Merge pull request #79 from LAMPSPUC/fix_simulation_no_slope

fix simulation for model with level and without slope

Author: andreramosfdc

Project.toml

@@ -1,7 +1,7 @@
 name = "StateSpaceLearning"
 uuid = "971c4b7c-2c4e-4bac-8525-e842df3cde7b"
 authors = ["andreramosfc <[email protected]>"]
-version = "2.0.9"
+version = "2.0.10"
 
 [deps]
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"

src/models/structural_model.jl

@@ -95,10 +95,10 @@ mutable struct StructuralModel <: StateSpaceLearningModel
 
         @assert !has_intercept(exog) "Exogenous matrix must not have an intercept column"
         @assert 1 <= stochastic_start < length(y) "stochastic_start must be greater than or equal to 1 and smaller than the length of the time series"
-        @assert level in ["deterministic", "stochastic", "none"] "level must be either deterministic, stochastic or no"
-        @assert slope in ["deterministic", "stochastic", "none"] "slope must be either deterministic, stochastic or no"
-        @assert seasonal in ["deterministic", "stochastic", "none"] "seasonal must be either deterministic, stochastic or no"
-        @assert cycle in ["deterministic", "stochastic", "none"] "cycle must be either deterministic, stochastic or no"
+        @assert level in ["deterministic", "stochastic", "none"] "level must be either deterministic, stochastic or none"
+        @assert slope in ["deterministic", "stochastic", "none"] "slope must be either deterministic, stochastic or none"
+        @assert seasonal in ["deterministic", "stochastic", "none"] "seasonal must be either deterministic, stochastic or none"
+        @assert cycle in ["deterministic", "stochastic", "none"] "cycle must be either deterministic, stochastic or none"
         @assert seasonal != "none" ? length(y) > minimum(freq_seasonal) : true "Time series must be longer than the seasonal period if seasonal is added"
 
         if typeof(freq_seasonal) <: Vector
@@ -1591,7 +1591,7 @@ function simulate_states(
         )
     end
 
-    model.slope ? slope = ones(T, N_scenarios) .* slope : nothing
+    slope = ones(T, N_scenarios) .* slope
     model.level ? trend = ones(T, N_scenarios) .* trend : nothing
     model.seasonal ? seasonals = [ones(T, N_scenarios) .* s for s in seasonals] : nothing
     for t in (T + 1):(T + steps_ahead)