Using Machine Learning Approaches to precisely Predict RainFall

Abstract

Meteorologists continually seek techniques to grasp the Earth's atmosphere and build precise weather forecast models. In weather prediction, several approaches were employed. Machine approaches are recently thought to be precise procedures and have been widely utilised for weather prediction, as an alternative to conventional methods. The precipitation rate has an important impact on agriculture and the bio-sector in the weather system. This article aims at the development of multiple linear regression model to forecast the rainfalls (PRCP). It depends on several meteorological characteristics such as temperature, wind velocity and wind speed. The information utilised in this research is published on the National Climate Data Center website. A Python code was created in order to develop the pattern for artificial neural networks using the package Pytorch. Measured by comparing the mean square error value of the training data to test data, the efficiency of the model has been measurable. The findings achieved demonstrate that, while using the same amount of data throughout trainings and testing periods, the average square error has been improved by 85 percent during trial time. However, the amount of data during the test phase goes above the quantity of data in the training phase to 59 percent