华东师范大学学报(自然科学版) ›› 2006, Vol. 2006 ›› Issue (6): 41-46.

• 地理学 河口海岸学 • 上一篇    下一篇

基于贝叶斯方法的中尺度对流系统移动方向研究

苏君毅, 邱 洁, 过仲阳, 戴晓燕   

  1. 华东师范大学 地理信息科学教育部重点实验室，上海 200062
  • 收稿日期:2005-07-05 修回日期:2005-11-21 出版日期:2006-11-25 发布日期:2012-11-09
  • 通讯作者: 过仲阳

Study on the Trajectories of MCS Based on Bayesian Classification(Chinese)

SU Jun-yi, QIU Jie, GUO Zhong-yang, DAI Xiao-yan   

  1. Key Laboratory of Geographic Information Science, Ministry of Education, East China Normal University, Shanghai 200062, China
  • Received:2005-07-05 Revised:2005-11-21 Online:2006-11-25 Published:2012-11-09
  • Contact: GUO Zhong-yang

摘要： 在朴素贝叶斯分类的基础上建立了一种增强型分类器系统，并在对1997～2002年夏季青藏高原上MCS(Mesoscale Convective System)进行自动追踪的基础上，对MCS的移动方向与其周边环境物理量场的分布特征进行了分类研究.进而，将分类结果与决策树、人工神经网络分类方法进行了比较.研究表明，与其他分类方法相比，使用增强型的贝叶斯分类器预测MCS的移动路径具有较好的效果，这为揭示高原上MCS的移动规律、提高长江中下游地区灾害天气预报的准确率提供了一种有效的方法.

关键词: 青藏高原, 中尺度对流系统, 贝叶斯分类, 青藏高原, 中尺度对流系统, 贝叶斯分类

Abstract: In this paper, a Boosting Classifier based on Naive Bayesian Classification was built and applied to classify the trajectories of MCS, using a dataset of environmental physical field values around MCS, based on the automated tracking of MCS over the Tibetan Plateau in summer from 1997 to 2000. Furthermore, results comparing several classification methods found the Boosting Bayesian Classifier to be comparable in performance with decision tree and neural network classifiers in the application of prediction of the trajectories of MCS. So it is proven to be an effective method to reveal the trajectories of MCS over the Tibetan Plateau and improve the accuracy of forecasting the disaster weather in Yangtze River Basin.

Key words: Mesoscale Convective System, Bayesian Classification, Tibetan Plateau, Mesoscale Convective System, Bayesian Classification

中图分类号: