Acceso abierto

Development of an adaptive neuro-fuzzy inference system (ANFIS) model to predict sea surface temperature (SST)

Semih Kale
Semih Kale
   | 26 nov 2020
Oceanological and Hydrobiological Studies's Cover Image
Acerca de este artículo
Artículo anterior
Artículo siguiente
Cite
Article Category: Research Article
Publicado en línea: 26 nov 2020
Páginas: 354 - 373
Recibido: 22 may 2020
Aceptado: 19 jun 2020
DOI: https://doi.org/10.1515/ohs-2020-0031
© 2020 Faculty of Oceanography and Geography, University of Gdańsk, Poland
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1

Location of the study area
Location of the study area

Figure 2

ANFIS structure
ANFIS structure

Figure 3

Comparison of ANFIS-SC and ANFIS-GP models for the observed and predicted SST values in the Çanakkale Strait
Comparison of ANFIS-SC and ANFIS-GP models for the observed and predicted SST values in the Çanakkale Strait

Figure 4

Scatter plots of ANFIS-SC and ANFIS-GP models for the observed and predicted SST values in the Çanakkale Strait
Scatter plots of ANFIS-SC and ANFIS-GP models for the observed and predicted SST values in the Çanakkale Strait

Figure 5

3D surface visualization of the relationship between the input and output for the ANFIS-SC4 model
3D surface visualization of the relationship between the input and output for the ANFIS-SC4 model

Figure 6

3D surface visualization of the relationship between the input and output for the ANFIS-GP4 model
3D surface visualization of the relationship between the input and output for the ANFIS-GP4 model

Figure 7

3D surface visualization of the relationship between the input and output for ANFIS-SC3 and ANFIS-GP3 models
3D surface visualization of the relationship between the input and output for ANFIS-SC3 and ANFIS-GP3 models

Figure 8

3D surface visualization of the relationship between the input and output for ANFIS-SC2 and ANFIS-GP2 models
3D surface visualization of the relationship between the input and output for ANFIS-SC2 and ANFIS-GP2 models

Performance of different ANFIS models in modelling SST (°C) for the Çanakkale Strait

ANFIS Model Stage MAE MSE RMSE MAPE NSE R-squared
Training 1.004 2.190 1.480 6.366 0.885 0.922
SC1 Validation 0.459 1.219 1.104 2.596 0.608 0.831
Total 1.463 3.409 1.846 8.963 0.846 0.846
Training 0.800 1.412 1.188 4.725 0.926 0.932
SC2 Validation 0.327 0.751 0.867 1.919 0.758 0.770
Total 1.127 2.163 1.471 6.644 0.902 0.902
Training 0.792 1.368 1.169 4.669 0.928 0.934
SC3 Validation 0.328 0.746 0.864 1.924 0.760 0.770
Total 1.120 2.114 1.454 6.594 0.904 0.905
Training 0.571 0.725 0.851 3.323 0.962 0.962
SC4 Validation 0.205 0.254 0.504 1.196 0.918 0.900
Total 0.776 0.978 0.989 4.519 0.956 0.956
Training 1.057 2.440 1.562 6.664 0.872 0.907
GP1 Validation 0.453 1.164 1.079 2.569 0.625 0.842
Total 1.510 3.604 1.898 9.233 0.837 0.838
Training 0.812 1.480 1.217 4.894 0.922 0.929
GP2 Validation 0.339 0.865 0.930 1.975 0.722 0.734
Total 1.152 2.346 1.532 6.869 0.894 0.894
Training 0.793 1.443 1.201 4.763 0.924 0.931
GP3 Validation 0.341 0.862 0.928 1.972 0.723 0.736
Total 1.134 2.305 1.518 6.735 0.896 0.896
Training 0.486 0.533 0.730 2.709 0.972 0.972
GP4 Validation 0.262 0.677 0.823 1.556 0.782 0.753
Total 0.748 1.210 1.100 4.265 0.945 0.946

Basic statistics of the measured factors in the Çanakkale meteorological observation station

Variable Mean Unit SE SD Maximum Minimum Correlation with SST
Air Temperature 18.90 °C 0.30 5.37 28.50 5.30 0.91
Evaporation 169.05 mm 4.23 75.32 366.60 29.50 0.75
Precipitation 33.73 mm 2.16 38.50 222.20 0.00 −0.48
SST 18.96 °C 0.26 4.71 26.70 7.50 1.00
eISSN:
1897-3191
Idioma:
Inglés
Calendario de la edición:
4 veces al año
Temas de la revista:
Chemistry, other, Geosciences, Life Sciences
RSS Feed de revista