Simplify the model with t-statistics
Besides p-values, t-statistics can also help decide the necessity of model parameters. In this exercise, you will practice using t-statistics to assess the significance of model parameters.
The t-statistic is computed as the estimated parameter value subtracted by its expected mean (zero in this case), and divided by its standard error. The absolute value of the t-statistic is a distance measure, that tells you how many standard errors the estimated parameter is away from 0. As a rule of thumb, if the t-statistic is larger than 2, you can reject the null hypothesis.
You will work with the same GARCH model as the previous exercise. You can access the model fitting summary in gm_result
.
This exercise is part of the course
GARCH Models in Python
Exercise instructions
- Get the model parameters, standard errors and t-statistic, and save them in the DataFrame
para_summary
. - Compute t-statistics manually using parameter values and their standard errors, and save the calculation result in
calculated_t
. - Print and review
calculated_t
. - Print and review
para_summary
.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Get parameter stats from model summary
para_summary = pd.DataFrame({'parameter':gm_result.____,
'std-err': gm_result.____,
't-value': gm_result.____})
# Verify t-statistic by manual calculation
calculated_t = para_summary['____']/para_summary['____']
# Print calculated t-value
print(____)
# Print parameter stats
print(____)