Effective Features and Hybrid Classifier for Rainfall Prediction

DOI

10.1080/18756891.2014.960234How to use a DOI?

Keywords

rainfall prediction, hybrid classifier, feature indicator, ABC, genetic, FFNN

Abstract

Rainfall prediction has emerged as a challenging time-series prediction problem in recent years. In this paper, we propose a novel rainfall prediction technique using effective feature indicators and a hybrid technique. Our proposed model consists of three phases, namely, layer model simulation, training phase and testing phase. At the outset, the input rainfall dataset is preprocessed using the feature indicators. There are five feature indicators used in the preprocessing step namely, channel index (CI), ulcer index (UI), rate of change (ROC), relative strength index (RSI) and average directional movement index (ADX). Subsequently, feature matrices are formed based on the preprocessed rainfall data. Once the feature matrix is formed, the prediction is done based on the hybrid classifier. In the hybrid classifier, artificial bee colony algorithm is combined with the genetic algorithm for training the feed forward neural network. The performance of the algorithm is analyzed with the help of real datasets gathered from Rayalaseema, Aandhra and Telangana regions. Finally, from comparative analysis it is established that the proposed rainfall prediction yields better result (MAC=4.0672) when compared with Artificial Bee Colony with Neural Network.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - JOUR
AU  - KavithaRani B.
AU  - A. Govardhan
PY  - 2014
DA  - 2014/10/01
TI  - Effective Features and Hybrid Classifier for Rainfall Prediction
JO  - International Journal of Computational Intelligence Systems
SP  - 937
EP  - 951
VL  - 7
IS  - 5
SN  - 1875-6883
UR  - https://doi.org/10.1080/18756891.2014.960234
DO  - 10.1080/18756891.2014.960234
ID  - B.2014
ER  -