| 青岛市大气污染时间序列分析预报方法研究 |
| |
| 引用本文： | 黄磊,王赐震,王建华,于鹏.青岛市大气污染时间序列分析预报方法研究[J].中国海洋大学学报(自然科学版),2001,31(1):14-20. |
| |
| 作者姓名： | 黄磊 王赐震 王建华 于鹏 |
| |
| 作者单位： | 1. 青岛海洋大学大气科学系,
2. 青岛市环境科学研究所, |
| |
| 基金项目： | 青岛市科委“大气污染预报研究”课题资助 |
| |
| 摘 要： | 根据青岛市大气污染监测资料 ,采用时间序列分析方法 ,建立多种预报模型 ,有原序列周期外延法、均生函数周期外延法、均生函数逐步回归法以及自回归预报法等 ,最后提出一种综合预报模型。连续预报试验表明 ,综合预报模型优于任何个别预报模型 ,有较好的预报能力。利用马尔可夫概型对污染状态 (轻、中、重 )进行了分级预报试验 ,也获得良好的效果。
|
| 关 键 词： | 大气污染 预报 时间序列分析 综合模型 马尔可夫概型 |
| 文章编号： | 1001-1862(2001)01-014-07 |
| 修稿时间： | 2000年7月6日 |
Research on the Time Series Analysis Forecast Method for Air Pollution in Qingdao |
| |
| Huang Lei,Wang Cizhen.Research on the Time Series Analysis Forecast Method for Air Pollution in Qingdao[J].Periodical of Ocean University of China,2001,31(1):14-20. |
| |
| Authors: | Huang Lei Wang Cizhen |
| |
| Abstract: | Based on the observations of air pollutants from 1995 to 1999 in Qingdao, 5 kinds of air pollution forecast methods for statistical time series analysis are established. The forecast methods include the extrapolation of periods of time series, the extrapolation of periods of Mean Generation Function (MGF), the MGF step wise regress method and the autoregress forecast method. We further put forward a synthetic forecast model. Continuous forecast experiments show that the latter is superior to the former 4 models. Markov chain models are also used to forecast the three states of pollution (light, intermediate and heavy) and the forecast results are satisfactory. |
| |
| Keywords: | air pollution forecast time series analysis synthetic model Markov chain models |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
