Download PDFOpen PDF in browser

Comparison of Accuracy of Artificial Neural Network (ANN) and Kriging Methods for Estimating Chlorine Concentration

8 pagesPublished: September 20, 2018

Abstract

Groundwater is one of the major sources of fresh water. Maintenance and management of this vital resource is so important especially in arid and semi-arid regions. Reliable and accurate groundwater quality assessment is essential as a basic data for any groundwater management studies. The aim of this study is to compare the accuracy of two Artificial Neural Network (ANN) and Kriging methods in predicting chlorine in groundwater. In case of ANN, we created an appropriate emulator, which minimize the prediction error by changing the parameters of the neural network, including the number of layers. The best Kriging model is also obtained by changing the variogram function, such that the Gaussian variogram has the least error in interpolation of the amount of chlorine. To evaluate the accuracy of these two methods, the mean square error (MSE) and Coefficient of determination (R2) are used. The data set consists of the amount of chlorine, in a monthly basis, measured at 112 observation wells from 1999 to 2015 in aquifer Qazvin, Iran. MSE values for ANN and Kriging are 14.8 and 15.4, respectively, which indicate that the ANN has a better performance and is more capable of predicting chlorine values in comparison with Kriging.

Keyphrases: Artificial Neural Network, Groundwater, Kriging methods, pollution, prediction, quality control

In: Goffredo La Loggia, Gabriele Freni, Valeria Puleo and Mauro De Marchis (editors). HIC 2018. 13th International Conference on Hydroinformatics, vol 3, pages 772--779

Links:
BibTeX entry
@inproceedings{HIC2018:Comparison_of_Accuracy_of,
  author    = {Faezeh Ghaleh Navi and Hamed Mazandarani Zadeh and Dragan Savic},
  title     = {Comparison of Accuracy of Artificial Neural Network (ANN) and Kriging Methods for Estimating Chlorine Concentration},
  booktitle = {HIC 2018. 13th International Conference on Hydroinformatics},
  editor    = {Goffredo La Loggia and Gabriele Freni and Valeria Puleo and Mauro De Marchis},
  series    = {EPiC Series in Engineering},
  volume    = {3},
  pages     = {772--779},
  year      = {2018},
  publisher = {EasyChair},
  bibsource = {EasyChair, https://easychair.org},
  issn      = {2516-2330},
  url       = {https://easychair.org/publications/paper/s6C8},
  doi       = {10.29007/lpmh}}
Download PDFOpen PDF in browser