Feature Extraction and Tropical Cyclone Prediction System Using Correlation Coefficient and Logistic Regression

Since tropical storms (TS) are among the most hazardous natural disasters, timely and precise track forecasts are essential for preventing and mitigating disasters.In this study, we proposed a feature extraction method based on the correlation between two similar cyclones.The cyclone track is predicted using a multiple logistic regression technique that compares three sigmoid functions.We were able to estimate the latitude and longitude of a cyclone's location 24 hours ahead of time by using the last twelve hours of observations (two positions, at six hourly intervals, and the current position).We investigate the real-world transient circulation of TCs in the Northern Indian Ocean between 1945 and 2022.There is a mean error of 137.5 km in estimating the area between the historical track and the forecasted points.The tropical cyclone track was most significantly influenced by the minimum sea level pressure and maximum sustained wind speed for short-term forecasts.