Application Of ARIMA - BP Combination Model In HIV / AIDS And Its Application In AIDS Forecasting In

Objective:(1)By analyzing the HIV/AIDS patient data of Shaoyang city from 2004 to 2013, it can concluded that the popular trend and characteristics of Shaoyang city AIDS in recent 10 years.(2)The application of ARIMA model combined with BP neural network model of Shaoyang city from 2005 to 2013 HIV/AIDS patient data were analyzed, and through the combination model compared with the predicted effect of single model, provide a reference for further research.Methods:(1)Export the test data of HIV/AIDS from 2004-2013 through the reporting system and entry it into the EXCEL database, then using the SPSS18.0 statistical software describe it.(2)The ARIMA and combination model are established based on the monthly incidence rate of HIV/AIDS in Shaoyang, comparing the accuracy of the models.Results:(1)2004-2013,Shaoyang has reported a total of 1515 cases of HIV/AIDS cases, including 882 cases of HIV/AIDS patients and 497 cases of death. In recent 10 years, HIV/AIDS incidence increased year by year, from 0.81/100 in 2004 increased to 2013 per 44.5/100, the average annual growth rate of 49.12%;(2)Infected mainly by the group of 20 to 49 in the majority, however, with the passage of time, the 20- age group the proportion fell from 83.3% to 49.4%, from 16.7% to 47.2% in the 50- age group;(3)Component ratio is the highest professional vocational distribution more extensive, crowd for farming, accounting for 51% of all HIV/AIDS patients, and the proportion is rising trend year by year, by 2013 the professional composition ratio reached 64.1%,(4)Three approaches in Shaoyang city AIDS were reported, including sexual transmission as the main route of transmission, accounting for 79.4%, and sex transmitted predominantly unmarried of the opposite sex contact, account for sexually transmitted infection of the patients of 83.8%.(5)The relative error of the following models are 26.25% &39.89%, the infinitive coefficients are 0.786&0.852.Conclusion:(1)From 2004 to 2013 the occurrence number and incidence of the HIV/AIDS in Shaoyang increased year by year; 20- group was the major group of patients with HIV/AIDS in Shaoyang, but in recent 10 years, 50- group rapid growth; the major occupational structure of patients with HIV/AIDS in Shaoyang was farming; sexual contact become the major infection pathway of patients with HIV/AIDS in recent 10 years.(2)The ARIMA model of a single prediction for HIV/AIDS has larger limitations, not well guarantee the prediction effect, and after the introduction of the BP neural network, it can be directly to the adjustment of the ARIMA model to forecast the result. Due to the strong nonlinear mapping, respond to data capability is strong, and have achieved good prediction effect.