Font Size: a A A

A Study On Statistical Forecasts Of Maternal Mortality Rate, Infant Mortality Rate And Child Mortality Rate Under Age 5 By 2010 In China

Posted on:2006-02-26Degree:MasterType:Thesis
Country:ChinaCandidate:Y Y LiuFull Text:PDF
GTID:2144360155973557Subject:Epidemiology and Health Statistics
Abstract/Summary:PDF Full Text Request
Objective "National Plan of Action for Women Development of China" (2001-2010) and "National Plan of Action for Children Development of China (2001-2010)"(hereinafter "Two NPAs"), issued by the State Council of China in May 2001, put forward that the Maternal Mortality Rate (MMR), Infant Mortality Rate(IMR) and Child Mortality Rate under age 5(U5MR) in China"(hereinafter 'Three Mortalities') will have respectively dropped to 1/4 and 1/5 of those of the year of 2000 by the year of 2010. In May, 2002, the UN Special Session on Children launched the global goals in the document of "A World Fit For Children" with the reduction of MMR, IMR and U5MR dropped by 1/3 from 2000 to 2010, Which is the so-called Millennium Development Goals (MDGs). Entrusted by the National Working Committee on Children and Women under the State Council and the United Nations Children' Fund, from a statistical perspective, this study is to forecast the possibility of the MDGs to be realized by the year of 2010 in China.Methods Based on the data of Three Mortalities collected from 116 maternal and child health surveillance sites form 1990 to 2003, Time series analysis (Auto Regressive Integrated Moving Average model, ARIMA), Grey system forecasting (Grey Dynamics Model, GM), Simple average combined forecasts and the Combined forecasts based on radial basis function artificial neural network (RBFNN) were used to fit the dataand the prediction results were compared by some statistical indexes. Results In general, the year from 1990 to 2003 saw a slowing-down drop of the Three Mortalities in China, and the Three Mortalities of rural areas are higher than those in cities, and the jumping-off areas ranks on the top, then closely followed by inland areas, with the lowest rates for coastal areas. According to the comparison results of all the models involved, the Combined forecasts based on RBFNN is much better than other three models, and the results show that IMR and U5MR in cities will drop by 30.49% and 19.66% respectively from 2000 to 2010. While Simple average combined forecasts runs the least risk, which shows that U5MR in cities will drop by 18.84% from 2000 to 2010. The results imply that the goals of IMR and U5MR in cities may not be reached, but at the average level of whole country, the MDGs of Three Mortalities may hopefully be achieved by the year of 2010 in China.Conclusion The prediction results of Three Mortalities are comparatively more precise by fitting the Combining forecasts model based on RBFNN, and averagely speaking, from the entire country, by the year of 2010, MDGs of Three Mortalities in China may hopefully be achieved, while there are still some difficulties in urban areas in terms of IMR and U5MR.
Keywords/Search Tags:Statistical forecasting, ARIMA, GM (1,1), Combined forecasts, Simple average, radial basis function artificial neural network, MMR, IMR, U5MR
PDF Full Text Request
Related items