mossspider.dgm.generate_truth¶
- generate_truth(graph, p)¶
Simulates the true conditional mean outcome for a given network, distribution of W, and policy.
The true mean under the policy is simulated as
\[A = Bernoulli(p) \ Y = Bernoulli(expit(-2.5 + 1.5*W + 0.5*A + 1.5*map(A) + 1.5*map(W)))\]- Returns
- Return type
float
Examples
Loading the necessary functions
>>> from mossspider.dgm import uniform_network, generate_truth
Generating the uniform network
>>> G = uniform_network(n=500, degree=[0, 2])
Calculating truth for a policy via a large number of replicates
>>> true_p = [] >>> for i in range(1000): >>> y_mean = generate_truth(graph=G, p=0.5) >>> true_p.append(y_mean) >>> np.mean(true_p) # 'true' value for the stochastic policy
To reduce random error, a large number of replicates should be used