FORECASTING AMMONIA NITROGEN USING NONLINEAR GREY BERNOULLI MODEL IN BAOTOU REACH OF YELLOW RIVER, CHINA

摘要：

　　In this study, GM(1,1), grey Verhulst model (GVM) and nonlinear grey Bernoulli model (NGBM (1, 1)) were compared for short-term forecasting ammonia nitrogen (NH3-N) in Baotou reach of Yellow River.The model performance was evaluated by absolute percentage error (APE), mean absolute percentage error (MAPE), root mean square error (RMSE), mean absolute error (MAE), and Thei IC (TIC).The parameter r in NGBM (1, 1) was determined by nonlinear unconstrained optimization method, which is simple and effective.The empirical results indicate that GM (1, 1) outperforms GVM for in-sample forecasting and GVM outperforms GM (1, 1) for out-of-sample forecasting, and NGBM (1, 1) can achieve a good fitness with the minimum MAPE (10.72%) for in-sample forecasting and the minimum MAPE (5.4%) for out-of-sample forecasting.This demonstrates that NGBM (1,1) with an optimal parameter r can obviously improve forecast accuracy.

关键词： Ammonia Nitrogen GM (1,1) Grey Verhulst Model Nonlinear Grey Bernoulli Model Water Quality

作者: Yan An Zhihong Zou Xiaojing Wang

作者单位: School of Economics and Management, Beihang University, Beijing 100191, China

会议类型: 国际会议

会议名称: The 12th International Conference on Industrial Management(第十二届工业管理国际会议)

会议地点: 成都

会议语种:英文

页码: 476-481

在线出版日期: 2014-09-03（万方平台首次上网日期，不代表论文的发表时间）