Skip to main content
. 2018 Nov 22;4(4):e10827. doi: 10.2196/10827

Table 1.

Mutual Shannon information entropy, Autoregressive Integrated Moving Average or Autoregressive Integrated Moving Average with External Variable model parameters, and Akaike Information Criteria values in different quarters of 2016.

Quarters Original + Case Retweeting without commenting + Case Reply + Case
Q1

Mutual Info 0.04 0.01 0.09
ARIMA(X)a Par 2, 0, 3 2, 1, 3 2, 0, 2
dAICb –2.25c
(976.61, 974.36)
–1.88c
(1341.51, 1339.63)
–1.21c
(950.05, 948.84)
Q2

Mutual Info 0.13 0.17 0.29
ARIMA(X) Par 2, 1, 3 2, 1, 3 0, 1, 1
dAIC 0.96
(722.54, 723.50)
–0.88c
(1207.14, 1206.26)
1.88
(709.18, 711.06)
Q3

Mutual Info 0.02 0.08 0.02
ARIMA(X) Par 1, 1, 1 2, 1, 2 2, 1, 2
dAIC 1.95
(719.51, 721.46)
1.82
(1172.01, 1173,83)
–0.62c
(738.76, 738.14)
Q4

Mutual Info 0.01 0.07 0.01
ARIMA(X) Par 2, 0, 3 0, 1, 2 0, 0, 1
dAIC –0.59c
(453.28, 452.69)
1.62
(917.84, 919.46)
1.97
(353.23, 355.20)

aARIMA(X): Autoregressive Integrated Moving Average (with External Variable).

bdAIC: difference in Akaike information criterion.

cNegative dAIC value indicates better performance of the ARIMAX model compared with its corresponding ARIMA model; hence, including Zika case counts improves the model performance.