时序建模及其在农药降解动态模拟中的应用

Non stationary time series modeling and its application in the dynamic simulation of pesticide degradation。

用数学方法对农药残留量进行预测预报及动态模拟，对于合理使用农药，加强环境保护具有重要意义．作者用时间序列分析的理论与方法，研制农药残留定量预测模型，给出参数估计，提出非平稳时序叠合模型，对农药降解规律进行动态模拟，并以剩余平方和作为优化标准与其他方法进行比较．实例计算结果表明：时序模型预测值与实测值拟合程度优于其他模型。

Applying mathematical approach to make the prediction and dynamic simulation for the quantity of pesticide reminds possess an important significance for using pesticide rationally and enhancing the environment protection. The theory and methods of time series analysis, in this paper, were used to develop the quantitative predicting models for pesticide reminds. The estimation of parameters were given out. The regularity of pesticide degradation was simulated dynamically by non stationary time series additive model. Calculating results of the practical example showed that the fitness between predicting value and actually observed value, compared by the residual square sum as optimizing standards, was better than other method with easy to use.