Abstract:

The GM(1,1) grey method was applied to predict the tuberculosis epidemic in 2016-2021 where the national tuberculosis epidemic data in 2008-2015 obtained from the China Information System of Disease Prevention and Control was used as the initial input data. The Self Organizing Maps (SOM) neural network method was used for cluster analysis on the prediction results. According to the gray prediction model and the data related to tuberculosis epidemic in 2008-2015, the incidence and incidence rates of national and 31 provinces in China in 2016-2021 would be predicted. Through SOM neural network cluster analysis, 31 provinces could be divided into 4 levels with high to low emphasis. Class Ⅰ center included Qinghai, Tibet and Xinjiang, Class Ⅱ center included Heilongjiang, Hunan, Guangxi, Hainan and Guizhou, Class Ⅲ center included Liaoning, Anhui, Jiangxi, He’nan, Hubei, Guangdong, Chongqing, Sichuan, Yunnan and Shaanxi, Class Ⅳ center included Beijing, Tianjin, Hebei, Shanxi, Inner Mongolia, Jilin, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Gansu and Ningxia. According to the gray prediction, the tuberculosis epidemic in the next six years shows a downward trend. Control strategies can be timely adjusted and control forces can be allocated for different types of areas according to the results of clustering.

Key words: Tuberculosis,pulmonary, Information systems, Models,statistical, Neural networks (computer), Forecasting, Cluster analysis