A Hybrid Genetic Algorithm Based Rainfall Prediction Model Using Deep Neural Network

Abstract

This paper is intended to design and develop an efficient prediction algorithm in the context of forecasting in the precipitation as well as the intensity of the rainfall in a local region over a relative short period of time. Despite of several existing addressing the problem of forecasting rainfall, still it is remind as an open challenge in the context of dynamic data acquisition and aggression. Prediction of rain is main application of science and technology to predict the state of the atmosphere. The main objective of the study is to develop a Genetic Algorithm based approach that utilize dimensionality reduction technique and Multi-layer Perceptron for efficient and dynamic analysis of real time data. Furthermore a deep neural networks based framework is proposed to predict rainfall of a certain region. A Hybrid genetic algorithm presents a novel solution to predicting the rainfall in certain area or different regions by using Deep Neural Network

Forecast Rainfall Density by Utilizing Machine Learning Models

“…[9] have used BP neural network to forecast the rainfall for coastal areas. [4] proposed a method constructed on the Genetic Algorithm that makes use of a dimensionality reduction methodology as well as a Multilayer Perceptron (MLP) for the purpose of performing an analysis that is both dynamic and efficient using real-time data. [3] suggests using ensemble learning to improve the accuracy of forecasting rainfall.…”

“…An artificial neural network (ANN) [4], [16] that has a feed-forward architecture and is made up of numerous layers of neurons is called a multilayer perceptron (MLP). A MLP has three or more layers including one input, one output, and one or even more hidden layers.…”