Abstract: Objective:To reduce the mortality of children under 5 years of age,to understand the cause of death,and to analyse the monitoring data.
Methods:Describe the trends in mortality,perform analysis of death causes,and use the grey model［GM (1,1)］ and BP artificial neural network to predict mortality.
Results:The total mortality in children under five years old was decreased,being lower in the recent 4 years,with a certain degree of volatility and occasional increases.The main cause of death was premature low birth weight,congenital heart disease and other congenital anomalies,with the death due to injury ranking higher.The average error rate of the BP artificial neural network method was 2.68%,which was much lower than that of the grey model(14.06%).
Conclusion:We should strengthen prevention and health care work in premature low weight,congenital heart disease and other congenital anomalies and pay attention to prevention of child injury in order to ensure the existing achievement of mortality decline.BP artificial neural network method is better than the grey model method.

Key words: Death, Analysis of cause of death, Monitoring, Children