OPTIMASI KLASIFIKASI STATUS GIZI BALITA BERDASARKAN INDEKS ANTROPOMETRI MENGGUNAKAN ALGORITMA NAIVE BAYES CLASSIFICATION ADABOOST

  • Achmad Ridwan
  • Catur Supriyanto
  • Pulung Nurtantio Andono
Keywords: Toddler Nutritional Status, Anthropometric, Classification, NBC, Adaboost

Abstract

Body Mass Index (BMI) is commonly used as a measure to assess the nutritional status of infants. If there are two babies whose weight and height are the same may have different nutritional status. If it happens then the use BMI to measure nutritional status less relevant. Anthropometric measurement tool to be very instrumental for determining the nutritional status. The guidelines for determining the nutritional status Anthropometric parameters are selected and recommended which includes an assessment of the age, weight, height. On the contrary, along with the development of technology, increasing the amount of data that requires some methods to process and draw conclusions from such data and information. NBC algorithm is an algorithm of decision tree method has good performance in dealing with the classification of Toddler Nutritional Status based index Anthropometry, but NBC has a weakness in the class imbalance. Adaboost one boosting methods that could reduce imbalances class by giving weight to the level of classification Error which may alter the distribution of data. The use of Adaboost with reason this method can improve the accuracy in the process of classification and prediction by means generate a combination of a model, select the model that has the greatest weight. These experiments will apply the NBC algorithm used for classification of Toddler Nutritional Status based index Anthropometry and will be increased again by Adaboost method for being able to overcome the imbalance class thus increasing the probability value of each class and improve accuracy, it also lowers Error Classification. While that would be classified are five classes: normal, fat, very fat, thin, or very thin. The results of the experiment were obtained from NBC method to an accuracy of 88.60% and a classification Error of 11.40%, while the method by Adaboost (NBC + Adaboost) to an accuracy of 88.84% and 11.16% of the classification Error. So we can conclude NBC with Adaboost algorithm implementation on the Classification of Toddler Nutritional Status based index Anthropometry proved capable of overcoming the class imbalance and improve accuracy also lowers Error Classification.

Published
2019-09-26