模糊时间序列模型在短期气候预测中的应用

Application of fuzzy time series model in short-term climate prediction

将模糊时间序列模型引入短期气候预报,利用重庆34个地面气象观测站的逐日观测资料(1971—2007年)和重庆市旱涝灾害监测预警决策服务系统计算的干旱指数和洪涝指数等资料,运用模糊时间序列模型分别对2001—2007年重庆市城口县1月降水、1月平均气温的预报结果(年度预测)和重庆市春旱指数的预报结果(年度预测)进行了模糊时间序列分析,预测了2004—2007年的发展趋势,用2004—2007年实测值与预测结果进行了比较,并与加权集成、人工神经网络集成、数据挖掘集成等模型进行了精度比较和分析.结果表明:模糊时间序列模型各项精度评定指标优良,并且计算简单,具有一定的实用价值.

The fuzzy time series model was introduced to short-term climate prediction. Daily ground observations （from 1971 to 2007 ） from 34 weather stations of Chongqing, and the drought index and the flood index computed by Chongqing drought and flood monitoring and warning decision making system are used. The precipitation, average temperature of Chengkou county of Chongqing in January, and the Spring drought index of Chongqing from 2001 to 2007 are analyzed by fuzzy time series model. The trend in precipitation, average temperature, and the Spring drought index of 2004--2007 are predicted by fuzzy time series method, and the results are compared with that of other models like weight integrating forecast model, artificial neural networks, data mining integrating forecast. The comparison shows that fuzzy time series model is better in accuracy and simpler for computation. The fuzzy time se ries method will be more valuable in future application.