Evaluation of Trace Element Concentrations in Groundwater and Classification of Endemic Disease Regions using Multilayer Perceptron Neural Network

Sahmurova A., Kilic N., Okan I., Karaca F., Ucan O. N.

JOURNAL OF RESIDUALS SCIENCE & TECHNOLOGY, vol.6, no.2, pp.83-88, 2009 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 6 Issue: 2
  • Publication Date: 2009
  • Journal Name: JOURNAL OF RESIDUALS SCIENCE & TECHNOLOGY
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.83-88
  • Istanbul University Affiliated: Yes

Abstract

In this study, trace elements were measured in the groundwater in Azerbaijan and the level of the fluoride was assessed. The endemic diseases in the regions of Azerbaijan were investigated by using these data. A Multilayer Perceptron Neural Network (MLPNN) was used to classify the regions with or without an endemic disease. MLPNN employing a backprobagation training algorithm was used to predict the presence or the absence of endemic disease potential in the regions. At the end of the classification process, percentages of the towns with or without an endemic disease were calculated as 100% and 68.75% respectively. Total classification accuracy of MLPNN was determined as 75%. Therefore, we can conclude that a MLPNN is one of the most promising methods for classification of regions with endemic diseases, based on the trace elements in the groundwater.