Weather Forecasting using Soft Computing Techniques

Weather has chaotic nature and so it has always been most challenging job for the scientists and so new methods and models have been developed in order to reach maximum accuracy in the forecasting of weather. Weather has its direct impact on the economy of the country as maximum share of GDP is taken by agriculture and other indirectly dependent fields. Weather and climate are different from each other on the bases of time. Weather is measured from days to months but climate change is for long range. Weather includes the following component: Rain, Cloud cover, Sun shine, Temperature, Wind, Lightening, Wind Direction, Humidity, Barometric Pressure, wind Speed, snow. Weather on Earth is considered to be Chaotic in nature and hence has large impact even if there is some minute change at any point on the surface. Scientist have worked rigorously in the field of climate change analysis and weather prediction using different mathematical models. In this paper a comparative study for a few models on the basis of the statistical values comparison is studied. The models are: Multi-Layer Perceptron, Linear Regression, Support Vector Regression and Gaussian Process. The paper aims to attain the highest efficiency in the prediction of the weather affecting parameters. The prediction of the weather affecting parameters has reached to highest accuracy for long range, medium range, short range and so the objective of the study is to scrutinize the best model for prediction.