Abstract—In this study, the patterns of specific humidity at 850 hPa during southwest monsoon over southern Thailand between the years 2000 to 2009 are investigated by self-organizing map (SOM). Learning rates and neighborhood functions are necessary parameters that influence the results. Bubble and Gaussian neighborhood functions and three learning rates (linear, inverse of time and power series) are analyzed by varying iteration. The quality of SOM is measured by the quantization error. The study finds that Gaussian function with linear learning rate gives the best result for pattern classification of the specific humidity.
Index Terms—Learning rates, neighborhood functions, self-organizing map (SOM), specific humidity.
W. Natita and W. Wiboonsak are with the Mathematics Department, King Mongkut’s University of Technology Thonburi, Bangkok, Thailand (e-mail: natita.wang@gmail.com, iwibhayu@kmutt.ac.th).
S. Dusadee is with the Joint Graduate School of Energy and Environment, King Mongkut’s University of Technology Thonburi, Bangkok, Thailand (e-mail: dusadee.suk@kmutt.ac.th).
[PDF]
Cite: W. Natita, W. Wiboonsak, and S. Dusadee, "Appropriate Learning Rate and Neighborhood Function of Self-organizing Map (SOM) for Specific Humidity Pattern Classification over Southern Thailand," International Journal of Modeling and Optimization vol. 6, no. 1, pp. 61-65, 2016.
