def __init__(self, file):
self.rooms = {}
self.sensors = {}
self.P = 0
self.measure = []
# Fill problem
self.read_file(file)
bayes_template = []
for room in self.rooms.values():
bayes_template.append((room.name + "_p", '', 0.5))
for room in self.rooms.values():
aux_str = room.name + "_p"
for parent in room.adj_rooms:
aux_str += " " + parent + "_p"
truth_table = self.create_truth_table(len(room.adj_rooms) +
1) if room.adj_rooms else {
True: 1,
False: 0
}
bayes_template.append((room.name, aux_str, truth_table))
for sensor in self.sensors.values():
bayes_template.append((sensor.id, sensor.room, {
True: sensor.TPR,
False: sensor.FPR
}))
bayes_template.append((sensor.id + "_p", sensor.room + "_p", {
True: sensor.TPR,
False: sensor.FPR
}))
for node in bayes_template:
print(node)
self.network_template = BayesNet(bayes_template)
def generate_bayesnet():
"""
Generates a BayesNet object representing the Bayesian network in Part 2
returns the BayesNet object
"""
bayes_net = BayesNet()
# load the dataset, a list of DataPoint objects
data = pickle.load(open("data/bn_data.p", "rb"))
# BEGIN_YOUR_CODE ######################################################
raise NotImplementedError
# END_YOUR_CODE ########################################################
return bayes_net
def create(self):
baye = []
for step in range(self.time_step + 1):
for room in self.room_list:
# Create parents at t=0
if step == 0:
baye.append((room.name + '_' + str(step), '', 0.5))
if room.sensor:
s_name = room.sensor.name + '_' + str(step)
r_name = room.name + '_' + str(step)
baye.append((s_name, r_name,
self.get_prob(sensor=room.sensor)))
elif step != self.time_step and step != 0:
# Start building the child nodes
parent = None
parent = room.name + '_' + str(step - 1)
# Add parents of current node
for neighboor in room.neighbours:
parent = parent + ' ' + neighboor + '_' + str(step - 1)
# Append to list, a tuple witt all the info of the current node
r_name = room.name + '_' + str(step)
baye.append((r_name, parent, self.get_prob(room=room)))
# Do sensors now
if room.sensor: #and any(item == step for item in room.sensor.time):
s_name = room.sensor.name + '_' + str(step)
r_name = room.name + '_' + str(step)
baye.append((s_name, r_name,
self.get_prob(sensor=room.sensor)))
# if step == self.time_step:
# parent = None
# parent = room.name + '_' + str(step)
# for neighboor in room.neighbours:
# parent = parent + ' ' + neighboor + '_' + str(step)
# r_name = room.name + '_' + str(step+1)
# baye.append((r_name, parent, self.get_prob(room = room)))
else:
parent = None
parent = room.name + '_' + str(step - 1)
for neighboor in room.neighbours:
parent = parent + ' ' + neighboor + '_' + str(step - 1)
# Append to list, a tuple witt all the info of the current node
r_name = room.name + '_' + str(step)
baye.append((r_name, parent, self.get_prob(room=room)))
print(baye)
return BayesNet(baye)
def create(self):
baye = []
# Step in this case goes only until T-1
for step in range(self.time_step):
for room in self.room_list:
# Create parents at t=0
if step == 0:
# Add initial probability
baye.append((room.name + '_' + str(step), '', 0.5))
# Add respective sensors if any
if room.sensor:
for s in room.sensor:
s_name = s.name + '_' + str(step)
r_name = room.name + '_' + str(step)
baye.append(
(s_name, r_name, self.get_prob(sensor=s)))
else:
# Start building the child nodes
parent = None
ax = 1
# Add parents of current node
parent = room.name + '_' + str(step - 1)
for neighboor in room.neighbours:
parent = parent + ' ' + neighboor + '_' + str(step - 1)
ax += 1
# Append to list, a tuple witt all the info of the current node
r_name = room.name + '_' + str(step)
baye.append((r_name, parent, self.get_prob(room=room)))
# Add respective sensors of the room if any
if room.sensor:
