What is the purpose of providing observation data in gcm.conventional_samples()?
CrisprM opened this issue · comments
CrisprM commented
Ask your question
Hi
About simulating the Impact of Interventions, I would like to understand the difference between using ""observed_data"" and ""num_samples_to_draw"" in gcm.conventional_samples()?
I think both generate data using fitted structural equations, what is the effect of providing observed_data? Can you give me an example that shows the difference?
Thank you very much.
samples_1 = gcm.interventional_samples(causal_model,
{'X': lambda x: 1},
observed_data=data)
samples_2 = gcm.interventional_samples(causal_model,
{'X': lambda x: 1},
num_samples_to_draw=1000)