Combining Support Vector Machine and Polynomial Regressing to Predict Tropical Cyclone Track

Tropical cyclone track prediction is the essential role in disaster prevention and mitigation. Early stage of the cyclone formation is difficult to detect automatically. Many image processing and classification methods are used to detect the region of interest from satellite images, but they are not perfect and sometimes misdetection can be occurred in the early stage of the cyclone formation. In this research, the Tropical Cyclone tracking system is developed by combining the Support Vector Machine (SVM) and polynomial regressing model. After estimating the center of the tropical cyclone by SVM, cyclone track is predicted by using polynomial regressing model. Ground truth data for cyclone tracking is obtained from JTWC best track. Five tropical cyclones that have been affected to Myanmar are tested to predict the cyclone track.