论文部分内容阅读
害虫发生是一种复杂、动态时间序列数据,单一预测模型都是基于线性或非线性数据,不能同时捕捉害虫发生的线性和非线性规律,很难达到理想的预测精度。本研究首先采用差分自回归移动平均模型对昆虫发生时间序列进行线性建模,然后采用支持向量机对非线性部分进行建模,最后得到两种模型的组合预测结果。将组合模型应用到松毛虫Dendrolimus punctatus发生面积的预测,实验结果表明组合模型的预测精度明显优于单一模型,发挥了两种模型各自的优势。组合模型是一种切实可行的害虫预测预报方法。
Pest occurrence is a complex and dynamic time-series data. The single prediction model is based on linear or non-linear data and can not simultaneously capture the linear and nonlinear laws of pest occurrence. It is difficult to achieve the desired prediction accuracy. In this study, the differential autoregressive moving average model was used to model the time series of insects. Then the nonlinear part was modeled by support vector machine. Finally, the combined forecasting results of the two models were obtained. The combined model was applied to predict the area of Dendrolimus punctatus. The experimental results show that the prediction accuracy of the combined model is obviously superior to that of the single model, and the respective advantages of the two models are exerted. The combinatorial model is a feasible pest forecasting method.