Modeling rainfall-runoff process using artificial neural network with emphasis on parameter sensitivity

被引:35
作者
Vidyarthi, Vikas Kumar [1 ]
Jain, Ashu [2 ]
Chourasiya, Shikha [3 ]
机构
[1] SRM IST NCR Campus, Modinagar 201204, UP, India
[2] Indian Inst Technol Kanpur, Dept Civil Engn, Kanpur 208016, Uttar Pradesh, India
[3] Cent Univ Jharkhand, Water Engn & Management, Ranchi, Bihar, India
关键词
Rainfall-runoff; Artificial neural network; Gradient descent; Levenberg-Marquardt; Hydrology; PREDICTION;
D O I
10.1007/s40808-020-00833-7
中图分类号
X [环境科学、安全科学];
学科分类号
08 ; 0830 ;
摘要
The gradient descent (GD) and Levenberg-Marquardt (LM) algorithms are commonly adopted methods for training artificial neural network (ANN) models for modeling various earth system and environmental processes. The performance of these algorithms is sensitive to the initialization of their parameters. The initialization of the algorithm's parameters for modeling different physical processes also varies process to process. However, there is a minority that tried to verify the sensitivity of the parameters of the algorithm than the sensitivity of the input data to the model. This work investigates the sensitivity of the popular ANN training algorithms to initial weights while modeling one of the earth system processes, i.e., the rainfall-runoff (RR) process. A novel methodology consisting of basic statistics for assessment of sensitivity of ANN parameters is proposed. The rainfall and flow data derived from three contrasting catchments are employed to establish the conclusions of this study. The results indicate that the RR model trained by LM algorithm is more robust in achieving performance with less variance irrespective of the existence of randomness in initialization of parameters than that of the GD trained models.
引用
收藏
页码:2177 / 2188
页数:12
相关论文
共 33 条
[1]  
Abrahart RJ, 2000, HYDROL PROCESS, V14, P2157, DOI 10.1002/1099-1085(20000815/30)14:11/12<2157::AID-HYP57>3.0.CO
[2]  
2-S
[3]   A comparison between neural-network forecasting techniques - Case study: River flow forecasting [J].
Atiya, AF ;
El-Shoura, SM ;
Shaheen, SI ;
El-Sherif, MS .
IEEE TRANSACTIONS ON NEURAL NETWORKS, 1999, 10 (02) :402-409
[4]   Identifying Prominent Explanatory Variables for Water Demand Prediction Using Artificial Neural Networks: A Case Study of Bangkok [J].
Babel, Mukand Singh ;
Shinde, Victor R. .
WATER RESOURCES MANAGEMENT, 2011, 25 (06) :1653-1676
[5]   Rainfall-Runoff Modeling: Comparison of Two Approaches with Different Data Requirements [J].
Bhadra, A. ;
Bandyopadhyay, A. ;
Singh, R. ;
Raghuwanshi, N. S. .
WATER RESOURCES MANAGEMENT, 2010, 24 (01) :37-62
[6]   Integration of artificial neural networks with conceptual models in rainfall-runoff modeling [J].
Chen, JY ;
Adams, BJ .
JOURNAL OF HYDROLOGY, 2006, 318 (1-4) :232-249
[7]   An artificial neural network approach to rainfall-runoff modelling [J].
Dawson, CW ;
Wilby, R .
HYDROLOGICAL SCIENCES JOURNAL-JOURNAL DES SCIENCES HYDROLOGIQUES, 1998, 43 (01) :47-66
[8]  
Deb K., 2010, OPTIMIZATION ENG DES
[9]  
Govindaraju RS, 2000, J HYDROL ENG, V5, P115
[10]   PHYSICALLY BASED HYDROLOGIC MODELING .2. IS THE CONCEPT REALISTIC [J].
GRAYSON, RB ;
MOORE, ID ;
MCMAHON, TA .
WATER RESOURCES RESEARCH, 1992, 28 (10) :2659-2666