Expert System to Identify Risk Factors of Toddler’s Nutrition Status with Case Based Reasoning

Abstract

In 2012, Indonesia was the 5th most malnourished country in the world. This rank is affected by the population of Indonesia which ranked fourth in the world. Toddler malnutrition is a hot issue in Indonesia, and it is the basis of programs that supported by goverment to remedies these problems. The number of malnourished children in Indonesia is currently around 900 thousand people. The amount is 4.5 percent of the number of Indonesian children, which is 23 million people. For this reason it is important to predict the nutritional status of children so that preventive measures can be taken to reduce the number of malnutrition status in children in Indonesia. This study aims to apply the Modified K-Nearest Neighboar (M-KNN) method to identify risk factors for toddler nutritional status. The data used in this study is a combination of two types of data sources (primary and secondary data), where the data is obtained from posyandu in Malang. This study uses anthropometric assessment variables for weight and age. The steps taken include: data input, determination of the value of k, calculating the value of validity and the value of weight voting. Furthermore, to measure the performance of the proposed method, measurement is carried out by calculating the accuracy value of the predicted results. From the results of testing with variations in the value of k obtained an accuracy value of 75% using 295 nutritional status data of toddlers, with neighbors k which is the best value of k = 4.