When using anova_lm from statsmodels with a single multiple linear regression model, how are statistics for each predictor calculated ?
E.g.
model = sm.formula.ols(mpg ~ cylinders + displacement + horsepower + weight + acceleration + year + origin, data = auto)
anova_lm(model)
How are sum_sq, mean_sq and F values calculated for each independent predictor (e.g.horsepower) ?