for s in room.sensor:
s_name = s.name + '_' + str(step)
r_name = room.name + '_' + str(step)
baye.append(
(s_name, r_name, self.get_prob(sensor=s)))
# Uncomment to see bayes net
#print(baye)
return BayesNet(baye)
print(spam_filter.filter(["i", "am", "spam", "spam", "i", "am"])) # spam list
print(spam_filter.filter(["do", "i", "like", "green", "eggs", "and", "ham"])) # ham list
print(spam_filter.filter(["i", "do", "not", "like", "that", "spamiam"]))
# Problem 2a
# TODO: copy into Jupyter notebook
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask
# Utility variables
T, F = True, False
grass = BayesNet([
('Cloudy', '', 0.5),
('Sprinkler', 'Cloudy', {T:0.1, F:0.5}),
('Rain', 'Cloudy', {T:0.8, F:0.2}),
('WetGrass', 'Sprinkler Rain', {(T, T):0.99, (T,F):0.9, (F,T):0.9, (F,F):0.0}),
])
print("\n\nProblem 2")
print("i. " + enumeration_ask('Cloudy', dict(), grass).show_approx())
print("ii. " + enumeration_ask('Sprinkler', dict(Cloudy=T), grass).show_approx())
print("iii. " + enumeration_ask('Cloudy', dict(Sprinkler=T, Rain=F), grass).show_approx())
print("iv. " + enumeration_ask('WetGrass', dict(Cloudy=T, Sprinkler=T, Rain=T), grass).show_approx())
print("v. " + enumeration_ask('Cloudy', dict(WetGrass=F), grass).show_approx())
# Burglary example [Figure 14.2]
from probability import BayesNet
T, F = True, False
death = BayesNet([('Overweight', '', 0.339), ('HighStress', '', 0.72),
('HeartAttack', 'Overweight HighStress', {
(T, T): 0.72,
(T, F): 0.15,
(F, T): 0.12,
(F, F): 0.003
}), ('OldAge', '', 0.147),
('Death', 'HeartAttack OldAge', {
(T, T): 0.749,
(T, F): 0.25,
(F, T): 0.0009,
(F, F): 0.001
})])
death.label = 'Death and Heart Attacks'
examples = {
death: [
{
'variable': 'Overweight',
'evidence': {
'Death': T,
},
},
{
'variable': 'Death',
'evidence': {
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask
# Utility variables
T, F = True, False
# From AIMA code (probability.py) - Fig. 14.2 - burglary example
happy = BayesNet([
('Sunny', '', .7),
('Raise', '', .01),
('Happy', 'Sunny Raise', {
(T, T): 1.0,
(T, F): 0.7,
(F, T): 0.9,
(F, F): .1
}),
])
print("\nP(Raise | Sunny)")
print(enumeration_ask('Raise', dict(Sunny=T), happy).show_approx())
# P(Raise | sunny) = alpha * <P(Raise) * P(Sunny | Raise), P(Not raise) * P(Sunny | not raise)>
# = alpha * <.01 * .7, .99 * .7>
# = alpha * <.007, .693>
# = <.01, .99> -independent events so they don't affect eachother.
#
#
print("\nP(Raise | Happy and Sunny)")
print(enumeration_ask('Raise', dict(Happy=T, Sunny=T), happy).show_approx())
# P(Raise | Happy and Sunny) = alpha * <P(Happy | Sunny and Raise) * P(Sunny) * P(Raise), P(Happy | Sunny and Not raise) * P(Sunny) * P( No raise)
# Cancer example
from probability import BayesNet
T, F = True, False
cancer = BayesNet([
('Pollution', '', 0.10),
('Smoker', '', 0.90),
('LungCancer', 'Pollution Smoker', {
(T, T): 0.987,
(T, F): 0.10,
(F, T): 0.87,
(F, F): 0.09
}),
('XRay', 'LungCancer', {
T: 0.90,
F: 0.10
}),
('Dyspnoea', 'LungCancer', {
T: 0.70,
F: 0.30
}),
('Death', 'LungCancer', {
T: 0.87,
F: 0.13
}),
])
cancer.label = 'Lung Cancer Example'
examples = {
cancer: [{
'variable': 'LungCancer',
T, F = True, False
depressedStudent = BayesNet([
('Sad', '', 0.45),
('Homework', '', 0.32),
('Tired', '', 0.7),
('Sleeping', 'Sad Homework Tired',
{(T, T, T): 0.12,
(T, T, F): 0.05,
(T, F, T): 0.65,
(T, F, F): 0.33,
(F, T, T): 0.10,
(F, T, F): 0.03,
(F, F, T): 0.89,
(F, F, F): 0.01}),
('Moving', 'Sleeping', {T: 0.5, F: 0.87}),
('Drooling', 'Sleeping', {T: 0.56, F: 0.05}),
('Outside', 'Sleeping Sad',
{(T, T): 0.005,
(T, F): 0.01,
(F, T): 0.28,
(F, F): 0.43}),
('Exercise', 'Outside Moving',
{(T, T): 0.6,
(T, F): 0.0,
(F, T): 0.34,
(F, F): 0.0}),
])
depressedStudent.label = 'Depressed Student: is he sleeping, moving, drooling or outside?'
