Prediction of new cases of HIV infection and AIDS in Chongqing: data fitting and comparative analysis of 3 prediction models

• Main Authors: XIANG Ying, YU Hongyue, ZHANG Wei
• Format: Article
• Language: zho
• Published: Editorial Office of Journal of Third Military Medical University 2019-02-01
• Series: Di-san junyi daxue xuebao
• Subjects: human immunodeficiency virus; acquired immune deficiency syndrome; epidemic situation; prediction; data fitting
• Online Access: http://aammt.tmmu.edu.cn/Upload/rhtml/201810141.htm

Objective To establish prediction models for the total number of newly reported cases of HIV infection and AIDS in Chongqing and assess the performance of these models using the data in the recent 2 years. Methods The ARIMA model, exponential smoothing method and trend extrapolation method were used to fit the data of the total number of newly reported cases of HIV infection and AIDS in Chongqing collected from 2006 to 2015. The number of newly reported cases in 2016 and 2017 were used to verify the performance of the 3 prediction models to establish the optimal model for epidemic prediction of AIDS in Chongqing. Results Based on the results of model identification and diagnosis, 3 time series models, namely ARIMA (0, 1, 1) (0, 1, 1) model, simple seasonal exponential smoothing model and quadratic curve model, were constructed to fit and predict the total number of new HIV/AIDS cases in Chongqing. The standard errors of the 3 models for fitting the epidemic trends were 62.30, 45.51 and 227.29, respectively, and their average relative errors for predicting the total number of new cases in the next two years were 4.51%, 2.14% and 12.74%, respectively. Conclusion The 3 prediction models have different fitting effects and performance in predicting HIV/AIDS epidemic in Chongqing. Among them, the simple seasonal exponential smoothing method can more accurately fit and predict the total number of new HIV/AIDS cases in Chongqing.