Fig 1.
Trend of syphilis incidence in China from 2005–2012 by year
Table 1.
Syphilis incidence seasonal indices.
Fig 2.
Seasonal indices of each type of syphilis
Table 2.
Linear regression model for each syphilis time series removed seasonality.
Table 3.
Available ARIMA models fitted for each syphilis series.
Fig 3.
Cross correlation analysis between primary syphilis and other syphilis time series (highest values appear when the lag is 0)
Fig 4.
Cross correlation analysis between secondary syphilis and other syphilis time series (highest values appear when the lag is 0 except between secondary and congenital syphilis).
Table 4.
The highest cross-correlation between syphilis incidence time series (the values in the brackets are the lags, the lags are 0 except between secondary and congenital syphilis).
Table 5.
Estimation of available ARIMAX models for each series.
Fig 5.
Primary syphilis incidence fitting and testing performance by ARIMA and ARIMAX (U95 and L95 refer to the upper and lower 95% confidential interval respectively.
The vertical gray line separates modelling from estimates.)
Fig 6.
Secondary syphilis incidence fitting and testing performance by ARIMA and ARIMAX (U95 and L95refers to the upper and lower 95% confidential interval respectively.
The data were divided into modeling and forecasting groups with a vertical line; the left is the modeling part, and the right is the forecasting part.)
Fig 7.
Tertiary syphilis incidence fitting and testing performance by ARIMA and ARIMAX (U95 and L95refers to the upper and lower 95% confidential interval respectively.
The data were divided into modeling and forecasting groups with a vertical line; the left is the modeling part, and the right is the forecasting part.)
Fig 8.
Latent syphilis incidence fitting and testing performance by ARIMA and ARIMAX (U95 and L95refers to the upper and lower 95% confidential interval respectively.
The data were divided into modeling and forecasting groups with a vertical line; the left is the modeling part, and the right is the forecasting part.)
Fig 9.
Congenital syphilis incidence fitting and testing performance by ARIMA and ARIMAX (U95 and L95refers to the upper and lower 95% confidential interval respectively.
The data were divided into modeling and forecasting groups with a vertical line; the left is the modeling part, and the right is the forecasting part.)
Table 6.
Comparison of the performances of the ARIMA model and ARIMAX model.