def naive():
bn = hh.BayesNet('A', 'B', 'C')
bn.P['A'] = pd.Series({True: .1, False: .9})
bn.P['B'] = pd.Series({True: .3, False: .7})
bn.P['C'] = pd.Series({True: .5, False: .5})
bn.prepare()
return bn
def load_sprinkler() -> hh.BayesNet:
"""Load the water sprinkler network.
This example is taken from figure 14.12(a) of Artificial Intelligence: A Modern Approach.
Example:
>>> import hedgehog as hh
>>> bn = hh.load_sprinkler()
>>> bn.query('Rain', event={'Sprinkler': True})
Rain
False 0.7
True 0.3
Name: P(Rain), dtype: float64
"""
bn = hh.BayesNet(('Cloudy', 'Sprinkler'), ('Cloudy', 'Rain'),
('Sprinkler', 'Wet grass'), ('Rain', 'Wet grass'))
# P(Cloudy)
bn.P['Cloudy'] = pd.Series({False: .5, True: .5})
# P(Sprinkler | Cloudy)
bn.P['Sprinkler'] = pd.Series({
(True, True): .1,
(True, False): .9,
(False, True): .5,
(False, False): .5
})
# P(Rain | Cloudy)
bn.P['Rain'] = pd.Series({
(True, True): .8,
(True, False): .2,
(False, True): .2,
(False, False): .8
})
# P(Wet grass | Sprinkler, Rain)
bn.P['Wet grass'] = pd.Series({
(True, True, True): .99,
(True, True, False): .01,
(True, False, True): .9,
(True, False, False): .1,
(False, True, True): .9,
(False, True, False): .1,
(False, False, True): 0,
(False, False, False): 1
})
bn.prepare()
return bn
def test_cpt_with_index_names():
"""
From https://github.com/MaxHalford/hedgehog/issues/19
"""
edges = pd.DataFrame({"parent": ["A", "B"], "child": "C"})
bn = hh.BayesNet(*edges.itertuples(index=False, name=None))
bn.P['A'] = pd.Series({True: 0.7, False: 0.3})
bn.P['B'] = pd.Series({True: 0.4, False: 0.6})
PC = pd.DataFrame({
"B": [True, True, True, True, False, False, False, False],
"A": [True, True, False, False, True, True, False, False],
"C": [
True,
False,
True,
False,
True,
False,
True,
False,
],
"p": [1, 0, 0, 1, 0.5, 0.5, 0.001, 0.999],
})
bn.P["C"] = PC.set_index(["B", "A", "C"])["p"]
bn.prepare()
pd.testing.assert_series_equal(
bn.query("C", event={
"B": False,
"A": True
}),
pd.Series([0.5, 0.5],
name="P(C)",
index=pd.Index([False, True], name="C")))
def __init__(self):
bn = hh.BayesNet(
(["player_emotion_face", "player_emotion_voice"
], "player_emotion"),
(["player_offer", "player_emotion"], "robot_decision"),
(["player_offer", "player_emotion"], "robot_offer"))
bn.P["player_emotion_face"] = pd.Series({
"positive": 0.3,
"negative": 0.3,
"neutral": 0.4
})
bn.P["player_emotion_voice"] = bn.P["player_emotion_face"]
bn.P["player_emotion"] = pd.Series({
("positive", "positive", "positive"):
1,
("positive", "positive", "negative"):
0,
("positive", "positive", "neutral"):
0,
("positive", "negative", "positive"):
0.5,
("positive", "negative", "negative"):
0.5,
("positive", "negative", "neutral"):
0,
("positive", "neutral", "positive"):
0.5,
("positive", "neutral", "negative"):
0.0,
("positive", "neutral", "neutral"):
0.5,
# -----------------------------------------
("negative", "positive", "positive"):
