Classification of nutritional status of toddlers using fuzzy k-nearest neighbor in every class (FK-NNC)

Abstract

The purpose of this research is classify nutritional status of toddlers with algorithm fuzzy k-nearest in every class (FK-NNC). The FK-NNC algorithm is a modification concept of the nearest K neighbor for each class. The largest of the class on membership value will be selected as prediction classes. The optimal K value in FK-NNC algorithm uses k-fold cross validation. The optimal K values are searched by experimenting 1-fold cross validation, 4-fold cross validation and 10-fold cross validation. Accuracy rate of classify nutritional status of toddlers at the Wonorejo Health Center using the FK-NNC algorithm with an optimal K value. The results of this research obtained the optimal K value used in the FK-NNC algorithm at the Wonorejo Health Center is K=8 with 1-fold cross validation experiment. The value of K=8 was applied to the FK-NNC algorithm with a 1-fold cross validation experiment to predict class of nutritional status of toddlers in the Wonorejo Health Center. The percentage of accuracy produced as much as 100% with data proportion 90:10.