基于改进加权马尔科夫链的兰州市空气质量预测

Predicting Air Quality in Lanzhou with Improved Weighted Markov Chain

近年来城市空气污染问题逐渐成为社会关注的焦点,科学合理地预报城市空气质量不仅能为政府的环境管控措施提供依据,对预防大气污染灾害也具有重要的意义。以空气质量等级预测为切入点,针对单一AQI指标序列波动性大、预测信息不完善的缺陷,引入模糊级别特征值算法和空气质量指数评价方法,对兰州市8月份的小时、日空气质量等级和首要污染物进行预测分析。结果表明:改进加权马尔科夫链对兰州市AQI的预测结果与实际变化趋势较为一致,其预测时间越短,预测效果相对越好,且日预报较小时预报的效果好。该模型预测AQI等级和首要污染物的平均准确率为85.96%、79.07%。其中,PM_(2.5)、PM_(10)、NO_2、CO、O_3和SO_2的空气质量等级的平均预测准确率为84.75%、86.56%、88.71%、100%、94.56%和100%。该方法较适合预测城市的短期空气质量状况,在空气质量预报服务和预报产品中有一定的应用价值。

In recent years,the issue of urban air pollution has gradually become the focus of the society.Scientific and reasonable forecasting of urban air quality can not only provide a basis for the government＇s environmental control measures,but also have important significance for preventing air pollution disasters.This paper takes the air quality rating prediction as the entry point. In view of the large volatility and imperfect prediction information of a single AQI index sequence,we introduce the fuzzy-level eigenvalue algorithm and air quality index evaluation method,and predict the hourly and daily air quality levels and primary pollutants in August in Lanzhou City. The results show that the predicted results of improved weighted Markov chain for Lanzhou AQI are in line with the actual trend. The shorter prediction time has the better effect,and the daily forecast isbetter than hourly. The average accuracy of AQI ratings and primary pollutants is 85. 96% and 79. 07%. The average accuracy of air quality grades for PM2.5,PM10,NO2,CO,O3,and SO2 was 84. 75%,86. 56%,88. 71%,100%,94. 56%,and 100%. This method is more suitable for predicting the short-term air quality status of a city,and has certain application value in air quality forecasting service and forecasting products.