def test3(self):
examples = [
['a', 1, 0, 1, 0, 1],
['b', 1, 0, 0, 0, 1],
['c', 1, 0, 1, 0, 0],
['d', 0, 1, 1, 0, 0],
['e', 0, 1, 1, 1, 0],
]
nvars = len(examples[0]) - 2
attrs = list(range(1, nvars + 1))
npvals = [0] + [2] * nvars + [2]
qBounds = [0] + [0] * nvars + [0]
mdl = CrossValidate.makeNBClassificationModel(examples, attrs, npvals,
qBounds)
nWrong = 0
for eg in examples:
p = mdl.ClassifyExample(eg)
if p != eg[-1]:
nWrong += 1
self.assertEqual(nWrong, 1)
bitEx = []
for eg in examples:
newEg = [eg[0], None, eg[-1]]
bv = ExplicitBitVect(nvars)
for i in range(nvars):
if eg[i + 1]:
bv.SetBit(i)
newEg[1] = bv
bitEx.append(newEg)
attrs = list(range(nvars))
mdl2 = CrossValidate.makeNBClassificationModel(bitEx,
attrs,
npvals,
qBounds,
useSigs=True)
nWrong = 0
for eg in bitEx:
p = mdl2.ClassifyExample(eg)
if p != eg[-1]:
nWrong += 1
self.assertEqual(nWrong, 1)
# now compare:
for i in range(len(bitEx)):
eg = examples[i]
p1 = mdl.ClassifyExample(eg)
bitEg = bitEx[i]
p2 = mdl2.ClassifyExample(bitEg)
self.assertEqual(p1, p2)
v1 = mdl.GetClassificationDetails()
v2 = mdl.GetClassificationDetails()
self.assertAlmostEqual(v1, v2, 4)
def test3(self):
examples = [
['a', 1, 0, 1, 0, 1],
['b', 1, 0, 0, 0, 1],
['c', 1, 0, 1, 0, 0],
['d', 0, 1, 1, 0, 0],
['e', 0, 1, 1, 1, 0],
]
nvars = len(examples[0]) - 2
attrs = range(1, nvars + 1)
npvals = [0] + [2] * nvars + [2]
qBounds = [0] + [0] * nvars + [0]
mdl = CrossValidate.makeNBClassificationModel(examples, attrs, npvals, qBounds)
nWrong = 0
for eg in examples:
p = mdl.ClassifyExample(eg)
if p != eg[-1]:
nWrong += 1
self.assertEqual(nWrong, 1)
bitEx = []
for eg in examples:
newEg = [eg[0], None, eg[-1]]
bv = ExplicitBitVect(nvars)
for i in range(nvars):
if eg[i + 1]:
bv.SetBit(i)
newEg[1] = bv
bitEx.append(newEg)
attrs = range(nvars)
mdl2 = CrossValidate.makeNBClassificationModel(bitEx, attrs, npvals, qBounds, useSigs=True)
nWrong = 0
for eg in bitEx:
p = mdl2.ClassifyExample(eg)
if p != eg[-1]:
nWrong += 1
self.assertEqual(nWrong, 1)
# now compare:
for i in range(len(bitEx)):
eg = examples[i]
p1 = mdl.ClassifyExample(eg)
bitEg = bitEx[i]
p2 = mdl2.ClassifyExample(bitEg)
self.assertEqual(p1, p2)
v1 = mdl.GetClassificationDetails()
v2 = mdl.GetClassificationDetails()
self.assertAlmostEqual(p1, p2, 4)
def _test5(self) : # disabled because CMIM was removed
examples = [
['a', 1,0,1,0,1,1,0, 1],
['b', 1,0,0,0,1,0,0, 1],
['c', 1,0,1,0,1,1,0, 0],
['d', 0,1,1,0,1,0,0, 0],
['e', 0,1,1,1,0,1,0, 0],
]
nvars = len(examples[0])-2
attrs = range(1,nvars+1)
npvals = [0] + [2]*nvars + [2]
qBounds = [0] + [0]*nvars + [0]
bitEx = []
for eg in examples:
newEg = [eg[0],None,eg[-1]]
bv = ExplicitBitVect(nvars)
for i in range(nvars):
if eg[i+1]: bv.SetBit(i)
