Child Health Detection System Along with the Recommendations Using K-Nearest Neighbor Artificial Intelligence

Background: Children under 5 years of age were very vulnerable to disease. Therefore, child health is one of the main focuses in the health sector. The main triggers for under-five mortality were preventable and treatable diseases. There were 5 diseases mostly experienced by under-five children, namely pneumonia, diarrhea, malaria, measles and malnutrition. In Indonesia, child health detection is still performed manually by midwives using the IMCI cards, but during the Covid-19 pandemic, child health detection is not actively performed. The K-Nearest Neighbor Artificial Intelligence System can be used for child health detection.
Objective: To develop a child health detection system along with the recommendations using k-nearest neighbor artificial intelligence.
Methods: Research and Development (RnD) and system accuracy test. The study subjects involved secondary data of 1000 IMCI register data. Data were tested using the intra-class correlation coefficient and the K-Nearest Neighbor algorithm.
Results: The results of the child health detection system using k-nearest neighbor artificial intelligence obtained a system accuracy rate of 100% and the sick child health recommendation system using k-nearest neighbor artificial intelligence obtained a system accuracy rate of 89.3%. It was proven that k-nearest neighbor artificial intelligence was able to develop a child health detection system along with the recommendations which can be used by parents.