Comparison of ARIMAX and multivariate LSTM model in predicting daily death toll in Yancheng City / 公共卫生与预防医学

Objective To compare the effects of Autoregressive Integrated Moving Average model-X (ARIMAX) and multivariate Long Short Term Memory Network (multivariate LSTM) in the prediction of daily total death toll in Yancheng City. Methods Based on total death toll data, meteorological data and air quality data from January 1st, 2014 to June 30th,2017 in Yancheng City, Jiangsu province, ARIMAX model and multivariate LSTM model were established to predict the daily total death toll from July 1st,2017 to July 14th,2017. RMSE, MAE and MAPE were used as evaluation indexes to compare the prediction effects of these two models. Results RMSE, MAE and MAPE of ARIMAX model and multivariate LSTM model were 20.742、15.094、9.921 and 47.182、35.863、19.633, respectively. Conclusion ARIMAX model is better than multivariate LSTM model to predict the daily death toll in Yancheng city.

Prediction and analysis of influenza-like illness and meteorological factors by ARIMAX model in Urumqi / 公共卫生与预防医学

Objective To analyze the influence of meteorological factors on the number of influenza-like illness (ILI) cases in Urumqi, Xinjiang, and establish an ARIMAX (autoregressive integrated moving average model-X) model to make short-term prediction of the number of ILI cases, so as to provide theoretical basis for the prevention and control of influenza in Urumqi. Methods The number of ILI cases in Urumqi from January 2015 to September 2017 and meteorological data of the same period were used to establish ARIMAX model and predict the number of ILI cases in Urumqi from October 2017 to March 2018. Results The ARIMA (0,1,1) (1,1,0)12 model was established from January 2015 to September 2017, AIC = 200.09. According to residual correlation function (CCF), there was a positive correlation between monthly average relative humidity and ILI cases, and a negative correlation between monthly sunshine hours and ILI cases. The average monthly relative humidity and monthly sunshine hours were taken as influencing variables to establish the ARIMAX model. Among them, the ARIMAX model incorporating the lagging order of 0 of monthly sunshine hours had the smallest AIC (AIC=197.63), and all parameters of the model were statistically significant. Compared with the univariate time series ARIMA model, the mean absolute percentage error (MAPE) of fitting was reduced by 1.3687%, the predicted MAPE was reduced by 5.25%, and the prediction accuracy was improved. Conclusion The ARIMAX model with meteorological factors established in this study can better predict the incidence trend of ILI cases in a short time, providing evidence for influenza surveillance and prevention and control.