examples = {
D1 = Decision()
D1.addChildren(C1)
D1.addChildren(C2)
return 'Value for D1 is ' + str(D1.value()) + ' and ' + 'Action from D1 is ' + D1.action
T, F = True, False
chatbot = BayesNet([
('Accurate', '', 0.9),
('ProblemSize', '', 0.9), # 0.9 is probabililty that problem size is small
('ConversationLength', 'ProblemSize', {T: (0.40, 0.40, 0.20), F: (0.20, 0.30, 0.50)}, 3),
('Frustrated', 'Accurate ProblemSize ConversationLength',
{(T, T, 0): 0.20, (T, T, 1): 0.30, (T, T, 2): 0.60,
(T, F, 0): 0.30, (T, F, 1): 0.60, (T, F, 2): 0.70,
(F, T, 0): 0.40, (F, T, 1): 0.50, (F, T, 2): 0.80,
(F, F, 0): 0.50, (F, F, 1): 0.80, (F, F, 2): 0.90}),
('Resolved', 'Accurate ConversationLength',
{(T, 0): 0.30, (T, 1): 0.50, (T, 2): 0.70,
(F, 0): 0.20, (F, 1): 0.30, (F, 2): 0.40})
])
psize_dict = {True: 'Small', False: 'Big'}
clength_dict = {0: 'Short', 1: 'Medium', 2: 'Long'}
def q8(variable, conditions, value):
# Do enumeration ask
probdist = enumeration_ask(variable, conditions, chatbot)
from probability import BayesNet
T, F = True, False
test = BayesNet([
('Study', '', 0.3),
('Sleep', '', 0.6),
('PlayGames', '', 0.1),
('Pass', 'Study Sleep PlayGames',
{(T, T, T): 0.85,
(T, T, F): 0.98,
(F, T, T): 0.1,
(F, F, T): 0.001,
(F, F, F): 0.01,
(T, F, T): 0.51,
(F, T, F): 0.21,
(T, F, F): 0.78}),
('BadDream', 'Pass', {T: 0.4, F: 0.6}),
('BeatGame', 'Pass', {T: 0.01, F: 0.2}),
('GoodDream', 'Pass', {T: 0.70, F: 0.1})
])
test.label = 'Pass Test'
examples = {
test: [
{'variable': 'Study',
'evidence': {'BadDream':T, 'GoodDream':T},
},
{'variable': 'Study',
'evidence': {'BadDream':F, 'GoodDream':T},
# Cancer example
from probability import BayesNet
T, F = True, False
cancer = BayesNet([
('Pollution', '', 0.10),
('Smoker','', 0.90),
('LungCancer', 'Pollution Smoker',
{(T, T): 0.987,
(T, F): 0.10,
(F, T): 0.87,
(F, F): 0.09}),
('XRay', 'LungCancer', {T: 0.90, F: 0.10}),
('Dyspnoea', 'LungCancer', {T: 0.70, F: 0.30}),
('Death', 'LungCancer', {T: 0.87, F: 0.13}),
])
cancer.label = 'Lung Cancer Example'
examples = {
cancer: [
{'variable': 'LungCancer',
'evidence': {'Death':T, 'Pollution':T},
},
{'variable': 'Death',
'evidence': {'LungCancer':F, 'Smoker':T},
},
{'variable': 'Smoker',
'evidence': {'LungCancer':T, 'Pollution':T},
},
{'variable': 'Pollution',
from probability import BayesNet
T, F = True, False
snowDay = BayesNet([
('snow', '', 0.022),
('windChill', '', 0.017),
('advisory', 'snow windChill',
{(T, T): 0.62,
(T, F): 0.48,
(F, T): 0.41,
(F, F): 0.0001}),
('noSchool', 'advisory', {T: 0.70, F: 0.005}),
('extremeWarning', 'advisory', {T: 0.16, F: 0.002})
])
snowDay.label = 'snowDay shows various weather probabilities that effect the chances of a snow day in Buffalo, New York.'
examples = {
snowDay: [
{'variable': 'snow',
'evidence': {'noSchool':T, 'extremeWarning':T},
},
{'variable': 'windChill',
'evidence': {'noSchool': F, 'advisory': T},
},
{'variable': 'advisory',
'evidence': {'noSchool': T, 'windChill': F},
},
{'variable': 'extremeWarning',
'evidence': {'snow': T, 'windChill': T},
},
def __init__(self, file):
""" Initializes the problem with an opened file.
Creates the bayesian network onwhich every room is a child of itself
and the adjacent rooms at timestep t-1 and is parent of sensors that
exist on the room and make measures at time t.
"""
self.rooms = {}
self.sensors = {}
self.P = 0
self.measure = []
# Fills the problem variables
self.read_file(file)
# Constructs the bayesian network
str_temp = "{}_t_{}"
bayes_template = []
# First constructs the base network at time 0
for room in self.rooms.values():
bayes_template.append((str_temp.format(room.name, 0), '', 0.5))
for sensor_id in self.measure[0]:
sensor = self.sensors[sensor_id]
bayes_template.append(
(str_temp.format(sensor.id,
0), str_temp.format(sensor.room, 0), {
True: sensor.TPR,
False: sensor.FPR
}))
# Then adds every timestep connecting rooms from the timestep before and sensors needed
for k in range(1, len(self.measure)):
for room in self.rooms.values():
aux_str = str_temp.format(room.name, k - 1)
