将使用 PyMC3 创建的模型绘制为图形

Question

我正在使用以下代码通过 PyMC3 创建一个简单的模型：

import pymc3 as pm
import theano.tensor as tt

with pm.Model() as model:
    p = pm.Uniform("freq_cheating", 0, 1)
    p_skewed = pm.Deterministic("p_skewed", 0.5*p + 0.25)

    yes_responses = pm.Binomial("number_cheaters", 100, p_skewed, observed= 50) 

    step = pm.Metropolis()
    trace = pm.sample(25000, step=step)
    burned_trace50 = trace[2500:]

是否可以将此模型绘制为 DAG？

Answer 1

Added in version 3.5 是 model_to_graphviz 方法，它正是这样做的。对于上面的示例，您将使用

pm.model_to_graphviz(model)

将使用 PyMC3 创建的模型绘制为图形

Plotting a Model created with PyMC3 as a graph

python

directed-acyclic-graphs

pymc3