基于支持向量机的中国农业机械总动力预测  

Prediction on total power of Chinese agricultural machinery based on support vector machine

作　　者：李建伟[1] 周洪[2] 赵汉雨[1] 陈沛然[3] 

机构地区：[1]河南农业大学机电工程学院,河南郑州450002 [2]河南工程学院电气信息工程学院,河南郑州451191 [3]北京航空航天大学自动化科学与电气工程学院,北京100191

出　　处：《河南农业大学学报》2013年第3期296-300,共5页Journal of Henan Agricultural University

基　　金：河南省科技攻关项目(102102210535)

摘　　要：为更好地预测中国农业机械总动力的发展趋势,引入了基于支持向量机的预测方法.以1979—2008年中国农业机械总动力的统计数据为训练样本,以2009年和2010年的统计数据为检验样本,采用新陈代谢法建立了基于支持向量机的我国农业机械总动力预测模型.为了验证该方法的有效性和优越性,同时采用新陈代谢法分别建立了基于普通BP神经网络和改进的BP神经网络的预测模型.仿真预测与检验样本预测的结果表明,基于支持向量机的预测精度明显高于普通BP神经网络和改进的BP神经网络预测模型.在此基础上,计算出2011年至2015年中国农业机械总动力的预测值分别为97 859.1,103 053.7,108 480.7,112 794.7,115 096.8万kW,指出了其具有增长趋缓的变化趋势.In order to predict the development trends of total power of Chinese agricultural machinery better, a method based on support vector machine was presented. In the building of prediction model based on support vector machine, the metabolic method was used. Statistical data of Chinese agricul- tural machinery＇ s total power from 1979 to 2008 were used as training samples, and statistical data of 2009 and 2010 were used as test samples. To verify the effectiveness and superiority of this method, prediction models based on normal BP neural network and improved BP neural network were also estab- lished using metabolic method. Comparison of predicted results based on historical data shows that the prediction accuracy based on support vector machine was significantly higher than that of normal BP neural network and the improved BP neural network. Then, predictive values of Chinese total power of agricultural machinery from 2011 to 2015 were respectively 97 859.1,103 053.7,108 480.7,112 794.7, and 115 096.8 ten thousand kW）, and the changing trend of slowdown in the growth was also indica- ted.

关 键 词：农业机械总动力 支持向量机 BP神经网络 

分 类 号：S232.3[农业科学—农业机械化工程] TP183[农业科学—农业工程]