Automated quality control of tropical cyclone winds through data mining

The analysis of tropical cyclones (TC) depends heavily on the quality of the incoming data set. With the advances in technology, the sizes of these data sets also increase. There is a great demand for an efficient and effective unsupervised quality control tool. Towards such a demand, data mining algorithms like spatial clustering and specialized distance measures can be applied to perform this task. This paper reports our findings on the studies on utilizing a density-based clustering algorithm with three different distance measures on a series of TC data sets.