for parent in room.adj_rooms:
aux_str += " " + str_temp.format(parent, k - 1)
truth_table = self.create_truth_table(
len(room.adj_rooms) + 1) if room.adj_rooms else {
True: 1,
False: 0
}
bayes_template.append(
(str_temp.format(room.name, k), aux_str, truth_table))
for sensor_id in self.measure[k]:
sensor = self.sensors[sensor_id]
bayes_template.append(
(str_temp.format(sensor.id,
k), str_temp.format(sensor.room, k), {
True: sensor.TPR,
False: sensor.FPR
}))
self.network_template = BayesNet(bayes_template)
# Burglary example [Figure 14.2]
from probability import BayesNet
T, F = True, False
nashvilleWeather = BayesNet([
('Rain', '', 0.36),
('Freezing', '', 0.27),
('CarAccident', 'Rain Freezing',
{(T, T): 0.95,
(T, F): 0.79,
(F, T): 0.29,
(F, F): 0.001}),
('InjuriesReported', 'CarAccident', {T: 0.89, F: 0.09}),
('DeathReported', 'CarAccident', {T: 0.95, F: 0.02})
])
nashvilleWeather.label = 'Nashville Weather Correlation With Accidents (Hypothetical)'
examples = {
nashvilleWeather: [
{'variable': 'Rain',
'evidence': {'InjuriesReported': T, 'DeathReported': T},
},
{'variable': 'Freezing',
'evidence': {'InjuriesReported': T, 'DeathReported': T},
},
{'variable': 'CarAccident',
'evidence': {'InjuriesReported': T, 'Rain': F},
},
{'variable': 'DeathReported',
'evidence': {'Rain': T, 'Freezing': T}
@author: kvlinden
@edited: Quentin Barnes
@version Jan 2, 2013
'''
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask
# Utility variables
T, F = True, False
# From lab 05 - Cs 344
burglary = BayesNet([('Sunny', '', 0.70), ('Raise', '', 0.01),
('Happy', 'Sunny Raise', {
(T, T): 1.0,
(T, F): 0.7,
(F, T): 0.9,
(F, F): 0.1
})])
print('P(Raise | sunny): ')
print(enumeration_ask('Raise', dict(Sunny=T), burglary).show_approx())
print('\nP(Raise | happy ∧ sunny): ')
print(enumeration_ask('Raise', dict(Happy=T, Sunny=T), burglary).show_approx())
print('\nP(Raise | happy): ')
print(enumeration_ask('Raise', dict(Happy=T), burglary).show_approx())
print('\nP(Raise | happy ∧ ¬sunny): ')
print(enumeration_ask('Raise', dict(Happy=T, Sunny=F), burglary).show_approx())
@version March 6, 2020
'''
import sys
sys.path.append('/home/neg6/cs344/cs344-code/tools/aima')
sys.path.append(
'D:\\School\\2019-20 Spring\\CS 344 AI\\cs344-code\\tools\\aima')
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask
# Utility variables
T, F = True, False
# From AIMA code (probability.py) - Fig. 14.2 - burglary example
happy = BayesNet([('Sunny', '', 0.7), ('Raise', '', .01),
('Happy', 'Sunny Raise', {
(T, T): 1,
(T, F): .7,
(F, T): .9,
(F, F): .1
})])
# P(Raise | sunny)
# <.01, .99>
print("P(Raise | sunny):")
print(enumeration_ask('Raise', dict(Sunny=T), happy).show_approx())
# P(Raise | happy ^ sunny)
# x * <P(raise | happy) * P(raise | sunny) * P(raise), P(!raise | happy) * P(!raise | sunny) * P(!raise)>
# x * <(.9 * .01 * .01), (.07 * .99 * .99)>
# x * <.00009, .06861>
# <.00009/.0687, .06861/.0687>
# <.00131, .99869>
print("P(Raise | happy, sunny):")
from probability import BayesNet
T, F = True, False
test = BayesNet([('Study', '', 0.3), ('Sleep', '', 0.6),
('PlayGames', '', 0.1),
('Pass', 'Study Sleep PlayGames', {
(T, T, T): 0.85,
(T, T, F): 0.98,
(F, T, T): 0.1,
(F, F, T): 0.001,
(F, F, F): 0.01,
(T, F, T): 0.51,
(F, T, F): 0.21,
(T, F, F): 0.78
}), ('BadDream', 'Pass', {
T: 0.4,
F: 0.6
}), ('BeatGame', 'Pass', {
T: 0.01,
F: 0.2
}), ('GoodDream', 'Pass', {
T: 0.70,
F: 0.1
})])
test.label = 'Pass Test'
examples = {
test: [
{
'variable': 'Study',
@author: Sarah Hendriksen (srh34)
@startedby: kvlinden
@version Jan 2, 2013
'''
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask
# Utility variables
T, F = True, False
# From AIMA code (probability.py) - Fig. 14.2 - burglary example
happiness = BayesNet([
('Sunny', '', 0.70),
('Raise', '', 0.01),
('Happy', 'Sunny Raise', {
(T, T): 1.0,
(T, F): 0.70,
(F, T): 0.90,
(F, F): 0.10
}),
])
# Exercise 5.3
# a)
# i)
print(enumeration_ask('Raise', dict(Sunny=T), happiness).show_approx())
# Output - False: .99, True: .01
# Hand-computed -
# P(R | s) = < .01, .99 > --> they are independent events so it is just P(R)
# ii)
print(
