Asymmetric Effects of Weather-Integrated Human Brucellosis Forecasting System Using a New Nonlinear Autoregressive Distributed Lag Model
Table 2
Long- and short-term estimates of the best NARDL and ARDL methods.
NARDL model
ARDL model
Variable
Coefficient
P
Variable
Coefficient
P
Long-run estimate
Long-run estimate
ΔAT(+)
0.018
0.603
ΔAT
0.081
<0.001
ΔAT(−)
−0.001
0.973
ΔAP
−0.0004
0.677
ΔAP(+)
−0.0004
0.696
ΔASH
0.001
0.530
ΔAP(−)
−0.0004
0.693
ΔAWV
1.449
0.022
ΔASH(+)
0.002
0.376
ΔARH
0.030
0.004
ΔASH(−)
0.002
0.290
ΔAAP
−0.084
0.012
ΔAWV(+)
0.738
0.044
Short-run estimate
ΔAWV(−)
0.875
0.031
ΔΔAWV
0.137
0.691
ΔARH(+)
0.031
0.001
ΔΔAWV, 1-month lag
−0.648
0.050
ΔARH(−)
0.031
0.002
ΔΔARH
0.027
0.007
ΔAAP(+)
−0.003
0.913
ΔAAP(−)
−0.005
0.846
Short-run estimate
ΔΔAT(−)
0.089
0.029
ΔΔASH(+)
0.000
0.889
ΔΔASH(+), 1-month lag
0.002
0.168
ΔΔAWV(+)
−0.177
0.671
ΔΔAWV(−)
0.047
0.909
ΔΔAWV(−), 1-month lag
−0.809
0.045
ΔΔARH(−)
0.002
0.850
Adjustment for seasonality as a dummy variable. NARDL, nonlinear autoregressive distributed lag model; ARDL, autoregressive distributed lag model; AT, average temperature; AP, aggregate precipitation; ASH, aggregate sunshine hours; AWV, average wind velocity; ARH, average relative humidity; AAP, average air pressure; HB, human brucellosis; and VIF, variance inflation factor.
