Research Article

Application of Extreme Learning Machine Algorithm for Drought Forecasting

Table 8

Pearson’s correlation coefficient between the observed and predicted values of SPTI 1–12 month time scales using ELM, MLP, and ARIMA algorithms at selected meteorological stations.

ScaleAstoreChhorChitralKalatKohatMianwaliMultanMuzaffarabadSialkot
ELMMLPARIMAELMMLPARIMAELMMLPARIMAELMMLPARIMAELMMLPARIMAELMMLPARIMAELMMLPARIMAELMMLPARIMAELMMLPARIMA

SPTI-10.870.590.560.860.620.540.930.740.690.860.630.520.840.480.390.880.590.430.840.410.320.840.560.480.880.660.60
SPTI-30.960.880.860.950.860.830.970.930.900.940.860.840.920.830.800.950.850.800.910.780.750.930.850.820.940.860.83
SPTI-60.960.940.900.960.920.880.970.960.930.950.930.910.940.900.870.940.910.870.930.890.890.930.890.860.950.920.89
SPTI-90.960.940.910.960.280.840.970.960.920.970.950.930.950.770.890.960.950.920.940.910.860.950.930.890.950.940.91
SPTI-120.960.950.960.950.930.950.940.940.950.980.970.980.960.950.960.970.960.970.770.920.930.950.940.950.950.930.94