Predicting the number of neonatal deaths in Thailand using grey system theory

Abstract

The infant mortality rate is an indicator of a country's level of development. Forecasting can be used as a guideline for health resource administration. This research aimed to predict the number of neonatal deaths in Thailand by health region in fiscal year 2024 using Grey System Theory. Data was collected to develop the model from the statistics report on the number of neonates who died less than or equal to 28 days after birth from the fiscal year 2014 to 2024 from the Ministry of Public Health (processing date: March 9, 2024). The results showed that the GM (1,1) models had a development coefficient (a) of nine health regions between , the models were suitable for short-term forecasts. The improved GM (1,1) expanded with a periodic correction model had a lower mean absolute percentage error (MAPE) than the GM (1,1) model in every health region. The MAPE in all health region was less than 10 in the criteria for high accuracy, except for 1, 8, and 13, which were between 10 and 20 in the good prediction criteria. Predicting the number of neonatal deaths in the fiscal year 2024, every health region decreased from the fiscal year 2023, except health region 8, 10, and 11, which increased by 3.89, 6.97, and 6.80, respectively. The proportion of the number of neonatal deaths for fiscal year 2024 (processed only up to March 9, 2024) as a percentage of the predicted value in every health region was less than 43.99 percent, of which the proportion by the number of days in fiscal year 2024, except health region 1, 3, 6, and 12, accounted for 57.65, 46.54, 46.03, and 49.20 percent, respectively.