Quantitative Rainfall Prediction: Deep
Neural Network-Based Approach
Debraj Dhar, Sougato Bagchi, Chayan Kumar Kayal, Soham Mukherjee
and Sankhadeep Chatterjee
Abstract Forecasting the weather has always been a challenge using conventional
methods of climatology, analogue and numerical weather prediction. To improvise
the prediction of weather much further, the proposed method can be used. In this
work, authors proposed a method which uses the advantages of deep neural network
to achieve high degree of performance and accuracy compared to the old conventional
ways of forecasting the weather. It is done by feeding the perceptrons of the DNN
some specific features like temperature, relative humidity, vapor and pressure. The
output generated is a highly accurate amount of the rainfall based on the given input
data.
Keywords Forecast · Rainfall prediction · Deep neural network · Perceptrons
1 Introduction
A meteorologist’s biggest job is to predict how the weather will change depending
upon the climate changing parameters and it has always been a challenging job to
predict this on a higher scale of accuracy [1]. Using DNN (deep neural network) [2],
the task of predicting the weather [3] can be achieved with much greater accuracy.
D. Dhar · S. Bagchi · C. K. Kayal · S. Mukherjee · S. Chatterjee (B )
Department of Computer Science & Engineering, University
of Engineering & Management, Kolkata, India
e-mail: chatterjeesankhadeep.cu@gmail.com
D. Dhar
e-mail: debrajdhar100@gmail.com
S. Bagchi
e-mail: sougato97@gmail.com
C. K. Kayal
e-mail: chayankayal32@gmail.com
S. Mukherjee
e-mail: mukherjee.soham56@gmail.com
© Springer Nature Singapore Pte Ltd. 2019
M. Chakraborty et al. (eds.), Proceedings of International Ethical Hacking
Conference 2018, Advances in Intelligent Systems and Computing 811,
https://doi.org/10.1007/978-981-13-1544-2_37
455