| 仝艳时,陈鹏,疏学明,吴振波.时间序列模型在火警短期预测中的应用[J].火灾科学,2008,17(4):216-221. |
| 时间序列模型在火警短期预测中的应用 |
| Application of time series model in short-term prediction fire alarming |
| 投稿时间：2008-09-02 修订日期：2008-09-27 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI：10.3969/j.issn.1004-5309.年.期.顺序 |
| 基金项目:“十一五”科技支撑计划应急平台重大项目资助课题 |
|
| 中文关键词: 时间序列 ARIMA模型 预测 火警 |
| 英文关键词:Time series ARIMA model Forecasting Fire |
| 摘要点击次数: 539 |
| 全文下载次数: 769 |
| 中文摘要: |
| 基于某市的119火警数据,采用时间序列模型中的ARIMA算法进行了分析。计算结果表明,该市119火警数据的天序列服从ARIMA(0,1,2)模型,周序列服从ARIMA(1,0,0)模型。从模型对数据的拟合效果来看,ARI-MA模型较为准确的反映了数据序列的发展趋势。采用这两种模型对数据序列分别进行了短期的预测,其预测值与实际结果在趋势上基本一致,研究表明采用周序列预测的效果要好于天序列。基于时序模型的火警预测方法是实现对火灾应急处置中关口前移的重要措施。 |
| 英文摘要: |
| In this paper,ARIMA algorithm of time series model is used to process the fire alarming data of one certain city.It is shown that day series of fire alarming data suit for ARIMA(0,1,2)model,while the week series suit for ARIMA(1,0,0)model.From the fitting results of the models,the trend of the data series is reflected accurately by the ARIMA models.With the models,short-term forecasting of fire alarming is made.It is certificated that the forecast and the actual results are basically the same in the trend.The mean relative error of forecasting by week is 23%,while by day is 50%,so it is concluded that forecasting by week is better than that by day. |
| 关闭 |
|
|
|
