如何从Statsmodels的VAR包中便捷提取R2值？

Hey, great question! You absolutely don’t have to compute R² values manually for your VAR model in statsmodels—there’s a straightforward way to pull them right from the results object.

Accessing R² Values Directly #

After fitting your VAR model (like you did with results = model.fit(2)), you can use two built-in attributes to get the R² metrics:

1. Unadjusted R²: Use the rsquared attribute. This returns a pandas Series where each entry corresponds to the R² of the regression equation for one of your endogenous variables.

   # Get unadjusted R² for all equations
   r_squared = results.rsquared
   print(r_squared)
   

2. Adjusted R²: For R² corrected for the number of predictors, use rsquared_adj:

   # Get adjusted R² for all equations
   adjusted_r_squared = results.rsquared_adj
   print(adjusted_r_squared)
   

Example Usage #

If your VAR model uses variables like 'GDP' and 'Inflation', the output might look like this:

GDP         0.852
Inflation   0.724
dtype: float64

You can also grab the R² for a specific variable directly by indexing the Series:

# Get R² only for the GDP equation
gdp_r2 = results.rsquared['GDP']
print(gdp_r2)

That’s all—no manual calculations required! 😊

内容的提问来源于stack exchange，提问作者GroJ