0.5,
("negative", "positive", "negative"):
0.5,
("negative", "positive", "neutral"):
0,
("negative", "negative", "positive"):
0,
("negative", "negative", "negative"):
1,
("negative", "negative", "neutral"):
0,
("negative", "neutral", "positive"):
0,
("negative", "neutral", "negative"):
0.5,
("negative", "neutral", "neutral"):
0.5,
# -----------------------------------------
("neutral", "positive", "positive"):
0.5,
("neutral", "positive", "negative"):
0,
("neutral", "positive", "neutral"):
0.5,
("neutral", "negative", "positive"):
0,
("neutral", "negative", "negative"):
0.5,
("neutral", "negative", "neutral"):
0.5,
("neutral", "neutral", "positive"):
0,
("neutral", "neutral", "negative"):
0,
("neutral", "neutral", "neutral"):
1,
})
bn.P["player_offer"] = pd.Series({
"1": 0.03,
"2": 0.07,
"3": 0.1,
"4": 0.2,
"5": 0.2,
"6": 0.2,
"7": 0.1,
"8": 0.07,
"9": 0.03,
})
bn.P["robot_decision"] = pd.Series({
("positive", "1", "yes"):
0,
("positive", "1", "no"):
1,
("positive", "2", "yes"):
0,
("positive", "2", "no"):
1,
("positive", "3", "yes"):
0,
("positive", "3", "no"):
1,
("positive", "4", "yes"):
0.2,
("positive", "4", "no"):
0.8,
("positive", "5", "yes"):
0.4,
("positive", "5", "no"):
0.6,
("positive", "6", "yes"):
1,
("positive", "6", "no"):
0,
("positive", "7", "yes"):
1,
("positive", "7", "no"):
0,
("positive", "8", "yes"):
1,
("positive", "8", "no"):
0,
("positive", "9", "yes"):
1,
("positive", "9", "no"):
0,
# -----------------------
("negative", "1", "yes"):
0.4,
("negative", "1", "no"):
0.6,
("negative", "2", "yes"):
0.5,
("negative", "2", "no"):
0.5,
("negative", "3", "yes"):
0.6,
("negative", "3", "no"):
0.4,
("negative", "4", "yes"):
0.7,
("negative", "4", "no"):
0.3,
("negative", "5", "yes"):
0.8,
("negative", "5", "no"):
0.2,
("negative", "6", "yes"):
0.9,
("negative", "6", "no"):
0.1,
("negative", "7", "yes"):
0.9,
("negative", "7", "no"):
0.1,
("negative", "8", "yes"):
1,
("negative", "8", "no"):
0,
("negative", "9", "yes"):
1,
("negative", "9", "no"):
0,
# -----------------------
("neutral", "1", "yes"):
0,
("neutral", "1", "no"):
1,
("neutral", "2", "yes"):
0,
("neutral", "2", "no"):
1,
("neutral", "3", "yes"):
0.1,
("neutral", "3", "no"):
0.9,
("neutral", "4", "yes"):
0.3,
("neutral", "4", "no"):
0.7,
("neutral", "5", "yes"):
0.5,
("neutral", "5", "no"):
0.5,
("neutral", "6", "yes"):
0.7,
("neutral", "6", "no"):
0.3,
("neutral", "7", "yes"):
0.9,
("neutral", "7", "no"):
0.1,
("neutral", "8", "yes"):
1,
("neutral", "8", "no"):
0,
("neutral", "9", "yes"):
1,
("neutral", "9", "no"):
0,
})
bn.P["robot_offer"] = pd.Series({
("positive", "1", "1"):
0.95,
("positive", "1", "2"):
0.05,
("positive", "1", "3"):
0,
("positive", "1", "4"):
0,
("positive", "1", "5"):
0,
("positive", "1", "6"):
0,
("positive", "1", "7"):
0,
("positive", "1", "8"):
0,
("positive", "1", "9"):
0,
("positive", "2", "1"):
0.9,
("positive", "2", "2"):
0.1,