from probability import BayesNet
T, F = True, False
grass = BayesNet([('Raining', '', 0.1), ('Sprinklers', '', 0.4),
('GrassWet', 'Raining Sprinklers', {
(T, T): 0.99,
(T, F): 0.82,
(F, T): 0.94,
(F, F): 0.001
}), ('DadIsHappy', 'GrassWet', {
T: 0.90,
F: 0.05
}),
('AquaphobicNeighborIsNotHappy', 'GrassWet', {
T: 0.70,
F: 0.01
})])
grass.label = 'WetGrass'
examples = {
grass: [
{
'variable': 'Raining',
'evidence': {
'DadIsHappy': T,
'AquaphobicNeighborIsNotHappy': T
},
},
{
'variable': 'Sprinklers',
# Burglary example [Figure 14.2]
from probability import BayesNet
T, F = True, False
coffee = BayesNet([
('Tired', '', 0.24),
('Working', '', 0.45),
('Coffee', 'Tired Working',
{(T, T): 0.97,
(T, F): 0.38,
(F, T): 0.29,
(F, F): 0.05}),
('Energy', 'Coffee', {T: 0.83, F: 0.21}),
('Jitters', 'Coffee', {T: 0.19, F: 0.01})
])
coffee.label = 'Probability of getting coffee'
examples = {
coffee: [
{'variable': 'Tired',
'evidence': {'Energy':T, 'Jitters':T},
},
{'variable': 'Tired',
'evidence': {'Energy':F, 'Jitters':T},
},
{'variable': 'Working',
'evidence': {'Energy':T, 'Jitters':T},
},
],
}
from probability import BayesNet
T, F = True, False
compile_error = BayesNet([
('MissingSemicolon', '', 0.5),
('ExtraSemicolon', '', 0.2),
('IndexOutOfRange', '', 0.1),
('CompileError', 'MissingSemicolon ExtraSemicolon IndexOutOfRange',
{
(T, T, T): 0.2,
(T, T, F): 0.5,
(T, F, T): 0.1,
(T, F, F): 0.7,
(F, T, T): 0.1,
(F, T, F): 0.4,
(F, F, T): 0.75,
(F, F, F): 0.09
}),
('CantFindSemiColon', 'CompileError', {T: 0.7, F: 0.3}),
('ArrayIsNull', 'CompileError', {T: 0.3, F: 0.85}),
('OhYeahForgotThat', 'CompileError', {T: 0.8, F: 0.1})
])
compile_error.label = 'Programming Compile Error Correlation With Cause (Hypothetical)'
examples = {
compile_error: [
{'variable': 'ExtraSemicolon',
'evidence': {'CantFindSemiColon': T, 'OhYeahForgotThat': T}
},
@author: Sarah Hendriksen (srh34)
@startedby: kvlinden
@version Jan 2, 2013
'''
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask
# Utility variables
T, F = True, False
# From AIMA code (probability.py) - Fig. 14.2 - burglary example
cancer = BayesNet([('Cancer', '', 0.01), ('Test1', 'Cancer', {
T: 0.90,
F: 0.20
}), ('Test2', 'Cancer', {
T: 0.90,
F: 0.20
})])
# Exercise 5.2
# a)
print(enumeration_ask('Cancer', dict(Test1=T, Test2=T), cancer).show_approx())
# Output - False: 0.83, True: 0.17
# Hand-coding -
# P(C | t1 = T, t2 = T) = alpha P(C, t1 = T, t2 = T)
# = alpha < P(C) * P(t1 = T | C) * P(t2 = T | C),
# P(not C) * P(t1 = T | not C) * P(t2 = T | not C)>
# = alpha < .01 * .9 * .9, .99 * .2 * .2 >
# = alpha < .0081, .0396 >
# = < .170, .830 >
def Q1_1_2():
print("=" * 80)
print("ANSWER 1.1.2")
T, F = True, False
bayes_net = BayesNet([('AI', '', 0.8), ('FossilFuel', '', 0.4),
('RenewableEnergy', 'AI FossilFuel', {
(T, T): 0.2,
(T, F): 0.7,
(F, T): 0.02,
(F, F): 0.5
}), ('Traffic', 'FossilFuel', {
T: 0.95,
F: 0.1
}),
('GlobalWarming', 'RenewableEnergy Traffic', {
(T, T): 0.6,
(T, F): 0.4,
(F, T): 0.95,
(F, F): 0.55
}),
('Employed', 'AI GlobalWarming', {
(T, T): 0.01,
(T, F): 0.03,
(F, T): 0.03,
(F, F): 0.95
})])
#print(bayes_net.variable_node('GlobalWarming').cpt)
p_employed = enumeration_ask(X='Employed',
e={
'AI': True,
'FossilFuel': True
},
bn=bayes_net)
print('-' * 80)
print(f"given AI=true and FossilFuel=True:",
f"\n\t\tP(Employed)\t\t=\t\t{p_employed.show_approx()}")
print('-' * 80)
p_global_warming = elimination_ask(X='GlobalWarming',
e={
'Employed': False,
'Traffic': False
},
bn=bayes_net)
print('-' * 80)
print(f"Given Employed=False and Traffic=False, \n\t\tP(GlobalWarming)\t=",
f"\t\t{p_global_warming.show_approx()}")
print('-' * 80)
p_ai = elimination_ask(X='AI',
e={
'RenewableEnergy': True,
'GlobalWarming': True,
'Employed': True,
'Traffic': True,
'FossilFuel': True
},
bn=bayes_net)
print('-' * 80)
print(
f"Given RenewableEnergy=T, GlobalWarming=T",
f"Employed=T, Traffic=T, FossilFuel=T, \n\t\tP(AI)\t\t\t=\t\t{p_ai.show_approx()}"
)
print('-' * 80)
