基于时间序列的流域氟化物预测模型研究

Study on prediction model of watershed fluoride based on time series

采用时间序列分析方法拟合具有趋势性和季节性的流域水体中氟化物含量并进行预测。利用清水江流域2013—2018年的每月氟化物监测数据,用差分和季节差分方法对监测数据进行平稳化,采用ARIMA乘积季节模型(p,d,q)(P,D,Q)s拟合序列,应用残差和BIC进行模型参数调整,建立氟化物时间序列预测模型,并对测试集月均氟化物浓度进行了预测。预测结果虽然与实测结果有差距,对可能产生预测误差的原因进行了讨论分析,表明本模型受环保控污政策及季节气候变化、人类活动等因素影响显著。

Time series analysis was used to fit and predict fluoride content in trendy and seasonal watershed water bodies.By using the monthly fluoride monitoring data of Qingshui river basin from 2013 to 2018,the monitoring data were smoothed by differential and seasonal difference methods.The ARIMA product seasonal model(p,d,q)(P,D,Q)s was used to fit the sequence,the model parameters were adjusted by using residual and BIC to establish the fluoride time series prediction model,and the monthly average fluoride concentration of the test set was predicted.Although there is a difference between the predicted results and the measured results,this study discusses the possible reasons for the prediction error.The results show that this model is influenced significantly by environmental pollution control policy,seasonal climate change and human activities.