3 ms·

Causal inference in Python

• dash2 10y ago
  I'd love for someone to explain about this "inferring causality from observational data". I've read a little before, and it sounded really exciting, but I wondered if it was just the same as instrumental variables with multiple instruments. Anyone know more?
  • axplusb 10y ago
    There are actually two sides of what is referred to as causal inference. Either (a) inferring a causal graph from the data, or (b) given a graph and data, measuring the causal effect of variables among each other. The broad idea in (a) is to start with a fully connected graph, and eliminate edges between nodes that can be tested as independent, or independent conditionally on other nodes. This gives you a non-directed graph which can be oriented by several methods (identifying V-structures, looking at residuals of regressions of X on Y vs Y on X). The theory in (b) actually generalizes instrumental variables and lays out graphical configurations where you can measure the causal effect of a variable onto another variable, and how to compute that effect. A great reference: https://www.amazon.com/Causality-Reasoning-Inference-Judea-Pearl/dp/052189560X https://www.amazon.com/Causality-Reasoning-Inference-Judea-P... A nice introduction: https://www.youtube.com/watch?v=RPgvfSeQB8A https://www.youtube.com/watch?v=RPgvfSeQB8A
  • closed 10y ago
    Statistical modeling often involves two parts, model selection and model inference. As it turns out, many models have a causal interpretation. For example, a simple linear model could be interpreted as x causes y using the equation y = b*x + error. If you get a bunch of variables and relate them through linear equations where some cause others plus some error, then different patterns of causal relations imply different covariance matrices. Classically, people have used these covariance matrices to choose between possible causal models. There are different approaches, but a common one in the behavioral sciences is to choose a few causal models to represent theories, and then perform model selection (like choosing between multiple regression models with different variables). To answer your question, though, instrumental variables are a specific causal pattern in a model, but there can be other models, such as those with latent variables.
• ertand 10y ago
  There is also Tetrad. It's written in Java and has a GUI. https://github.com/cmu-phil/tetrad https://github.com/cmu-phil/tetrad
• ebola1717 10y ago
  Haha, was wondering if this was by Adam Kelleher. I've been to a few of the meetups he hosted about causal inference. Really smart guy.
• rrggrr 10y ago
  I always wish the readme included a real-world example to help make the libraries more accessible.