("positive", "2", "3"):
0,
("positive", "2", "4"):
0,
("positive", "2", "5"):
0,
("positive", "2", "6"):
0,
("positive", "2", "7"):
0,
("positive", "2", "8"):
0,
("positive", "2", "9"):
0,
("positive", "3", "1"):
0.8,
("positive", "3", "2"):
0.15,
("positive", "3", "3"):
0.05,
("positive", "3", "4"):
0,
("positive", "3", "5"):
0,
("positive", "3", "6"):
0,
("positive", "3", "7"):
0,
("positive", "3", "8"):
0,
("positive", "3", "9"):
0,
("positive", "4", "1"):
0.6,
("positive", "4", "2"):
0.2,
("positive", "4", "3"):
0.1,
("positive", "4", "4"):
0.1,
("positive", "4", "5"):
0,
("positive", "4", "6"):
0,
("positive", "4", "7"):
0,
("positive", "4", "8"):
0,
("positive", "4", "9"):
0,
("positive", "5", "1"):
0.3,
("positive", "5", "2"):
0.3,
("positive", "5", "3"):
0.2,
("positive", "5", "4"):
0.1,
("positive", "5", "5"):
0.1,
("positive", "5", "6"):
0,
("positive", "5", "7"):
0,
("positive", "5", "8"):
0,
("positive", "5", "9"):
0,
("positive", "6", "1"):
0.1,
("positive", "6", "2"):
0.2,
("positive", "6", "3"):
0.2,
("positive", "6", "4"):
0.2,
("positive", "6", "5"):
0.2,
("positive", "6", "6"):
0.1,
("positive", "6", "7"):
0,
("positive", "6", "8"):
0,
("positive", "6", "9"):
0,
("positive", "7", "1"):
0.1,
("positive", "7", "2"):
0.1,
("positive", "7", "3"):
0.2,
("positive", "7", "4"):
0.2,
("positive", "7", "5"):
0.2,
("positive", "7", "6"):
0.1,
("positive", "7", "7"):
0.1,
("positive", "7", "8"):
0,
("positive", "7", "9"):
0,
("positive", "8", "1"):
0.1,
("positive", "8", "2"):
0.1,
("positive", "8", "3"):
0.1,
("positive", "8", "4"):
0.1,
("positive", "8", "5"):
0.1,
("positive", "8", "6"):
0.2,
("positive", "8", "7"):
0.2,
("positive", "8", "8"):
0.1,
("positive", "8", "9"):
0,
("positive", "9", "1"):
0.1,
("positive", "9", "2"):
0.1,
("positive", "9", "3"):
0.1,
("positive", "9", "4"):
0.1,
("positive", "9", "5"):
0.1,
("positive", "9", "6"):
0.1,
("positive", "9", "7"):
0.1,
("positive", "9", "8"):
0.2,
("positive", "9", "9"):
0.1,
# -----------------------
("negative", "1", "1"):
0,
("negative", "1", "2"):
0,
("negative", "1", "3"):
0,
("negative", "1", "4"):
0,
("negative", "1", "5"):
0.05,
("negative", "1", "6"):
0.15,
("negative", "1", "7"):
0.2,
("negative", "1", "8"):
0.3,
("negative", "1", "9"):
0.3,
("negative", "2", "1"):
0,
("negative", "2", "2"):
0,
("negative", "2", "3"):
0,
("negative", "2", "4"):
0.05,
("negative", "2", "5"):
0.15,
("negative", "2", "6"):
0.2,
("negative", "2", "7"):
0.2,
("negative", "2", "8"):
0.2,
("negative", "2", "9"):
0.2,
("negative", "3", "1"):
0,
("negative", "3", "2"):
0,
("negative", "3", "3"):
0.05,
("negative", "3", "4"):
0.15,
("negative", "3", "5"):
0.1,
("negative", "3", "6"):
0.2,
("negative", "3", "7"):
0.2,
("negative", "3", "8"):
0.2,
("negative", "3", "9"):
0.2,
("negative", "4", "1"):
0,
("negative", "4", "2"):
0,
("negative", "4", "3"):
0,
("negative", "4", "4"):
0.05,
