3. BBN Generation

BBN generation is available. We support singly- and mulit-connected BBNs.

3.1. Singly-connected BBN

Use generate_singly_bbn() to generate a singly-connected BBN and then use create_reasoning_model() to create a reasoning model.

from pybbn.factory import create_reasoning_model
from pybbn.generator import generate_singly_bbn
import numpy as np

g, p = generate_singly_bbn(
    n=10, max_iter=20, max_values=2, max_alpha=10, rng=np.random.default_rng(seed=37)
)

model = create_reasoning_model(g, p)

The singly-connected BBN graph looks like the following.

import networkx as nx
import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(5, 5))

pos = nx.nx_agraph.graphviz_layout(g, prog="dot")
nx.draw(g, pos=pos, with_labels=True, node_color="#e0e0e0")

fig.tight_layout()

Querying of the posteriors proceeds as usual.

import pandas as pd

q1 = model.pquery()

e = {"4": model.create_observation_evidences("4", "s0")}
q2 = model.pquery(evidences=e)

e = {"4": model.create_observation_evidences("4", "s1")}
q3 = model.pquery(evidences=e)

p1 = pd.Series({n: q1[n].iloc[0]["__p__"] for n in model.d.nodes()})
p2 = pd.Series({n: q2[n].iloc[0]["__p__"] for n in model.d.nodes()})
p3 = pd.Series({n: q3[n].iloc[0]["__p__"] for n in model.d.nodes()})

pd.DataFrame([p1, p2, p3]).T.rename(columns={0: "q1", 1: "q2", 2: "q3"})

	q1	q2	q3
0	0.500393	0.500424	0.500380
1	0.370587	0.370672	0.370552
2	0.500547	0.500535	0.500551
3	0.221217	0.222402	0.220730
4	0.291306	1.000000	0.000000
5	0.672492	0.630880	0.689596
6	0.246865	0.246865	0.246865
7	0.603267	0.597744	0.605537
8	0.304271	0.304277	0.304269
9	0.887248	0.961834	0.856590

from pybbn.serde import model_to_dict
import json

with open("../../_support/bbn/singly-connected.json", "w") as fp:
    json.dump(model_to_dict(model), fp, indent=1)

3.2. Multi-connected BBN

Use generate_multi_bbn() to generate a singly-connected BBN and then use create_reasoning_model() to create a reasoning model.

from pybbn.generator import generate_multi_bbn

g, p = generate_multi_bbn(
    n=10, max_iter=5, max_values=2, max_alpha=10, rng=np.random.default_rng(seed=37)
)

model = create_reasoning_model(g, p)

fig, ax = plt.subplots(figsize=(5, 5))

pos = nx.nx_agraph.graphviz_layout(g, prog="dot")
nx.draw(g, pos=pos, with_labels=True, node_color="#e0e0e0")

fig.tight_layout()

import pandas as pd

q1 = model.pquery()

e = {"0": model.create_observation_evidences("0", "s0")}
q2 = model.pquery(evidences=e)

e = {"0": model.create_observation_evidences("0", "s1")}
q3 = model.pquery(evidences=e)

p1 = pd.Series({n: q1[n].iloc[0]["__p__"] for n in model.d.nodes()})
p2 = pd.Series({n: q2[n].iloc[0]["__p__"] for n in model.d.nodes()})
p3 = pd.Series({n: q3[n].iloc[0]["__p__"] for n in model.d.nodes()})

pd.DataFrame([p1, p2, p3]).T.rename(columns={0: "q1", 1: "q2", 2: "q3"})

	q1	q2	q3
0	0.026385	1.000000	0.000000
1	0.564311	0.210227	0.573907
2	0.486509	0.402171	0.488795
3	0.352396	0.357359	0.352262
4	0.492260	0.477167	0.492669
5	0.446751	0.412918	0.447668
6	0.458902	0.490701	0.458041
7	0.584410	0.767471	0.579449
8	0.634409	0.590927	0.635587
9	0.491198	0.502911	0.490881

with open("../../_support/bbn/multi-connected.json", "w") as fp:
    json.dump(model_to_dict(model), fp, indent=1)