Analisis Laju Pembelajaran (a) pada Self Organizing Maps (SOM) dalam Pengenalan Karakter

Abstract

Self-organizing Maps is one of the methods in Artificial Neural Networks are able to learn on their own and are often used for pattern recognition or image .. The most decisive in the learning process Self Organizing Maps is alpha/learning rate a (t). This multiplication factor determining the speed of learning networks. But unknown to what extent the influence of alpha/learning rate a (t) is the ANN Self Organizing Maps to achieve convergent. The analysis is performed on the character of letters that make up words. Analysis showed that the influence of alpha/learning rate a (t) is the ANN Self Organizing Maps are very large, where the value of the variable alpha/learning rate a (t) is given very predispose ANN Self Organizing Maps achieve convergent, At the time value of the variable alpha/learning rate a (t) is close to zero then it will converge more quickly achieved whereas if the value of the variable alpha /learning rate a (t) is close to I converges difficult to achieve.