# this bit will yield perfect accuracy if
# the attrs argument isn't being used properly:
newEg[1] = bv
bitEx.append(newEg)
attrs = range(nvars)
mdl2 = CrossValidate.makeNBClassificationModel(bitEx,attrs,npvals,qBounds,
useSigs=True,useCMIM=2)
nWrong = 0
for eg in bitEx:
p = mdl2.ClassifyExample(eg)
if p != eg[-1]: nWrong +=1
self.assertEqual(nWrong,1)
def test4(self):
examples = [
['a', 1, 0, 1, 0, 1],
['b', 1, 0, 0, 0, 1],
['c', 1, 0, 1, 0, 0],
['d', 0, 1, 1, 0, 0],
['e', 0, 1, 1, 1, 0],
]
nvars = len(examples[0]) - 2
attrs = range(1, nvars + 1)
origNVars = nvars
nvars = 10
npvals = [0] + [2] * nvars + [2]
qBounds = [0] + [0] * nvars + [0]
bitEx = []
for eg in examples:
newEg = [eg[0], None, eg[-1]]
bv = ExplicitBitVect(nvars)
for i in range(origNVars):
if eg[i + 1]:
bv.SetBit(i)
# this bit will yield perfect accuracy if
# the attrs argument isn't being used properly:
if eg[-1]:
bv.SetBit(origNVars)
newEg[1] = bv
bitEx.append(newEg)
attrs = range(origNVars)
mdl2 = CrossValidate.makeNBClassificationModel(bitEx,
attrs,
npvals,
qBounds,
useSigs=True)
nWrong = 0
for eg in bitEx:
p = mdl2.ClassifyExample(eg)
if p != eg[-1]:
nWrong += 1
self.assertEqual(nWrong, 1)
def _test5(self): # disabled because CMIM was removed # pragma: nocover
examples = [
['a', 1, 0, 1, 0, 1, 1, 0, 1],
['b', 1, 0, 0, 0, 1, 0, 0, 1],
['c', 1, 0, 1, 0, 1, 1, 0, 0],
['d', 0, 1, 1, 0, 1, 0, 0, 0],
['e', 0, 1, 1, 1, 0, 1, 0, 0],
]
nvars = len(examples[0]) - 2
npvals = [0] + [2] * nvars + [2]
qBounds = [0] + [0] * nvars + [0]
bitEx = []
for eg in examples:
newEg = [eg[0], None, eg[-1]]
bv = ExplicitBitVect(nvars)
for i in range(nvars):
if eg[i + 1]:
bv.SetBit(i)
# this bit will yield perfect accuracy if
# the attrs argument isn't being used properly:
newEg[1] = bv
bitEx.append(newEg)
attrs = list(range(nvars))
mdl2 = CrossValidate.makeNBClassificationModel(bitEx,
attrs,
npvals,
qBounds,
useSigs=True,
useCMIM=2)
nWrong = 0
for eg in bitEx:
p = mdl2.ClassifyExample(eg)
if p != eg[-1]:
nWrong += 1
self.assertEqual(nWrong, 1)
def test4(self) :
examples = [
['a', 1,0,1,0, 1],
['b', 1,0,0,0, 1],
['c', 1,0,1,0, 0],
['d', 0,1,1,0, 0],
['e', 0,1,1,1, 0],
]
nvars = len(examples[0])-2
attrs = range(1,nvars+1)
origNVars=nvars
nvars = 10
npvals = [0] + [2]*nvars + [2]
qBounds = [0] + [0]*nvars + [0]
bitEx = []
for eg in examples:
newEg = [eg[0],None,eg[-1]]
bv = ExplicitBitVect(nvars)
for i in range(origNVars):
if eg[i+1]: bv.SetBit(i)
# this bit will yield perfect accuracy if
# the attrs argument isn't being used properly:
if eg[-1]: bv.SetBit(origNVars)
newEg[1] = bv
bitEx.append(newEg)
attrs = range(origNVars)
mdl2 = CrossValidate.makeNBClassificationModel(bitEx,attrs,npvals,qBounds,
useSigs=True)
nWrong = 0
for eg in bitEx:
p = mdl2.ClassifyExample(eg)
if p != eg[-1]: nWrong +=1
self.failUnless(nWrong==1)
