Author/Authors :
Javan, S Environmental Health Department - Medical Sciences Faculty - Neyshabur University of Medical Sciences, Neyshabur , Gholamalizadeh Ahangar, A Soil Sciences Department - Soil & Water Engineering Faculty - Zabol University, Zabol , Hassani, A.H Environmental Engineering Department - Environment & Energy Faculty - Tehran Science & Research Branch - Islamic Azad University, Tehran , Soltani, J Water Engineering Department - Water Engineering Faculty - Abureyhan Campus - University of Tehran, Tehran
Abstract :
Aims Artificial Neural Networks (ANNs) are powerful tools that are commonly used today
in prediction deposit-related sciences. The research aimed at predicting various five links of
heavy metals using the properties of deposit.
Materials & Methods 180 samples of surface sediments were taken from the Chahnimeh
reservoir and they were transferred to lab under standard conditions. Total Zinc concentration,
deposit properties and Zinc five bonds with deposit were measured. Efficiency of the ANN and
Multi-Layer Perceptron (MLP) model were evaluated to estimate the Zn bonds following the
measurement of parameters in the laboratory.
Findings Five links were predicted with the aid of ANNs and MLP model. Deposit properties
and total concentrations of heavy metals were considered as input and each of bonds were
considered as output
Conclusion Ultimately, the ANN showed good performance in the predicting the determination
of coefficients or R2 (0.98 to 1) and root mean square error or RMSE (0.7 to 0.01).
Keywords :
Artificial Neural Networks , Heavy Metals , Sediment Pollution , Chahnimeh
