By using 3-layer DBN to extract features embed in original image data, then fed the features into a extreme learning machine(ELM) for classification. Due to the difficulty in hiiden-nodes selection, then PSO was applied to select it automatically ,and PSO aimed to minimize the fitness function that is the accuracy of DBN-ELM crossvalidation classifcition
quan1995417 / pso-dbn Goto Github PK
View Code? Open in Web Editor NEWThis project forked from byq-luo/pso-dbn
By using 3-layer DBN to extract features embed in original image data, then fed the features into a extreme learning machine(ELM) for classification. Due to the difficulty in hiiden-nodes selection, then PSO was applied to select it automatically ,and PSO aimed to minimize the fitness function that is the accuracy of DBN-ELM crossvalidation classifcition