Radfard Majid, Soleimani Hamed, Nabavi Samira, Hashemzadeh Bayram, Akbari Hesam, Akbari Hamed, Adibzadeh Amir
Research Center for Health Sciences, Institute of Health, Department of Environmental Health, School of Health, Shiraz University of Medical Sciences, Shiraz, Iran.
Department of Environmental Health, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran.
Data Brief. 2018 Sep 5;20:1462-1467. doi: 10.1016/j.dib.2018.08.205. eCollection 2018 Oct.
In this article the data of the groundwater quality of Aras catchment area were investigated for estimating the sodium absorption ratio (SAR) in the years 2010-2014. The artificial neural network (ANN) is defined as a system of processor elements, called neurons, which create a network by a set of weights. In the present data article, a 3-layer MLP neural network including a hidden layer, an input layer and an output layer had been designed. The number of neurons in the input and output layers of network was considered to be 4 and 1, respectively, due to having four input variables (including: pH, sulfate, chloride and electrical conductivity (EC)) and only one output variable (sodium absorption ratio). The impact of pH, sulfate, chloride and EC were estimated to be 11.34%, 72.22%, 94% and 91%, respectively. ANN and multiple linear regression methods were used to estimate the rate of sodium absorption ratio of groundwater resources of the Aras catchment area. The data of both methods were compared with the model׳s performance evaluation criteria, namely root mean square error (RMSE), mean absolute error (%) and correlation coefficient. The data showed that ANN is a helpful and exact tool for predicting the amount SAR in groundwater resources of Aras catchment area and these results are not comparable with the results of multiple linear regressions.
在本文中，对阿拉斯集水区2010 - 2014年的地下水水质数据进行了调查，以估算钠吸附比（SAR）。人工神经网络（ANN）被定义为一种处理器元件系统，称为神经元，它通过一组权重创建一个网络。在当前的数据文章中，设计了一个包含隐藏层、输入层和输出层的三层多层感知器（MLP）神经网络。由于有四个输入变量（包括：pH值、硫酸盐、氯化物和电导率（EC））且只有一个输出变量（钠吸附比），网络输入层和输出层的神经元数量分别被设定为4个和1个。pH值、硫酸盐、氯化物和电导率的影响估计分别为11.34%、72.22%、94%和91%。使用人工神经网络和多元线性回归方法来估算阿拉斯集水区地下水资源的钠吸附比速率。将两种方法的数据与模型的性能评估标准，即均方根误差（RMSE）、平均绝对误差（%）和相关系数进行比较。数据表明，人工神经网络是预测阿拉斯集水区地下水资源中SAR量的一种有用且准确的工具，并且这些结果与多元线性回归的结果不可比。
