Hybrid Learning Machine Metaheuristic Model for Estimating Groundwater Level Level

Document Type : Original Research

Authors
S. khosravi 1
A. Robati 2
1 Islamic Azad University, Kerman branch
2 Islamic Azad University Kerman Branch
Abstract
Groundwater is the most reliable source of supply for potable water and supports a wide array of economic and environmental services. There is a significant concern that groundwater levels are declining due to intense aquifer use. The sustainable management of groundwater resources requires good planning and concerted efforts. To manage groundwater resources, it is necessary to predict the groundwater levels and its fluctuations. The prediction groundwater level can guide water managers and engineers effectively. On the other hand, there are multifarious types of equipment for measuring levels of groundwater. Sophisticated water level loggers or divers can measure the groundwater level automatically. Sounding devices with acoustic and light signals are also used to check groundwater levels. The use of devices for measuring the level of groundwater is time-consuming and costly. To reduce the time and cost of the groundwater level measuring process, many methods of Artificial Intelligence (AI) have been utilized for estimating the groundwater level. Among the AI methods, SVMs has great ability in predicting non-linear hydrological processes. Support vector machines (SVMs) is as an intelligent computational method for predicting hydrological processes. Recently, (SVMs) have been successfully applied in classification problems, regression and predicting; as techniques of machine learning, statistics and mathematical analysis. The SVM is based on the structural risk minimization (SRM), which can escape from various difficulties, such as the necessity of a large number of control parameters and a local minimum in artificial neural networks (ANNs). The weighted least squares support vector machines (WLSSVM) was first introduced by Suykens et al., and has proved to be much more robust in several fields, especially for noise mixed data, than least squares version of SVM (LSSVM). Their powerful scientific research provides motivation for employing WLSSVM method in estimating groundwater level. The accurate value of WLSSVM parameters effect on the estimation, these optimal parameters can be achieved optimization algorithms. Therefore, weighted least square support vector machine (WLS-SVM) model was coupled with particle swarm optimization (PSO) and gravitational search algorithm (GSA) as metaheuristic algorithms for estimating well water level. In this study, an attempt has been made to use the hybrid model with high accuracy to estimate the groundwater level. In order to estimate the groundwater level, ten wells data in Bagheyn plain of Kerman province is considered during ten-year time series. The estimated value obtained by the WLSSVM-PSO and WLSSVM-GSA models are compared with the observed value, and showed the estimated results have nearly coincidence with observed values. Numerical results show the merits of the suggested technique for groundwater level simulation. In order to verify the hybrid learning machine metaheuristic model, Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Average Absolute Error (AAE), and Model Efficiency (EF) are computed, and these statistical indicators stand on the good acceptable range, and find WLSSVM-GSA is more accurate than WLSSVM-PSO. The results demonstrate that the new hybrid WLSSVM-GSA model has high efficiency and accuracy with observed values, and the modelling method is an innovative and powerful idea in estimating well water level.

Subjects

[1] Zinivand A. Groundwater Level Modeling with Combining Artificial Neural Network and Wavelet (Case Study: Sharif Abad Plain). MSc Thesis, University of Qom, Qom. 2014.
[2] Deka PC. Support vector machine applications in the field of hydrology: a review. Applied soft computing. 2014; 19: 372-86.
[3] Yoon H, Hyun Y, Ha K, Lee K-K, Kim G-B. A method to improve the stability and accuracy of ANN-and SVM-based time series models for long-term groundwater level predictions. Computers & geosciences. 2016; 90: 144-55.
[4] Eskandari A, Solgi A, Zarei H. Simulating Fluctuations of Groundwater Level Using a Combination of Support Vector Machine and Wavelet Transform. Irrigation Sciences and Engineering Journal. 2016; 41(1):165-80.

[5] Ch S, Mathur S. Groundwater level forecasting using SVM-PSO. International Journal of Hydrology Science and Technology. 2012; 2(2):202-18.
[6] Mallikarjuna B, Sathish K, Krishna PV, Viswanathan R. The effective SVM-based binary prediction of ground water table. Evolutionary Intelligence. 2020:1-9.
[7] Ghourdoyee Milan S, Aryaazar N, Javadi S, Razdar B. Simulation of groundwater head using LS-SVM and comparison with ANN & MLR. Hydrogeology. 2020;5(1):118-33.
[8] Suykens JA, Van Gestel T, De Brabanter J. Least squares support vector machines: World scientific; 2002.
[9] Suykens JA, De Brabanter J, Lukas L, Vandewalle J. Weighted least squares support vector machines: robustness and sparse approximation. Neurocomputing. 2002;48(1-4):85-105.
[10] Aljarah I, Ala’M A-Z, Faris H, Hassonah MA, Mirjalili S, Saadeh H. Simultaneous feature selection and support vector machine optimization using the grasshopper optimization algorithm. Cognitive Computation. 2018; 10(3): 478-95.
[11] Faris H, Hassonah MA, Ala’M A-Z, Mirjalili S, Aljarah I. A multi-verse optimizer approach for feature selection and optimizing SVM parameters based on a robust system architecture. Neural Computing and Applications. 2018; 30(8): 2355-69.
[12] De Brabanter K, Pelckmans K, De Brabanter J, Debruyne M, Suykens JA, Hubert M, et al., editors. Robustness of kernel based regression: a comparison of iterative weighting schemes. International conference on artificial neural networks; 2009: Springer.
[13] Deng N, Tian Y, Zhang C. Support vector machines: optimization based theory, algorithms, and extensions: CRC press; 2012.
[14] Vapnik V. The nature of statistical learning theory: Springer science & business media; 2013.
[15] Cortes C, Vapnik V. Support-vector networks. Machine learning. 1995; 20(3): 273-97.
[16] Quan T, Liu X, Liu Q. Weighted least squares support vector machine local region method for nonlinear time series prediction. Applied Soft Computing. 2010;10(2):562-6.
[17] David H. Early sample measures of variability. Statistical Science. 1998:368-77.
[18] Rousseeuw PJ, Leroy AM. Robust regression and outlier detection: John wiley & sons; 2005.
[19] Kennedy J, Eberhart R, Shi Y. Swarm Intelligence Morgan Kaufmann Publishers. San Francisco. 2001.
[20] Shi Y, Eberhart R, editors. A modified particle swarm optimizer. 1998 IEEE international conference on evolutionary computation proceedings IEEE world congress on computational intelligence (Cat No 98TH8360); 1998: IEEE.
[21] Rashedi E, Nezamabadi-Pour H, Saryazdi S. GSA: a gravitational search algorithm. Information sciences. 2009;179(13):2232-48.
[22] Hoang N-D, Pham A-D. Hybrid artificial intelligence approach based on metaheuristic and machine learning for slope stability assessment: A multinational data analysis. Expert Systems with Applications. 2016;46:60-8.
[23] Kalantari M, Akbarpour A, Khatibinia M. Numerical Modeling of Groundwater Flow in Unconfined Aquifer in Steady State with Isogeometric Method. 2018.
[24] Lal A, Datta B. Development and implementation of support vector machine regression surrogate models for predicting groundwater pumping-induced saltwater intrusion into coastal aquifers. Water Resources Management. 2018; 32(7): 2405-19.
Modares Civil Engineering journal
Volume 21, Issue 4
September and October 2021
Pages 75-88