("negative", "4", "5"):
0.15,
("negative", "4", "6"):
0.2,
("negative", "4", "7"):
0.2,
("negative", "4", "8"):
0.2,
("negative", "4", "9"):
0.2,
("negative", "5", "1"):
0,
("negative", "5", "2"):
0,
("negative", "5", "3"):
0,
("negative", "5", "4"):
0,
("negative", "5", "5"):
0.05,
("negative", "5", "6"):
0.15,
("negative", "5", "7"):
0.2,
("negative", "5", "8"):
0.3,
("negative", "5", "9"):
0.3,
("negative", "6", "1"):
0,
("negative", "6", "2"):
0,
("negative", "6", "3"):
0,
("negative", "6", "4"):
0,
("negative", "6", "5"):
0,
("negative", "6", "6"):
0.2,
("negative", "6", "7"):
0.2,
("negative", "6", "8"):
0.3,
("negative", "6", "9"):
0.3,
("negative", "7", "1"):
0,
("negative", "7", "2"):
0,
("negative", "7", "3"):
0,
("negative", "7", "4"):
0,
("negative", "7", "5"):
0,
("negative", "7", "6"):
0,
("negative", "7", "7"):
0.3,
("negative", "7", "8"):
0.4,
("negative", "7", "9"):
0.4,
("negative", "8", "1"):
0,
("negative", "8", "2"):
0,
("negative", "8", "3"):
0,
("negative", "8", "4"):
0,
("negative", "8", "5"):
0,
("negative", "8", "6"):
0,
("negative", "8", "7"):
0.1,
("negative", "8", "8"):
0.4,
("negative", "8", "9"):
0.5,
("negative", "9", "1"):
0,
("negative", "9", "2"):
0,
("negative", "9", "3"):
0,
("negative", "9", "4"):
0,
("negative", "9", "5"):
0,
("negative", "9", "6"):
0,
("negative", "9", "7"):
0,
("negative", "9", "8"):
0.5,
("negative", "9", "9"):
0.5,
# -----------------------
("neutral", "1", "1"):
0.03,
("neutral", "1", "2"):
0.07,
("neutral", "1", "3"):
0.1,
("neutral", "1", "4"):
0.2,
("neutral", "1", "5"):
0.2,
("neutral", "1", "6"):
0.2,
("neutral", "1", "7"):
0.1,
("neutral", "1", "8"):
0.07,
("neutral", "1", "9"):
0.03,
("neutral", "2", "1"):
0.03,
("neutral", "2", "2"):
0.07,
("neutral", "2", "3"):
0.1,
("neutral", "2", "4"):
0.2,
("neutral", "2", "5"):
0.2,
("neutral", "2", "6"):
0.2,
("neutral", "2", "7"):
0.1,
("neutral", "2", "8"):
0.07,
("neutral", "2", "9"):
0.03,
("neutral", "3", "1"):
0.03,
("neutral", "3", "2"):
0.07,
("neutral", "3", "3"):
0.1,
("neutral", "3", "4"):
0.2,
("neutral", "3", "5"):
0.2,
("neutral", "3", "6"):
0.2,
("neutral", "3", "7"):
0.1,
("neutral", "3", "8"):
0.07,
("neutral", "3", "9"):
0.03,
("neutral", "4", "1"):
0.03,
("neutral", "4", "2"):
0.07,
("neutral", "4", "3"):
0.1,
("neutral", "4", "4"):
0.2,
("neutral", "4", "5"):
0.2,
("neutral", "4", "6"):
0.2,
("neutral", "4", "7"):
0.1,
("neutral", "4", "8"):
0.07,
("neutral", "4", "9"):
0.03,
("neutral", "5", "1"):
0.03,
("neutral", "5", "2"):
0.07,
("neutral", "5", "3"):
0.1,
("neutral", "5", "4"):
0.2,
("neutral", "5", "5"):
0.2,
("neutral", "5", "6"):
0.2,
("neutral", "5", "7"):
0.1,
("neutral", "5", "8"):
0.07,
("neutral", "5", "9"):
0.03,
("neutral", "6", "1"):
0.03,
("neutral", "6", "2"):
0.07,
("neutral", "6", "3"):
0.1,
("neutral", "6", "4"):
0.2,
("neutral", "6", "5"):
0.2,
("neutral", "6", "6"):
0.2,
("neutral", "6", "7"):
0.1,
