Prediction of Nutritional Status of Toddlers Using C4.5 Algorithm

Abstract

In developing countries, chronic malnutrition leads to stunting. Stunting will become a public health problem in Indonesia resulting in a decline in the quality of Indonesia's human resources in the future. In the modernization era, determining the nutritional status of toddlers can be simplified automatically. The author's C4.5 algorithm was used to classify and predict the nutritional status of toddlers in Bringin sub-district, in addition to testing using the system that the author built, the author also conducted manual testing and testing using RapidMiner as a comparison. Based on the analysis of the results of 186 datasets tested using this system, predictions were made of toddlers with male gender categories, very underweight and height predicted that the toddlers were in the malnutrition category. For the performance of the dataset tested obtained an accuracy of 84.221%, there was a difference of 6% from the test results using RapidMiner which obtained an accuracy of 91.40%. The use of algorithms is highly recommended to classify the nutritional status of toddlers. The application of the toddler nutritional status classification system that the author designed is feasible because it can be faster and more effective in classifying the nutritional status of toddlers based on the zscore set and there is only a difference of 6% from the same dataset prediction testing using rapid miner. Based on the dataset of system test results, it can be concluded that 89.24% of toddlers in Bringin sub-district are well nourished.