On correlation and causation

 

Guillaume Filion’s new post in “The grand locus” talks about Bayesian networks and how they can help distinguish correlation from causation, two concepts that are often mistakenly put into the same box.

You can read the whole post for some examples about how and when Bayesian networks can demonstrate causation – and when they can’t. Spoiler: Filion concludes that, as the great statistician George Box said, “all models are wrong, but some are useful”, and that, at the end of the day, experimentation is needed to prove causal relationships.

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s