Data Mining for Tropical Cyclone Intensity Prediction
Document Type
Article
Publication Date
2005
Abstract
Tropical cyclone (TC) intensity prediction is far more challenging than the corresponding TC track prediction. One of the reasons is the lack of understanding of the coupling relationships of the physical processes controlling TC intensification. Data mining provides potential tools for relationship exploration from the ever-increasing amount of TC data. The data mining tools and algorithms require process-related features from raw observational data such as the TC best track data, TRMM data for 3D rainfall and 2D surface information. The data mining tools and algorithms can search for hidden relationships among the underlying features by means of statistical and logical inferences. In this article, we describe the data, definitions of features from the data, tools for data exploration, and several important results.
Recommended Citation
Tang, J., Yang, R., Kafatos, M. (2005). 7.5 Data mining for tropical cyclone intensity prediction. Sixth Conference on Coastal Atmospheric and Oceanic Prediction and Processes.
Peer Reviewed
1
Comments
This paper was presented at the Sixth Conference on Coastal Atmospheric and Oceanic Prediction and Processes in 2005.