8. Comparison of Neural Network Models for Weather Forecasting

Weather forecasting has a big impact on people’s lives from event planning to cultivation. Conventionally, it has been performed by simulating physical conditions of the atmosphere. Due to nonlinear and irregular kind of weather data, machine learning methodologies can be seen as an alternative of the physical model for forecasting weather. This paper traverses the potential of deep neural networks in the field of weather prediction. It compares the performance of two different neural network models. First model uses the feed- forward network while other uses recurrent neural network to feed the weather data. The models illustrate that neural network models are emulative with the conventional methods and can be perused as a better alternative to predict general meteorological conditions.