基于季节性AR(P)模型的水质预测

Water Quality Prediction Based on Seasonal AR(P) Model

自回归模型的建立是基于序列平稳性的假设,只能描述平稳序列的统计特性,而水质的月监测数据序列往往具有季节性变化的现象.文章介绍了平稳过程的相关理论及其检验方法并应用到黄河潼关、三门峡断面的水质序列的检验中,检验结果为非平稳序列,且序列具有明显季节性(月份)变化的特性.为此尝试建立季节性AR(P)模型来捕捉黄河水质的季节性变化规律,实践表明该模型预测总体效果是较为满意的.

The regression model is based on a series of assumptions, which can only describe the statistical characteristics of smooth steady sequence, while water quality monitoring data from the test sequence are often of seasonal phenomenon. This paper introduces the relevant theories and testing methods in the smooth process, as well as the application in the test sequence quality at Tongguan, Sanmenxia section of the Yellow River, the test results show non- stationary sequence of obvious seasonal （monthly） changes. The seasonal AR （1） model is attempted to establish to capture the seasonal changes in water quality in the Yellow River. Experiments show that the overall effect is satisfying.