Skip to main content

Targeted maximum likelihood estimation for network-dependent data

Project description


tests version docs Downloads


MossSpider provides an implementation of the targeted maximum likelihood estimator for network-dependent data (network-TMLE) in Python. Currently mossspider supports estimation of the conditional network mean for stochastic policies.

mossspider get its name from the spruce-fir moss spider, a tarantula that is both the world's smallest tarantula and native to North Carolina.



You can install via python -m pip install mossspider


The dependencies are: numpy, scipy, statsmodels, networkx, matplotlib. Notice that NetworkX must be at least 2.0.0 to work properly.

Getting started

To demonstrate mossspider, below is a simple demonstration of calculating the mean for the following data.

from mossspider import NetworkTMLE
from mossspider.dgm import uniform_network, generate_observed

First, we will use some built-in data generating functions

graph = uniform_network(n=500, degree=[1, 4])
graph_observed = generate_observed(graph)

Now, we can use NetworkTMLE to estimate the causal conditional mean under a stochastic policy. Here, the stochastic policy sets everyone's probability of action A=1 to 0.65.

ntmle = NetworkTMLE(network=graph_observed,
                    exposure='A',  # Exposure in graph
                    outcome='Y',   # Outcome in graph
                    verbose=True)  # Print model summaries
ntmle.exposure_model(model="W + W_sum")
ntmle.exposure_map_model(model='A + W + W_sum',  # Parametric model
                         measure='sum',          # Summary measure for A^s
                         distribution='poisson') # Model distribution to use
ntmle.outcome_model(model='A + A_sum + W + W_sum'), samples=500)

For full details on using mossspider, see the full documentation and worked examples available at MossSpider website.

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

mossspider-0.0.3.tar.gz (26.1 kB view hashes)

Uploaded Source

Built Distribution

mossspider-0.0.3-py3-none-any.whl (29.2 kB view hashes)

Uploaded Python 3

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