<![CDATA[Immune systems owned by children who are weaker than adults make children more susceptible to disease. Skin disease is one of them, this is because the skin is the sense of touch for humans. The similarity of symptoms of any skin disease makes the layman difficult to distinguish the illness in suffering whereas every type of disease has a different treatment. In this study implements artificial neural network method backpropagation to study the past data in order to diagnose skin diseases in children. The input used in the form of symptoms of all diseases amounted to 19 then represented into binary 0 and 1 where the value will be worth 1 if experiencing the symptoms and vice versa. The activation function used is sigmoid binner. The initial weights are obtained using Nguyen-Widrow which will then be done by repeatedly learning so that the result of the network that gives the correct response to the input. Based on the result of the test, the optimal parameters are 4 hidden neurons, learning rate 0.4 and epoch maximum 300000 and The results of the accuracy of the study reached 87.22% which indicates that this backpropagation method can be used in diagnosing skin diseases in children.]]>
Copyrights © 2017
