| 摘要: |
| 以FY-2G卫星反演产品为输入参数建立决策树模型,对2020年贵州降雹进行识别研究。收集了2020年贵州68个降雹点数据和68个未降雹点数据，从中随机选取58组降雹点和58组未降雹点数据用于建立决策树模型,剩余10组降雹点和10组未降雹点数据用于检验所建立模型的识别效果。结果表明,所建模型降雹识别准确率为80%,其中对10个降雹点识别准确率为70%,对10个未降雹点识别准确率为90%。 |
| 关键词: 决策树，冰雹，识别，检验 |
| DOI： |
| 投稿时间：2021-06-09修订日期：2021-08-03 |
| 基金项目:贵州省科技计划项目(黔科合基础-ZK[2021]一般217)：基于风云卫星观测资料的冰雹天气识别研究；贵州省气象局科研业务项目(黔气科登[2020]07-13号):基于Logistic回归模型的贵州降雹识别指标研究;中国气象局人工影响天气中心业务项目：FY3/4卫星云特性产品在西南防雹增雨中的应用示范(一期)。 |
|
| Hail Recognition in Guizhou Based on Decision Tree Model |
| pengyuxiang,wenjifen,lihao,liutao,tangpiru,guoxi |
| (The Weather Modification Office of Guizhou Province;The Guizhou Meteorological Information Center) |
| Abstract: |
| In this paper, based on the FY-2G satellite inversion products as input parameters, a decision tree model is established to identify hail in Guizhou in 2020.Data of 68 hail spots and 68 non-hail spots in Guizhou in 2020 were collected.The data of 58 groups of hailspots and 58 groups of non-hail spots were selected at random to establish the decision tree model, and the remaining 10 groups of hail spots and 10 groups of non-hail spots were used to suggest the recognition effect of the established model.The results show that the recognition accuracy of the model is 80%, among which the recognition accuracy of 10 hail spots is 70%, and the recognition accuracy of 10 non-hail spots is 90%. |
| Key words: decision tree； hail；identification； test |
