据部分时间序列数据贫信息、高噪声和非线性等特点,采用含边值修正的灰色模型进行预测,获取残差序列后运用支持向量回归(SVR)方法对模型进行残差修正得到复合的灰色支持向量回归模型.在支持向量回归中构造具有自适用性的动态惩罚参数Ci替代传统SVR中的不变参数来提高模型的准确性,同时构造算法决定ε以平滑过度调节.广东省工业生产指数的预测试验结果表明,复合模型具有比其他简单模型更理想的预测效果.
The common grey model satisfying with verge value condition was adopted to predict financial time series with some characteristics such as poor information, high noise, non-linearity and so on. Then, the model was revised by support vector regression based on the calculation of the residual error sequence between the predicted values and the original data. Auto-adaptive parameters Ci was adopted to replace C in the standard support vector regression to improve the forecasting accuracy. Meanwhile, an algorithm about ε was proposed to smooth overshooting. Experimental results show that the composite model can achieve more accurate prediction and smoothing overshooting than the other simple models in the predictions of IGIP of Guangdong Province.
出处
《经济数学》
北大核心
2009年第2期98-105,共8页
Journal of Quantitative Economics
