Comparative Assessment of Improved SVM Method under Different Kernel Functions for Predicting Multi-scale Drought Index

Autor: Chaitanya B. Pande, N. L. Kushwaha, Israel R. Orimoloye, Rohitashw Kumar, Hazem Ghassan Abdo, Abebe Debele Tolche, Ahmed Elbeltagi
Rok vydání: 2023
Předmět:
Zdroj: Water Resources Management. 37:1367-1399
ISSN: 1573-1650
0920-4741
Popis: Precise assessment, monitoring and forecasting of drought phenomena are crucial and play a vital role in agriculture and water resources management in the semi-arid region. In this study, Standardized Precipitation Index (SPI) was used to predict the drought in the upper Godavari River basin, India. Ten combinations were used to predict three SPI timescales (i.e., SPI − 3, SPI-6, and SPI-12). The historical data of SPI from 2000 to 2019 was divided into training (75% of the data) and testing (25% of the data) models for SPI prediction. The best subset regression method and sensitivity analysis were applied to estimate the most effective input variables for estimation of SPI 3, 6, and 12. The improved support vector machine using sequential minimal optimization (SVM-SMO) with various kernel functions i.e., SMO-SVM poly kernel, SMO-SVM Normalized poly kernel, SMO-SVM PUK (Pearson Universal Kernel) and SMO-SVM RBF (radial basis function) kernel was developed to estimate the SPI. The results were compared and analyzed using statistical indicators i.e., root mean square error (RMSE), mean absolute error (MAE), relative absolute error (RAE), root relative squared error (RRSE), and correlation coefficient (r). The main results showed that the SMO-SVM poly kernel model precisely predict the SPI-3 (R2 = 0.819) and SPI-12 (R2 = 0.968) values at Paithan station; the SPI-3 (R2 = 0.736) and SPI-6 (R2 = 0.841) values at Silload station, respectively. The SMO-SVM PUK kernel showed superiority in the prediction of SPI-6 (R2 = 0.846) at Paithan station and SPI-12 (R2 = 0.975) at the Silload station. The competition between SVM-SMO poly kernel and SVM-SMO PUK kernel was observed in the prediction of long setting time (i.e. SPI-6 and SPI-12), while SVM-SMO poly kernel is superior in the estimation of SPI-3 at both stations. The results of the study showed the efficacy of the SVM-SMO algorithm with various kernel functions in the estimation of multiscale SPI and can be helpful in decision making for water resource management and tackle droughts in the semi-arid region of central India.
Databáze: OpenAIRE
Externí odkaz: