SQINMA11(dgp_params, eps_params, M_rows, N_cols, dist, dist_ao, prerun)SQINMA11
Notedocblock
A struct to define a Spatial quadratic integer moving-average SQINMA(1, 1):
\(\qquad X_{t_1, t_2}=\beta_1 \circ \epsilon_{t_1-1, t_2}^a+\beta_2 \circ \epsilon_{t_1, t_2-1}^b+\beta_3 \circ \epsilon_{t_1-1, t_2-1}^c+\epsilon_{t_1, t_2}.\)
Xₜ₁,ₜ₂ = β₁ ⋅ εₜ₁₋₁,ₜ₂ᵃ + β₂ ⋅ εₜ₁,ₜ₂₋₁ᵇ + β₃ ⋅ εₜ₁₋₁,ₜ₂₋₁ᶜ + εₜ₁,ₜ₂
dgp_params::Tuple(β₁::Float64, β₂::Float64, β₃::Float64).Note that β₁, β₂, and β₃ ∈ [0, 1].eps_params::Tuple(a::Int, b::Int, c::Int).A tuple of the parameters of the DGP, indicating which error terms shall be squared. Note thateps_params∈ {1, 2}.M_rows::Int.The number of rows for the final “SOP” matrix. Note that the final spatial matrix (“picture”) equals m + 1.N_cols::Int.The number of columns for the final “SOP” matrix. Note that the final spatial matrix (“picture”) equals n + 1.dist::DistributionA distribution for . You can use any univariate distribution from theDistributions.jlpackage.dist_ao::Nothing.prerun::IntA value to initialize the DGP to achieve stationarity. These number of rows and columns will be discarded after the initialization.
sqinma11 = SQINMA11((0.5, 0.3, 0.2), (1, 1, 2), 10, 10, Normal(0, 1), nothing, 1)