("neutral", "6", "8"):
0.07,
("neutral", "6", "9"):
0.03,
("neutral", "7", "1"):
0.03,
("neutral", "7", "2"):
0.07,
("neutral", "7", "3"):
0.1,
("neutral", "7", "4"):
0.2,
("neutral", "7", "5"):
0.2,
("neutral", "7", "6"):
0.2,
("neutral", "7", "7"):
0.1,
("neutral", "7", "8"):
0.07,
("neutral", "7", "9"):
0.03,
("neutral", "8", "1"):
0.03,
("neutral", "8", "2"):
0.07,
("neutral", "8", "3"):
0.1,
("neutral", "8", "4"):
0.2,
("neutral", "8", "5"):
0.2,
("neutral", "8", "6"):
0.2,
("neutral", "8", "7"):
0.1,
("neutral", "8", "8"):
0.07,
("neutral", "8", "9"):
0.03,
("neutral", "9", "1"):
0.03,
("neutral", "9", "2"):
0.07,
("neutral", "9", "3"):
0.1,
("neutral", "9", "4"):
0.2,
("neutral", "9", "5"):
0.2,
("neutral", "9", "6"):
0.2,
("neutral", "9", "7"):
0.1,
("neutral", "9", "8"):
0.07,
("neutral", "9", "9"):
0.3,
})
bn.prepare()
self.bn = bn
self.emotionSimplifier = {
'happy': 'positive',
'surprised': 'neutral',
'calm': 'neutral',
'disgusted': 'negative',
'sad': 'negative',
'fearful': 'negative',
'angry': 'negative'
}
def load_asia() -> hh.BayesNet:
"""Load the Asia network.
Example:
>>> import hedgehog as hh
>>> bn = hh.load_asia()
>>> bn.query('Lung cancer', event={'Visit to Asia': True, 'Smoker': False})
Lung cancer
False 0.99
True 0.01
Name: P(Lung cancer), dtype: float64
"""
bn = hh.BayesNet(('Visit to Asia', 'Tuberculosis'),
('Smoker', ('Lung cancer', 'Bronchitis')),
(('Tuberculosis', 'Lung cancer'), 'TB or cancer'),
('TB or cancer', ('Positive X-ray', 'Dispnea')),
('Bronchitis', 'Dispnea'))
# P(Visit to Asia)
bn.P['Visit to Asia'] = pd.Series({True: .01, False: .99})
# P(Tuberculosis | Visit to Asia)
bn.P['Tuberculosis'] = pd.Series({
(True, True): .05,
(True, False): .95,
(False, True): .01,
(False, False): .99
})
# P(Smoker)
bn.P['Smoker'] = pd.Series({True: .5, False: .5})
# P(Lung cancer | Smoker)
bn.P['Lung cancer'] = pd.Series({
(True, True): .1,
(True, False): .9,
(False, True): .01,
(False, False): .99
})
# P(Bronchitis | Smoker)
bn.P['Bronchitis'] = pd.Series({
(True, True): .6,
(True, False): .4,
(False, True): .3,
(False, False): .7
})
# P(TB or cancer | Tuberculosis, Lung cancer)
bn.P['TB or cancer'] = pd.Series({
(True, True, True): 1,
(True, True, False): 0,
(True, False, True): 1,
(True, False, False): 0,
(False, True, True): 1,
(False, True, False): 0,
(False, False, True): 0,
(False, False, False): 1
})
# P(Positive X-ray | TB or cancer)
bn.P['Positive X-ray'] = pd.Series({
(True, True): .98,
(True, False): .02,
(False, True): .05,
(False, False): .95
})
# P(Dispnea | TB or cancer, Bronchitis)
bn.P['Dispnea'] = pd.Series({
(True, True, True): .9,
(True, True, False): .1,
(True, False, True): .7,
(True, False, False): .3,
(False, True, True): .8,
(False, True, False): .2,
(False, False, True): .1,
(False, False, False): .9
})
bn.prepare()
return bn
def load_alarm() -> hh.BayesNet:
"""Load Judea Pearl's famous example.
