WEATHER FORECASTING USING ARTIFICIAL NEURAL NETWORK

Abstract

There has been a huge growth of research within the past few years with regards to weather forecasting. An attempt was made to establish a direct relation between the input weather data and the targeted data. Due to the lack of regularity in the nature of the weather data, there has been increase in the focus over the irregular nature prediction of the weather data. Since the weather data has an irregular trend, Artificial Neural Network (ANN) has advanced and considered to be a favorable method to bring out the structural relation between the different entities. This paper looks over the applicability of Artificial Neural Networks approach by creating a effective and robust predictive model for the analysis of weather and also to compare and assess the performance of the model using different hidden layer, transfer functions and neurons to forecast the weather for a whole year (365 days).

Back-Propagation Algorithm and its use would be explained within the paper, based upon which the model has been created. The Back-Propagation Algorithm has been used to analysis the weather forecast. The paper also speaks about the output achieved in graphical format in the process while analyzing the weather. MATLAB R2021a has been used for processing the simulation.