| 佟金鹤，刘丽君.基于贝叶斯方法的海口空气质量等级预报模型[J].海南师范大学学报自科版,2025,38(2):238-243 |
| 基于贝叶斯方法的海口空气质量等级预报模型 |
| Forecast Model for Air Quality Level in Haikou Based on Bayesian Method |
| |
| DOI：10.12051/j.issn.1674-4942.2025.02.015 |
| 中文关键词: 空气质量等级预报 朴素贝叶斯方法 气温 北风分量 |
| 英文关键词: air quality level forecast Naive Bayes temperature north wind component |
| 基金项目:海南省自然科学基金项目（421QN0967，422RC802） |
|
| 摘要点击次数: 1237 |
| 全文下载次数: 509 |
| 中文摘要: |
| 为提升海口地区空气质量等级预报准确率，使用2019—2021年气象观测数据、再分析
数据和空气质量观测数据分析不同空气质量等级下气象要素的分布并建立空气质量等级的贝叶
斯预报模型，使用2022年数据进行检验。结果表明：相对湿度较低、北风分量适中、总云量低、无降
水以及适宜的气温是海口空气质量指数上升的有利条件；贝叶斯方法对空气质量等级有较好的预
报效果，在检验集中准确率达到78.08%，对空气质量一级天气预报效果最好，二级次之，三级最差；
模型对空气质量二级和三级的预报偏差主要源于较高的空报率，其中二级空报率为42%，三级为
70%；模型对空气质量三级天气的漏报率较低，仅为 10%，可在空气质量等级预报中发挥较好的
作用。 |
| 英文摘要: |
| Meteorological observation data, reanalysis data and air quality observation data in Haikou during 2019 to 2021
were used to analyze the distribution of meteorological factors under different air quality levels and establish a Bayesian
model for air quality level forecasting with data in 2022 as a test set. The results showed that an atmosphere condition of low
relative humidity, moderate northerly component, low total cloud cover, no precipitation, and suitable temperature could
contribute to the raising of Haikou’s air quality index (AQI). The Bayesian method had a good effect on air quality level
forecasting, with an accuracy rate of 78.08% in test set. The first-level air quality was predicted with the best performance,
followed by the second-level, and the third-level. The deviation of the model from the second and third levels was mainly
due to higher vacancy rates, with a second-level of 42% and a third-level of 70%. However, the model had a lower false
negative rate of only 10% for the third-level of air quality, indicating that this model could play a good role in the forecast⁃
ing of air quality levels. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
