Meteorological Data Analysis using Artificial Neural Networks

Abstract

This paper focuses on weather data analysis for Bangalore urban region(Karnataka,India) over a span of 30 years. The 30 years data is preprocessed to have average monthly temperature, vapor pressure, PET (Potential-Evapo Transpiration), cloud cover, rainfall. These features are considered as factors affecting the rainfall. The correlation between the above mentioned parameters with the monthly rainfall are found using spearman correlation. Artificial Neural Networks (ANN) is used to classify instances as less rain, medium and heavy rain. The results of accuracy, confusion matrix is tabulated. Also the optimal number epochs, number of neurons and number of hidden layers is also identified for the data. The graph of actual output and predicted output is plotted.