# Burglary example [Figure 14.2]
from probability import BayesNet
T, F = True, False
burglary = BayesNet([
('Burglary', '', 0.001),
('Earthquake', '', 0.002),
('Alarm', 'Burglary Earthquake',
{(T, T): 0.95,
(T, F): 0.94,
(F, T): 0.29,
(F, F): 0.001}),
('JohnCalls', 'Alarm', {T: 0.90, F: 0.05}),
('MaryCalls', 'Alarm', {T: 0.70, F: 0.01})
])
burglary.label = 'Burglary Example, Fig. 14.2'
examples = {
burglary: [
{'variable': 'Burglary',
'evidence': {'JohnCalls':T, 'MaryCalls':T},
},
{'variable': 'Burglary',
'evidence': {'JohnCalls':F, 'MaryCalls':T},
},
{'variable': 'Earthquake',
'evidence': {'JohnCalls':T, 'MaryCalls':T},
},
],
}
@date: 03/08/2019
"""
from probability import BayesNet, enumeration_ask, elimination_ask
# Utility variables
T, F = True, False
# From AIMA code (probability.py) - Fig. 14.12 - A multiply connected network with conditional probability tables
exercises = BayesNet([('Cloudy', '', 0.5),
('Sprinkler', 'Cloudy', {
T: 0.10,
F: 0.50
}), ('Rain', 'Cloudy', {
T: 0.80,
F: 0.20
}),
('WetGrass', 'Sprinkler Rain', {
(T, T): 0.99,
(T, F): 0.90,
(F, T): 0.90,
(F, F): 0.00
})])
print("P(Cloudy)")
print(enumeration_ask('Cloudy', dict(), exercises).show_approx())
print(elimination_ask('Cloudy', dict(), exercises).show_approx())
print("P(Sprinker | cloudy)")
print(enumeration_ask('Sprinkler', dict(Cloudy=T), exercises).show_approx())
print(elimination_ask('Sprinkler', dict(Cloudy=T), exercises).show_approx())
print("P(Cloudy| the sprinkler is running and it’s not raining)")
print(
# (T, F): 0.02,
# (F, T): 0.0368,
# (F, F): 0.04}),
# ('RedOnTop', 'RedKing', {T: 0.03846, F: 0.076923}),
# # ('MaryCalls', 'Alarm', {T: 0.70, F: 0.01})
# ])
# cards.label = 'Cards'
smoking = BayesNet([
('Female', '', 0.51),
('LGB', '', 0.054),
('Smoker', 'LGB Female',
{(T, T): 0.3047,
(T, F): 0.3046,
(F, T): 0.1885,
(F, F): 0.2202}),
('LungCancer', 'Smoker Female',
{(T, T): 0.095,
(T, F): 0.159,
(F, T): 0.004,
(F, F): 0.002}),
('PrematureDeath', 'LungCancer', {T: 0.86, F: 0.0134}),
# ('Smoker', 'LGB', {T: 0.205, F: 0.153})
])
smoking.label = 'Smoking Example'
examples = {
smoking: [
{'variable': 'LungCancer',
'evidence': {'LGB':T},
},
{'variable': 'LGB',
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask
# Utility variables
T, F = True, False
# From AIMA code (probability.py) - Fig. 14.2 - burglary example
cancer = BayesNet([('Cancer', '', .01), ('Test1', 'Cancer', {
T: .9,
F: .2
}), ('Test2', 'Cancer', {
T: .9,
F: .2
})])
print("\nP(Cancer | Test1 and Test2)")
print(enumeration_ask('Cancer', dict(Test1=T, Test2=T), cancer).show_approx())
print("\nP(Cancer | Test1 and negative Test2)")
print(enumeration_ask('Cancer', dict(Test1=T, Test2=F), cancer).show_approx())
#
# P(Cause | T1 And T2) = alpha * <P(C) * P(T1 | C) * P(T2 | C), P( not C) * P(T1 | not C) * P(T2 | not C)>
# = alpha * <.01 * .9 * .9, .99 * .2 * .2>
# = alpha * <.0081, .0396>
# = <.17, .83> (after normalization)
# P(C | T1 and T2) = alpha * <P(C) * P(T1 | C) * P(not T2 | C), P(not C) * P(T1 | not C) * P(not T2 | Not C)>
# = alpha * <.01 * .9 * .1, .99 * .2 * .8>
# = alpha * <.0009, .1584>
# = <.00565, .994> (after normalization)
#
CarAccidents = BayesNet([
('Texting', '', 0.002),
('Speeding', '', 0.002),
('Weather', '', 0.001),
('Alcohol', '', 0.003),
('AccidentA', 'Texting Speeding Weather Alcohol', {
(T, T, T, T): 0.90,
(F, T, T, T): 0.74,
(T, F, T, T): 0.59,
(T, T, F, T): 0.79,
(T, T, T, F): 0.58,
(F, F, T, T): 0.43,
(F, T, F, T): 0.63,
(F, T, T, F): 0.48,
(T, F, F, T): 0.42,
(T, F, T, F): 0.27,
(T, T, F, F): 0.47,
(T, F, F, F): 0.16,
(F, T, F, F): 0.31,
(F, F, T, F): 0.11,
(F, F, F, T): 0.32,
(F, F, F, F): 0.005
}),
# Attempt to try separating each kind of accident
# ('AccidentB', 'Texting Weather',
# {(T, T): 0.37,
# (T, F): 0.16,
# (F, T): 0.11,
# (F, F): 0.05}),
# ('AccidentC', 'Texting Alcohol',
# {(T, T): 0.58,
# (T, F): 0.16,
# (F, T): 0.32,
# (F, F): 0.05}),
# ('AccidentD', 'Speeding Weather',
# {(T, T): 0.52,
# (T, F): 0.31,
# (F, T): 0.11,
# (F, F): 0.05}),
# ('AccidentE', 'Speeding Alcohol',
# {(T, T): 0.73,
# (T, F): 0.31,
# (F, T): 0.32,
# (F, F): 0.05}),
# ('AccidentF', 'Weather Alcohol',
# {(T, T): 0.53,
# (T, F): 0.11,
# (F, T): 0.32,
# (F, F): 0.05}),
('HitCar', 'AccidentA', {