At the time of writing his seminal paper on Bayesian networks, Judea Pearl lived in California,
where earthquakes are quite common.
Example:
>>> import hedgehog as hh
>>> bn = hh.load_alarm()
>>> bn.query('John calls', 'Mary calls', event={'Burglary': True, 'Earthquake': False})
John calls Mary calls
False False 0.08463
True 0.06637
True False 0.25677
True 0.59223
Name: P(John calls, Mary calls), dtype: float64
"""
bn = hh.BayesNet(('Burglary', 'Alarm'), ('Earthquake', 'Alarm'),
('Alarm', 'John calls'), ('Alarm', 'Mary calls'))
# P(Burglary)
bn.P['Burglary'] = pd.Series({False: .999, True: .001})
# P(Earthquake)
bn.P['Earthquake'] = pd.Series({False: .998, True: .002})
# P(Alarm | Burglary, Earthquake)
bn.P['Alarm'] = pd.Series({
(True, True, True): .95,
(True, True, False): .05,
(True, False, True): .94,
(True, False, False): .06,
(False, True, True): .29,
(False, True, False): .71,
(False, False, True): .001,
(False, False, False): .999
})
# P(John calls | Alarm)
bn.P['John calls'] = pd.Series({
(True, True): .9,
(True, False): .1,
(False, True): .05,
(False, False): .95
})
# P(Mary calls | Alarm)
bn.P['Mary calls'] = pd.Series({
(True, True): .7,
(True, False): .3,
(False, True): .01,
(False, False): .99
})
bn.prepare()
return bn
def load_grades():
"""Load the student grades network.
Example:
>>> import hedgehog as hh
>>> bn = hh.load_grades()
>>> bn.nodes
['Difficulty', 'Grade', 'Intelligence', 'Letter', 'SAT']
>>> bn.children
{'Difficulty': ['Grade'], 'Intelligence': ['Grade', 'SAT'], 'Grade': ['Letter']}
>>> bn.parents
{'Grade': ['Difficulty', 'Intelligence'], 'SAT': ['Intelligence'], 'Letter': ['Grade']}
>>> bn.query('Letter', 'SAT', event={'Intelligence': 'Smart'})
Letter SAT
Strong Failure 0.153544
Success 0.614176
Weak Failure 0.046456
Success 0.185824
Name: P(Letter, SAT), dtype: float64
"""
bn = hh.BayesNet(('Difficulty', 'Grade'), ('Intelligence', 'Grade'),
('Intelligence', 'SAT'), ('Grade', 'Letter'))
# P(Difficulty)
bn.P['Difficulty'] = pd.Series({'Easy': .6, 'Hard': .4})
# P(Intelligence)
bn.P['Intelligence'] = pd.Series({'Average': .7, 'Smart': .3})
# P(Grade | Difficult, Intelligence)
bn.P['Grade'] = pd.Series({
('Easy', 'Average', 'A'): .3,
('Easy', 'Average', 'B'): .4,
('Easy', 'Average', 'C'): .3,
('Easy', 'Smart', 'A'): .9,
('Easy', 'Smart', 'B'): .08,
('Easy', 'Smart', 'C'): .02,
('Hard', 'Average', 'A'): .05,
('Hard', 'Average', 'B'): .25,
('Hard', 'Average', 'C'): .7,
('Hard', 'Smart', 'A'): .5,
('Hard', 'Smart', 'B'): .3,
('Hard', 'Smart', 'C'): .2
})
# P(SAT | Intelligence)
bn.P['SAT'] = pd.Series({
('Average', 'Failure'): .95,
('Average', 'Success'): .05,
('Smart', 'Failure'): .2,
('Smart', 'Success'): .8
})
# P(Letter | Grade)
bn.P['Letter'] = pd.Series({
('A', 'Weak'): .1,
('A', 'Strong'): .9,
('B', 'Weak'): .4,
('B', 'Strong'): .6,
('C', 'Weak'): .99,
('C', 'Strong'): .01
})
bn.prepare()
return bn