4.7 Article

Prediction of total dissolved solids, based on optimization of new hybrid SVM models

Journal

Publisher

PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.engappai.2023.106780

Keywords

Ensemble learning; Evolutionary algorithm; Machine learning; River water quality; Shannon entropy

Ask authors/readers for more resources

Accurate monitoring of water quality is crucial in arid and semi-arid countries like Iran, with Total Dissolved Solids (TDS) playing a significant role. This study developed several hybrid models to predict TDS in Babolrood River, Iran, using monthly data from 1968 to 2016. The results showed that the SVM-TLBO5 model improved predictions compared to the LS-SVR1 model, with improved MAE and SI values at different stations.
Accurate monitoring of water quality is of great importance, especially in arid and semi-arid countries such as Iran. The Total Dissolved Solids (TDS) plays quite a significant role in rivers water quality. In most studies sampling isn't considered due to difficulty of measuring elements its time-consuming and expensive nature. Herein several hybrid models including SVM-CA, SVM-HS and SVM-TLBO were developed to predict TDS in Babolrood River, Iran. The monthly measured and unpublished data of Ca, Mg, HCO3, Na, SO4, Cl, pH, and TDS, from 1968 to 2016 were used. Based on Shannon's entropy and correlation matrix approaches most influential inputs were identified in five scenarios. Results were analyzed using several statistical indicators, including SI, MAE, U95, R2, RMSE and Taylor diagram. SVM-TLBO5 model improved MAE by 66% and 81% compared to LS-SVR1 model at Quran-Talar and Koushtargah stations, respectively. Based on SI, SVM-TLBO5 model improved predictions 87% and 79% compared to LS-SVR1 at Quran-Talar and Koushtargah stations. RMSE, MAE, SI, R2, U95, and T-STAT at Quran-Talar station obtained equal 10.1 mg/l, 0.022, 35.14, 10.22 mg/l, 0.022 and 33.59. Also, the same values for the Koushtargah station obtained equal to 6.73 mg/l, 0.993, 1.56, 4.51 mg/l, 0.997 and 0.674.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available