Skip to main content

Information-Theoretic Causal Inference on Discrete Data

Project description


Caddie is a collection of bivariate discrete causal inference methods based on information-theoretic Additive Noise Models (ANM) and MDL-based instantiation of Algorithmic Independence of Conditionals (AIC).

Caddie Module Installation

The recommended way to install the caddie module is to simply use pip:

$ pip install caddie

Caddie officially supports Python >= 3.6.

How to use caddie?

>>> X = [1] * 1000
>>> Y = [-1] * 1000
>>> from caddie import cisc
>>> cisc.cisc(X, Y)                                                   # CISC
(0.0, 0.0)
>>> from caddie import anm, measures
>>> anm.fit_both_dir(X, Y, measures.StochasticComplexity)             # CRISP
(0.0, 0.0)
>>> anm.fit_both_dir(X, Y, measures.ChiSquaredTest)                   # DR
(1.0, 1.0)
>>> anm.fit_both_dir(X, Y, measures.ShannonEntropy)                   # ACID
(0.0, 0.0)
>>> from caddie import simulations
>>> simulations.simulate_decision_rate_against_data_type('/results/dir/') # for decision rate vs data type plots
>>> simulations.simulate_accuracy_against_sample_size('/results/dir/')    # for accuracy/decidability vs sample size plots

How to cite the paper?

Todo: Add the citation to thesis.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

caddie-0.1.6.tar.gz (7.3 kB view hashes)

Uploaded source

Built Distribution

caddie-0.1.6-py3-none-any.whl (9.1 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page