I have estimated several VAR models, which I would now like to compare looking at the residuals. I have tried several ways to access the residuals but none seem to work. Maybe it is because R does not seem to understand when I indedx $pval, as the picture here shows 
:
library(vars)
optimal_lag <- VARselect(energy.bv, lag.max = 10, type = "both")$selection["AIC(n)"]
model_season_none <- VAR(energy.bv, p = optimal_lag, type = "none", season = 12)
model_season_const <- VAR(energy.bv, p = optimal_lag, type = "const", season = 12)
model_season_trend <- VAR(energy.bv, p = optimal_lag, type = "trend", season = 12)
model_season_both <- VAR(energy.bv, p = optimal_lag, type = "both", season = 12)
summary(model_season_none)
# I checked there was no non finite values
if (any(!is.finite(model_season_none$residuals))) {
cat("Non-finite values exist in the residuals\n")
} else {
cat("No non-finite values in the residuals\n")
}
serial_test_none <- serial.test(model_season_none, lags.pt = 12,type = "PT.asymptotic")
serial_test_const <- serial.test(model_season_const, lags.pt = 12,type =
"PT.asymptotic")
serial_test_trend <- serial.test(model_season_trend, lags.pt = 12,type =
"PT.asymptotic")
serial_test_both <- serial.test(model_season_both, lags.pt = 12,type = "PT.asymptotic")
cat("P-values pour le test d'auto corrélation des résidus : \n")
cat("None Season: ", serial_test_none$pval, "\n")
cat("Constant Season: ", serial_test_const$pval, "\n")
cat("Trend Season: ", serial_test_trend$pval, "\n")
cat("Both: ", serial_test_both$pval, "\n")
Seeing as it didn't print anything (except the character strings, but not the p-values of the tests), I tried this :
print("P-values pour le test d'auto corrélation des résidus : ")
print(paste("None Season: ", serial_test_none$pval))
print(paste("Constant Season: ", serial_test_const$pval))
print(paste("Trend Season: ", serial_test_trend$pval))
print(paste("Both: ", serial_test_both$pval))
I didn't work too, so I tried this :
res_df <- data.frame(Model = c("None Season", "Constant Season", "Trend Season",
"Both"),
P_value = c(serial_test_none$pval, serial_test_const$pval,
serial_test_trend$pval, serial_test_both$pval))
# Access the p-values:
res_df$P_value[1] # None Season
res_df$P_value[2] # Constant Season
res_df$P_value[3] # Trend Season
res_df$P_value[4] # Both
Here is the message of error I got :
Error in data.frame(Model = c("None Season", "Constant Season", "Trend Season", :
les arguments impliquent des nombres de lignes différents : 4, 0
Could anyone help me to access the p-values of this test?
