A study of supervised machine learning techniques to predict cyclone.

The most frequent cause of natural disasters in India is tropical cyclones. Early warning, real-time monitoring, impact and damage assessment, and relief operations all depend on remote sensing and GIS. The Bay of Bengal is frequently the source of cyclones of various intensities. In this paper, the actual goal is to predicate the tropical cyclone in the tropical region of the Bay of Bengal, for that the main satellite data set is collected from the NASA data repository. The overall work is being done by using the tool MATLAB, where all machine learning classifications are applied to the data set and trained. As a result, which classifier has the highest accuracy will be considered as the best result. The primary goal of this research is to forecast cyclones using day-by-day satellite data of coastal regions and a variety of weather factors. The objective of this work is threefold. First to design a model to reduce the dependency on a single classification technique, secondly to avoid the over-fitting of data, as well as to improve the accuracy of the prediction. Two number of steps in collective learning: Multiple machine learning models created using the same or different machine learning algorithm. [ABSTRACT FROM AUTHOR]