
Research Article
A Comparison Between Stacked Auto-Encoder and Deep Belief Network in River Run-Off Prediction
@INPROCEEDINGS{10.1007/978-3-030-67101-3_6, author={Bui Tan Kinh and Duong Tuan Anh and Duong Ngoc Hieu}, title={A Comparison Between Stacked Auto-Encoder and Deep Belief Network in River Run-Off Prediction}, proceedings={Context-Aware Systems and Applications, and Nature of Computation and Communication. 9th EAI International Conference, ICCASA 2020, and 6th EAI International Conference, ICTCC 2020, Thai Nguyen, Vietnam, November 26--27, 2020, Proceedings}, proceedings_a={ICCASA \& ICTCC}, year={2021}, month={1}, keywords={Runoff prediction Stack autoencoder Deep belief network Srepok river}, doi={10.1007/978-3-030-67101-3_6} }
- Bui Tan Kinh
Duong Tuan Anh
Duong Ngoc Hieu
Year: 2021
A Comparison Between Stacked Auto-Encoder and Deep Belief Network in River Run-Off Prediction
ICCASA & ICTCC
Springer
DOI: 10.1007/978-3-030-67101-3_6
Abstract
The application of deep neural networks in forecasting hydrological time series data is increasingly popular, aiming to improve prediction accuracy in this challenging problem. As for river runoff prediction, Deep Belief Network (DBN) and Stacked Autoencoder (SAE) are two kinds of deep neural networks which are commonly used for extracting meaningful features from the data before prediction. In this study, we aim to compare the prediction performance of SAE model with that of DBN model on the runoff data of Srepok River in Central Highlands of Vietnam. Experiments are conducted by using historical data of the Srepok River that were collected in 11 years. The experimental results in this case study show that SAE brings out better prediction accuracy than DBN in terms of three evaluation criteria: correlation, root mean square error, and mean absolute percentage error.