Artificial Neural Networks and Backward Propagation Model for Weather Forecasting and Monitoring in Real-Time Environment - A Review

Abstract

Weather Forecasting rather than a binary decision is a statistical measure. In this paper, the authors have studied towards proposing a smart weather-monitoring model as it has become a necessity in today’s scenario. The proposed model checks the parameters such as max and min temperature and precipitation for a designated period of time and is monitored. An accurate and smart approximation based on the available data is accomplished using machine learning paradigms. A measure of more than 89% accuracy is obtained, depending upon the quality of data values. Weather Monitoring and Forecasting helps to determine the future conditions of the atmosphere. We have studies and prepared this paper on weather forecasting and monitoring by making use of ANN( Artificial Neural Network )methodology. We are focusing and concentrating on the implementation of a data-centric model using the various data mining methods and techniques available. Weather represents a process of constant changing and non-linear process and advanced methods of artificial neural network (ANN) can take care of such processes. The data mining methodology in collaboration with upcoming network of neurons which gives meaningful information for weather monitoring and forecasting which further reduces its cost as compared to other monitoring and prediction models.

Keywords: Weather forecasting, Artificial Intelligence, ANN, Numerical Weather Prediction, Back Propagation Approach