Fig 1.
Schematic diagram of RBFNN architecture.
Fig 2.
Flowchart of the differential search algorithm.
Fig 3.
Flowchart of the particle swarm optimization algorithm.
Fig 4.
Flowchart of the proposed PSODS algorithm.
Table 1.
Description of public datasets.
Fig 5.
RMSE for Boston House Pricing for by varying hidden layer neurons.
Fig 6.
RMSE for Concrete Compressive strength by varying hidden layer neurons.
Fig 7.
RMSE for Airfoil self—noise by varying hidden layer neurons.
Fig 8.
RMSE for Istanbul Stock Exchange by varying hidden layer neurons.
Fig 9.
RMSE for Forest Fires by varying hidden layer neurons.
Fig 10.
RMSE for Abalone by varying hidden layer neurons.
Fig 11.
RMSE for Auto MPG by varying hidden layer neurons.
Table 2.
Summary of results obtained for Boston House Pricing.
Table 3.
Summary of results obtained for Concrete Compressive strength.
Table 4.
Summary of results obtained for Airfoil self -noise.
Table 5.
Summary of results obtained for Istanbul Stock Exchange.
Table 6.
Summary of results obtained for Forest Fires.
Table 7.
Summary of results obtained for Abalone.
Table 8.
Summary of results obtained for Auto MPG.
Fig 12.
a) Convergence plot b) Successfully predicted samples for Boston House Pricing.
Fig 13.
a) Convergence plot b) Successfully predicted samples for Concrete strength.
Fig 14.
a) Convergence plot b) Successfully predicted samples for Airfoil self -noise.
Fig 15.
a) Convergence plot b) Successfully predicted samples for Istanbul Stock Exchange.
Fig 16.
a) Convergence plot b) Successfully predicted samples for Forest Fires.
Fig 17.
a) Convergence plot b) Successfully predicted samples for Abalone.
Fig 18.
a) Convergence plot b) Successfully predicted samples for Auto MPG.
Fig 19.
Normalized RMSE statistics using PSODS for % increase in testing samples.
Table 9.
Summary of results obtained for wind speed.
Fig 20.
a) Convergence plot b) Successfully predicted samples for wind speed.
Fig 21.
Normalized RMSE statistics for wind speed for % increase in testing samples.
Table 10.
Summary of test RMSE obtained using different neural networks.
Table 11.
Summary of successful prediction of samples.
Table 12.
Performance comparison of population initialization algorithm for PSODS-RBFNN.