83rd Annual

Monday, 10 February 2003
Cloud classification of satellite radiance data by Multicategory Support Vector Machines
Yoonkyung Lee, Ohio State Univ., Columbus, OH; and G. Wahba and S. A. Ackerman
Poster PDF (67.1 kB)
Two category Support Vector Machines (SVM) have been very popular in the machine learning community for the classification problem. Treating multicategory problems as a series of binary problems is very common in the SVM paradigm. However, this approach may fail under a variety of circumstances. We have proposed the Multicategory Support Vector Machine (MSVM), which extends the binary SVM to the multicategory case, and has good theoretical properties. The proposed method provides a unifying framework when there are either equal or unequal misclassification costs.

In this paper, we illustrate the potential of the multicategory SVM as an efficient cloud detection and classification algorithm for use in Earth Observing System models, which require knowledge of whether a radiance profile is cloud free, or not. If the profile is not cloud free, it is valuable to have information concerning the type of cloud, for example ice or liquid water. We have applied the MSVM to simulated MODIS type channels data to classify the radiance profiles as clear, liquid clouds, or ice clouds, and the results are promising. Although this study is not yet comprehensive, it is believed that the MSVM will be a very useful tool for classification problems in atmospheric sciences.

Supplementary URL: http://www.stat.wisc.edu/~wahba/trindex.html