支持向量机在耕地面积预测中的应用研究

Studyon Cultivated Land Prediction Base on Support Vector Machines

研究耕地面积准确预报问题,粮食产量稳定性受到耕地面积变化的影响。耕地面积是一种高度不稳定、复杂且难以预测,传统预测方法都是根据线性模型,忽略了耕地面积的非线性特征,导致预测精度不高。为了提高耕地面积预测的精度,提出一种基于支持向量机的耕地面积预测模型。利用相关分析和灰色关联分析对影响耕地面积变化的因子进行筛选,做为支持向量机的输入,耕地面积数量作为输出,可通过粒子群法对支持向量机的参数进行寻优,最后建立最优的耕地面积预测模型,对江苏无锡市耕地面积进行仿真。仿真结果表明,支持向量机模型比神经网络和其它预测模型有较高的预测精度,适用于耕地面积预测等非线性问题,为耕地面积预测提供了依据。

Cultivated land prediction is a complicated nonlinear dynamic system. It is very difficult to draw its in- herent rules using traditional timing prediction technique. In order to improve the analysis of the pattern of cultivated land prediction, a cultivated land prediction method is proposed based on the theory of support vector machines. First, using correlation analysis and grey correlation analysis to choose the influence factors of cultivated land as the input of support vector machine, and the cultivated land area as the output, and then through parameters optimization of particle swarm support vector machine, the optimal forecasting model is established, and the simulation experiment of cultivated land area in wuxi city, jiangsu, is carried out. Simulation results show that the support vector machine （SVM） model has higher prediction precision, higher generalization ability and forecast accuracy than the other mod- els, and therefore is suitable for orediction of the cultivated area.