T: 0.40,
F: 0.04
}),
('HitGuardRail', 'AccidentA', {
T: 0.50,
F: 0.02
}),
# Attempt for each of the aforementioned to have their own probabilities
# ('HitCar', 'AccidentB', {T: 0.60, F: 0.40}),
# ('HitGuardRail', 'AccidentB', {T: 0.80, F: 0.20}),
#
# # ('HitCar', 'AccidentC', {T: 0.60, F: 0.40}),
# ('HitGuardRail', 'AccidentC', {T: 0.80, F: 0.20}),
#
# ('HitCar', 'AccidentD', {T: 0.60, F: 0.40}),
# ('HitGuardRail', 'AccidentD', {T: 0.80, F: 0.20}),
#
# ('HitCar', 'AccidentE', {T: 0.60, F: 0.40}),
# ('HitGuardRail', 'AccidentE', {T: 0.80, F: 0.20}),
#
# ('HitCar', 'AccidentF', {T: 0.60, F: 0.40}),
# ('HitGuardRail', 'AccidentF', {T: 0.80, F: 0.20})
])
depressedStudent = BayesNet([
('Sad', '', 0.45),
('Homework', '', 0.32),
('Tired', '', 0.7),
('Sleeping', 'Sad Homework Tired', {
(T, T, T): 0.12,
(T, T, F): 0.05,
(T, F, T): 0.65,
(T, F, F): 0.33,
(F, T, T): 0.10,
(F, T, F): 0.03,
(F, F, T): 0.89,
(F, F, F): 0.01
}),
('Moving', 'Sleeping', {
T: 0.5,
F: 0.87
}),
('Drooling', 'Sleeping', {
T: 0.56,
F: 0.05
}),
('Outside', 'Sleeping Sad', {
(T, T): 0.005,
(T, F): 0.01,
(F, T): 0.28,
(F, F): 0.43
}),
('Exercise', 'Outside Moving', {
(T, T): 0.6,
(T, F): 0.0,
(F, T): 0.34,
(F, F): 0.0
}),
])
'''
This module implements the Bayesian network for exercise 5.2 in lab05 of CS344
@author: brentritzema
@version Mar 1, 2019
'''
from probability import BayesNet, enumeration_ask, elimination_ask
# Utility variables
T, F = True, False
cancer = BayesNet([
('Cancer', '', 0.01),
('Test1', 'Cancer', {T: 0.9, F: 0.2}),
('Test2', 'Cancer', {T: 0.9, F: 0.2}),
])
# a. P(Cancer | Test1 ^ Test2)
# This calculates the probability of cancer given that test1 and test2 were positive
print(enumeration_ask('Cancer', dict(Test1=T, Test2=T), cancer).show_approx())
print(elimination_ask('Cancer', dict(Test1=T, Test2=T), cancer).show_approx())
# b. P(Cancer | Test1 ^ -Test2)
# This calculates the probability of cancer given that test1 was positive and test2 was negative
print(enumeration_ask('Cancer', dict(Test1=T, Test2=F), cancer).show_approx())
print(elimination_ask('Cancer', dict(Test1=T, Test2=F), cancer).show_approx())
# These results make sense. One failed test reduces the likelihood of having cancer by two orders of magnitude
# Here is P(Cancer | Test1 ^ Test2) worked out by hand
# Burglary example [Figure 14.2]
from probability import BayesNet
T, F = True, False
burglary = BayesNet([('Burglary', '', 0.001), ('Earthquake', '', 0.002),
('Alarm', 'Burglary Earthquake', {
(T, T): 0.95,
(T, F): 0.94,
(F, T): 0.29,
(F, F): 0.001
}), ('JohnCalls', 'Alarm', {
T: 0.90,
F: 0.05
}), ('MaryCalls', 'Alarm', {
T: 0.70,
F: 0.01
})])
burglary.label = 'Burglary Example, Fig. 14.2'
famous = BayesNet([
# will you get famous?
# statistics recevied from https://nypost.com/2009/07/25/you-are-not-going-to-be-famous/ and
('liveInAmerica', '', 0.04),
('liveInNashville', '', 0.002),
('Famous', 'liveInAmerica liveInNashville', {
(T, T): 0.000000024,
(T, F): 0.000000024,
(F, T): 0.0000000012,
(F, F): 0.0000006
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask
# Utility variables
T, F = True, False
happyNetwork = BayesNet([('Sunny', '', 0.7), ('Raise', '', 0.01),
('Happy', 'Sunny Raise', {
(T, T): 1.0,
(T, F): 0.7,
(F, T): 0.9,
(F, F): 0.1
})])
# a.i
print(enumeration_ask('Raise', dict(Sunny=T), happyNetwork).show_approx())
# This is really a depressingly low probability, sitting at an even 0.01, which is just straight up the probability
# of getting a raise. Which is really low.
# a.ii
print(
enumeration_ask('Raise', dict(Happy=T, Sunny=T),
happyNetwork).show_approx())
# This is not very much more likely, sitting at .0142. The fact that you are happy does eliminate the times when
# you are not happy despite it being sunny from the equation.
# b.i
print(enumeration_ask('Raise', dict(Happy=T), happyNetwork).show_approx())
