def header(self,cells):
self = self or Rows()
self.indep = []
for c0,x in enumerate(cells):
if not "?" in x:
c = len(self._use)
self._use.append(c0)
self.name.append(x)
if "$" in x or "<" in x or ">" in x:
n1 = Num()
self.nums[c] = n1.nums([])
else:
s1 = Sym()
self.syms[c] = s1.syms([])
if "<" in x:
self.w[c] = -1
elif ">" in x:
self.w[c] = 1
elif "!" in x:
self._class = c
else:
self.indeps.append(c)
return self
def testing_Sym():
s1 = Sym()
s1 = s1.syms(
['y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'n', 'n', 'n', 'n', 'n'])
print(s1.symEnt())
result = round(s1.symEnt(), 4)
assert (result == 0.9403)
def baseSym():
s = Sym()
s = s.syms(
['y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'n', 'n', 'n', 'n', 'n'])
print(s.counts)
print("Entropy: " + str(round(s.symEnt(), 4)))
assert (round(s.symEnt(), 4) == 0.9403)
def __init__(i, a):
i.x, i.y, i.all = [], [], []
i.klass = None
for at, txt in enumerate(a):
new = Skip(at=at, txt=txt) if isSkip(txt) else (
Num(at=at, txt=txt) if isNum(txt) else Sym(at=at, txt=txt))
i.all += [new]
if not isSkip(txt):
if isX(txt): i.x += [new]
if isY(txt): i.y += [new]
if isKlass(txt): i.klass = new
def testSample():
"""Testing Entropy Sampler"""
s = Sym()
syms = ['y','y','y','y','y','y','y','y','y',
'n','n','n','n','n']
s.bulkAdd(syms)
print("Items = ", syms)
print("Entropy = ", '%.4f'%(s.symEnt()))
assert math.isclose(s.symEnt(), 0.9403, rel_tol = 0.01)
def header(self, cells):
for c0, x in enumerate(cells):
if not "?" in x:
c = len(self._use)
self._use.append(c0)
self.name.append(x)
if re.search("[<>$]", x):
self.nums[c] = Num()
else:
self.syms[c] = Sym()
if re.search("<", x): self.w[c] = -1
elif re.search(">", x): self.w[c] = 1
elif re.search("!", x): self._class = c
else: self.indeps.append(c)
return self
def header(self, cells):
for i, v in enumerate(cells):
if not re.match(r'^\?', v):
c = len(self._use)
self._use.append(i)
self.name.append(v)
if re.search('[<>$]', v):
self.nums[c] = Num([])
else:
self.syms[c] = Sym([])
if re.search('<', v):
self.w[c] = -1
elif re.search('>', v):
self.w[c] = 1
elif re.search('!', v):
self._class = c
else:
self.indeps.append(c)
def header(self, cells):
for c0, x in enumerate(cells):
if '?' not in x:
c = len(self._use)
self._use.append(c0)
self.name.append(x)
if re.search('[<>$]', x):
self.nums[c] = Num([])
else:
self.syms[c] = Sym([])
if re.match('<', x):
self.w[c] = -1
elif re.match(">", x):
self.w[c] = 1
elif re.match('!', x):
self.Class = c
else:
self.indeps.append(c)
return self
def header(data, cells):
for i, x in enumerate(cells):
if "%?" not in x:
data._use[i] = True
data.name.append(x)
if re.search(r"[<>$]", x):
data.nums[i] = Num()
else:
data.syms[i] = Sym()
if re.search(r"<", x):
data.w[i] = -1
elif re.search(r">", x):
data.w[i] = 1
elif re.search(r"!", x):
data.clss = i
else:
data.indeps[i] = True
else:
data._use[i] = False
def header(self, cells):
for index, cell in enumerate(cells):
if '?' not in cell:
self.valid_cols.append(index)
self.names[index] = cell
if re.match('[<>$]', cell):
self.nums[index] = Num()
else:
self.syms[index] = Sym()
if re.match('<', cell):
self.weights[index] = -1
elif re.match('>', cell):
self.weights[index] = 1
elif re.match('!', cell):
self.class_col = index
else:
self.indeps.append(index)
def header(self, cells):
'''Checks for certain symbols at the beginning of the column name and structure then into sym and num objects'''
self = self or Rows()
self.indep = []
for c0, x in enumerate(cells):
if not "?" in x:
c = len(self._use)
self._use.append(c0)
self.name.append(x)
# Col names beginning with $,<,> are set as numeric columns
if "$" in x or "<" in x or ">" in x:
self.nums[c] = Num([])
else:
self.syms[c] = Sym([])
if "<" in x:
self.w[c] = -1
elif ">" in x:
self.w[c] = 1
elif "!" in x:
self._class = c
else:
self.indeps.append(c)
return self
def sym_test():
"""
Testing sym.py ...
"""
sym = Sym()
sym_list = [
'y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'n', 'n', 'n', 'n', 'n'
]
sym.syms(sym_list)
print()
print('ENT', ':', sym.sym_ent())
print()
assert (close(sym.sym_ent(), 0.9403))
def __init__(i, down=-math.inf, up=math.inf):
i.down, i.up, i.also = down, up, Sym()
def testSym():
s = Sym(
['y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'n', 'n', 'n', 'n', 'n'])
assert abs(s.symEnt() - 0.9403) / 0.9403 < 0.01
print(s.symEnt())
def sym_test():
s = Sym(
['y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'n', 'n', 'n', 'n', 'n'])
print(s.sym_ent())
assert abs(s.sym_ent() - 0.9403) < 0.0001
def test_sym():
s = Sym(['y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'y',
'n', 'n', 'n', 'n', 'n'])
print(s.symEnt())
assert round(s.symEnt(), 4) == 0.9403
# vim: filetype=python ts=2 sw=2 sts=2 et : from sym import Sym s=Sym(all="aaaabbc") assert 4==s.seen["a"] assert 1.378 <= s.spread() <=1.38
def __add__(i, x):
for y in items(x): # x could a single thing or list of items
if y != my.data.ignore:
if not i.has:
i.has = Num() if nump(y) else Sym()
i.has + y
def testSym():
sy = Sym(
['y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'y', 'n', 'n', 'n', 'n', 'n'])
assert round(sy.symEnt(), 4) == 0.9403