Ejemplo n.º 1
0
def exact_bayesian():
    '''
    bayesian_inference
    '''
    A = Factor(name='A',
                values=[0, 1],
                cpd=[0.5, 0.5])
    B = Factor(name='B',
                values=[0, 1],
                cpd=[0.5, 0.5])
    C = Factor(name='C|A,B',
                values=[0, 1],
                cond=[A, B],
                cpd=[1, 0, 1, 0, 1, 0, 0, 1])
    print C.query(query=[A], evidence={C: 0, B: 0})
Ejemplo n.º 2
0
def bayesian_inference():
    '''
    bayesian_inference
    '''
    A = Factor(name='A',
                values=[0, 1],
                cpd=[0.6, 0.4])
    B = Factor(name='B',
                values=[0, 1],
                cpd=[0.7, 0.3])
    C = Factor(name='C|A,B', full='C',
                values=[0, 1],
                cond=[A, B],
                cpd=[0.7, 0.3, 0.9, 0.1, 0.8, 0.2, 0.6, 0.4])
    print C.joint().marginal(B.var[-1]).reduce(var=C.var[-1], value=0).norm()
Ejemplo n.º 3
0
def bayesian_inference():
    '''
    bayesian_inference
    '''
    A = Factor(name='A',
                values=[0, 1],
                cpd=[0.6, 0.4])
    B = Factor(name='B',
                values=[0, 1],
                cpd=[0.7, 0.3])
    C = Factor(name='C',
                values=[0, 1],
                cond=[A, B],
                cpd=[0.7, 0.3, 0.9, 0.1, 0.8, 0.2, 0.6, 0.4])
    print C
    print C.joint().marginal(B.var[-1]).reduce(var=C.var[-1], value=0)._norm()
    print C.query([A], {C: 0})
    print C.query([A], [C])
Ejemplo n.º 4
0
def student_example():
    '''
    bayesian_inference
    '''
    D = Factor(name='Difficulty', values=[0, 1], cpd=[0.6, 0.4])
    I = Factor(name='Intelligence', values=['low', 'high'], cpd=[0.7, 0.3])
    G = Factor(name='Grade|I,D', values=[1, 2, 3],
                cond=[D, I], cpd=[0.3, 0.4, 0.3,
                                0.05, 0.25, 0.7,
                                0.9, 0.08, 0.02,
                                0.5, 0.3, 0.2],)
    S = Factor(name='SAT|I', values=[0, 1], cond=[I],
                cpd=[0.95, 0.05, 0.2, 0.8])
    L = Factor(name='Letter|G', values=[0, 1], cond=[G],
                cpd=[0.1, 0.9, 0.4, 0.6, 0.99, 0.01])

    BN = Bayesian([D, I, S, G, L])

    print  BN.joint().query(query=[I], evidence={G:3, D:1})
    print  BN.joint().query(query=[I], evidence=[G, D])
    print  BN.joint().query(query=[I], evidence=[D])
    print  BN.joint().query(query=[I], evidence=[G])
Ejemplo n.º 5
0
def cancer_example():
    '''
    Cancer example
    '''
    C = Factor(name='Cancer', values=["no", "yes"], cpd=[0.999, 0.001])
    T = Factor(name='Test',
               values=["pos", "neg"],
               cond=[C],
               cpd=[0.2, 0.8, 0.9, 0.1])
    print C
    print T
    print T.joint()
    print T.uncond()
    R = T.joint().query2(query=[C], evidence={T: 'pos'})
    R.name = "Tested once"
    print R
Ejemplo n.º 6
0
def exact_bayesian():
    '''
    bayesian_inference
    '''
    A = Factor(name='A', values=[0, 1], cpd=[0.5, 0.5])
    B = Factor(name='B', values=[0, 1], cpd=[0.5, 0.5])
    C = Factor(name='C|A,B',
               full='C',
               values=[0, 1],
               cond=[A, B],
               cpd=[1, 0, 1, 0, 1, 0, 0, 1])
    print C.query2(query=[A], evidence={C: 0, B: 0})
Ejemplo n.º 7
0
def cancer_example():
    '''
    Cancer example
    '''
    C = Factor(name='Cancer', values=["no", "yes"],
                cpd=[0.999, 0.001])
    T = Factor(name='Test', values=["pos", "neg"],
                cond=[C], cpd=[0.2, 0.8, 0.9, 0.1])
##    print C
##    print T
##    print T.joint()
##    print T.uncond()
##    R = T.query(query=[C], evidence={T:'pos'})
    print "First algorithm"
    print T._query1(query=[C], evidence=[T])
    print "Second algorithm"
    print T._query2(query=[C], evidence=[T])
Ejemplo n.º 8
0
def cancer_example():
    '''
    Cancer example
    '''
    C = Factor(name='Cancer', values=["no", "yes"],
                cpd=[0.999, 0.001])
    T = Factor(name='Test', values=["pos", "neg"],
                cond=[C], cpd=[0.2, 0.8, 0.9, 0.1])
    print C
    print T
    print T.joint()
    print T.uncond()
    R = T.joint().query2(query=[C], evidence={T:'pos'})
    R.name = "Tested once"
    print R
Ejemplo n.º 9
0
def local_inference():
    A = Factor(name='A', values=[0, 1], cpd=[0.2, 0.8])
    B = Factor(name='B', values=[0, 1], cond=[A], cpd=[0.6, 0.4, 0.3, 0.7])
    C = Factor(name='C', values=[0, 1], cond=[B], cpd=[0.6, 0.4, 0.3, 0.7])
    D = Factor(name='D', values=[0, 1], cond=[C], cpd=[0.6, 0.4, 0.3, 0.7])
    E = Factor(name='E', values=[0, 1], cond=[D], cpd=[0.6, 0.4, 0.3, 0.7])
    F = Factor(name='F', values=[0, 1], cond=[E], cpd=[0.6, 0.4, 0.3, 0.7])
    G = Factor(name='G', values=[0, 1], cond=[F], cpd=[0.6, 0.4, 0.3, 0.7])

##    print A.uncond()
##    print B.uncond()
##    print C.uncond()
##    print D.uncond()
##    print E.uncond()
##    print F.uncond()

##    print B.uncond()
##    print B.query(query=[B], evidence=[])
##    print (B*A).marginal(var=A.cons)

##    print C.uncond()
##    print C.query(query=[C], evidence=[])

##    print C.query(query=[C], evidence=[B])
##    print C

    print C._query2(query=[C], evidence=[A, B])
##    print C.uncond()
##    print (C*B*A).marginal(A.cons).marginal(B.cons)
##    print (C*D).marginal(var=D.cons)
##    print C.query(query=[C], evidence=[A, B])


    A = Factor(name='A', values=[0, 1], cpd=[0.01, 0.99])
    B = Factor(name='B', values=[0, 1], cond=[A], cpd=[0.96, 0.04, 0.93, 0.07])
    C = Factor(name='C', values=[0, 1], cond=[B], cpd=[0.96, 0.04, 0.93, 0.07])
    D = Factor(name='D', values=[0, 1], cond=[C], cpd=[0.96, 0.04, 0.93, 0.07])
    E = Factor(name='E', values=[0, 1], cond=[D], cpd=[0.96, 0.04, 0.93, 0.07])
    F = Factor(name='F', values=[0, 1], cond=[E], cpd=[0.96, 0.04, 0.93, 0.07])
    G = Factor(name='G', values=[0, 1], cond=[F], cpd=[0.96, 0.04, 0.93, 0.07])