Artificial Intelligence Hybrid-Deep Learning Model for Groundwater Level Prediction Using MLP-ADAM

Groundwater is the largest storage of freshwater resources, which serves as the major inventory for mostof the human consumption through agriculture, industrial, and domestic water supply. In the fields ofhydrological, some researchers applied a neural network to forecast rainfall intensity in space-time andintroduced the advantages of neural networks compared to numerical models. Then, researches have beenconducted applying data-driven models. Some of them extended an Artificial Neural Networks (ANNs)model to forecast groundwater level in semi-confined glacial sand and gravel aquifer under variable state,pumping extraction, and climate conditions with significant accuracy. In this paper, a Multi-LayerPerceptron (MLP) model is applied to simulate groundwater level. The adaptive moment estimationoptimization algorithm (ADAM) is also used in this matter. This results in a hybrid MLP-ADAM modelwhereas the hyper-parameters of the MLP model are optimized with the ADAM method. Also, the casestudy aquifer is located in Najafabad plain, Gavkhoni catchment consists of ۲۱ basins. By applying thestated hybrid model for this region, the root mean squared error, mean absolute error, mean squared errorand the coefficient of determination ( ۲R ) are used to evaluate the accuracy of the simulated groundwaterlevel. The total value of ۲R and RMSE are ۰.۹۴۵۸ and ۰.۷۳۱۳ respectively which are obtained from themodel output. Results indicate that deep learning algorithms can demonstrate a high accuracy prediction.Although the optimization of parameters is insignificant in numbers, due to the value of time in modelingsetup, it is highly recommended to apply an optimization algorithm in modeling.

کلیدواژه ها:

Hybrid deep learning model ، Groundwater ، MLP ، ADAM