# This is yet slightly more likely again, since you aren't always happy when it is sunny, but you are almost always
# happy when you have gotten a raise. So the fact that you are happy makes it more likely that you got a raise than
# the fact that it is sunny does, even though the probability of having gotten a raise is still pretty miniscule.
# Burglary example [Figure 14.2]
from probability import BayesNet
T, F = True, False
burglary = BayesNet([('Burglary', '', 0.001), ('Earthquake', '', 0.002),
('Alarm', 'Burglary Earthquake', {
(T, T): 0.95,
(T, F): 0.94,
(F, T): 0.29,
(F, F): 0.001
}), ('JohnCalls', 'Alarm', {
T: 0.90,
F: 0.05
}), ('MaryCalls', 'Alarm', {
T: 0.70,
F: 0.01
})])
burglary.label = 'Burglary Example, Fig. 14.2'
storm = BayesNet([
('Thunderstorm', '', 0.005),
('Earthquake', '', 0.002),
('Rain', '', 0.02),
('Loud', 'Thunderstorm Earthquake Rain', {
(T, T, T): 0.99,
(T, F, T): 0.65,
(T, T, F): 0.95,
(T, F, F): 0.85,
(F, T, T): 0.80,
(F, T, F): 0.50,
# Accidents on the Nigerian Road
from probability import BayesNet
T, F = True, False
Aids = BayesNet([
('HIV', '', 0.1),
('Aids', '', 0.000982),
('Over30', '', 0.9),
('Dead', 'HIV Aids Over30',
{
(T, T, T): 0.90,
(T, T, F): 0.13,
(T, F, T): 0.85,
(T, F, F): 0.07,
(F, T, T): 0.0,
(F, T, F): 0.0,
(F, F, T): 0.99,
(F, F, F): 0.95
}),
('Jason', 'Dead', {T: 0.01, F: 0.90}),
('Lauren', 'Dead', {T: 0.20, F: 0.50}),
('Zach', 'Dead', {T: 0.49, F: 0.11})
])
Aids.label = 'Probability of Death with HIV/AIDS'
examples = {
Aids: [
{'variable': 'HIV',
It's taken from the AIMA Python code.
@author: kvlinden
@version Jan 2, 2013
'''
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask
# Utility variables
T, F = True, False
happiness = BayesNet([
('Sunny', '', 0.7),
('Raise', '', 0.01),
('Happy', 'Sunny Raise', {
(T, T): 1,
(T, F): 0.7,
(F, T): 0.9,
(F, F): 0.1
}),
])
# Exercise 5.3
# The Bayesian network shown on the right represents a happiness domain in which either the sun or a raise in pay can
# increase the happiness of an agent.
#
# a. Implement this network shown and use it to compute the following probabilities:
#
# i. P(Raise | sunny)
print("P(Raise | sunny): ",
enumeration_ask('Raise', dict(Sunny=T), happiness).show_approx())
# ii. P(Raise | happy ∧ sunny)
@version Jan 2, 2013
Updated by: Tyler Poel, for lab05 in CS 344, at Calvin University
Date: March 6, 2020
'''
from probability import BayesNet, enumeration_ask, elimination_ask, gibbs_ask, likelihood_weighting, rejection_sampling
# Utility variables
T, F = True, False
# From AIMA code (probability.py) - Fig. 14.2 - burglary example
burglary = BayesNet([
('Burglary', '', 0.001),
('Earthquake', '', 0.002),
('Alarm', 'Burglary Earthquake', {(T, T): 0.95, (T, F): 0.94, (F, T): 0.29, (F, F): 0.001}),
('JohnCalls', 'Alarm', {T: 0.90, F: 0.05}),
('MaryCalls', 'Alarm', {T: 0.70, F: 0.01})
])
# Compute P(Burglary | John and Mary both call).
print(enumeration_ask('Burglary', dict(JohnCalls=T, MaryCalls=T), burglary).show_approx())
# elimination_ask() is a dynamic programming version of enumeration_ask().
print(elimination_ask('Burglary', dict(JohnCalls=T, MaryCalls=T), burglary).show_approx())
# gibbs_ask() is an approximation algorithm helps Bayesian Networks scale up.
print(gibbs_ask('Burglary', dict(JohnCalls=T, MaryCalls=T), burglary).show_approx())
# See the explanation of the algorithms in AIMA Section 14.4.
# rejection_sampling() and likelihood_weighting() are also approximation algorithms.
print(rejection_sampling('Burglary', dict(JohnCalls=T, MaryCalls=T), burglary).show_approx())
print(likelihood_weighting('Burglary', dict(JohnCalls=T, MaryCalls=T), burglary).show_approx())
# Burglary example [Figure 14.2]
from probability import BayesNet
T, F = True, False
death = BayesNet([
('Overweight', '', 0.339),
('HighStress', '', 0.72),
('HeartAttack', 'Overweight HighStress',
{(T, T): 0.72,
(T, F): 0.15,
(F, T): 0.12,
(F, F): 0.003}),
('OldAge', '', 0.147),
('Death', 'HeartAttack OldAge',
{(T, T): 0.749,
(T, F): 0.25,
(F, T): 0.0009,
(F, F): 0.001})
])
death.label = 'Death and Heart Attacks'
examples = {
death: [
{'variable': 'Overweight',
'evidence': {'Death':T,},
},
{'variable': 'Death',
'evidence': {'HeartAttack':T, 'OldAge':F},
},
{'variable': 'HighStress',
