def get_large_sets(self):
"""
Calculates large sets from transactions. Returns dictionary where the key is arity and the value is list of item sets.
"""
large_sets = {}
counter = self.__count_items_in_transactions(self.__transactions)
minsup_count = self.__get_minsup_count()
L1 = self.__getL1(counter, minsup_count)
self.root = Root(self.__minsup, len(self.__transactions), L1,
self.__partial)
self.root.update_states(0)
parts_no = math.ceil(float(len(self.__transactions)) / float(self.__M))
finished = False
pass_counter = 0
if self.__randomize:
transaction_order = random.sample(xrange(len(self.__transactions)),
len(self.__transactions))
else:
transaction_order = xrange(len(self.__transactions))
if self.__partial:
M = self.__M
else:
M = int(math.ceil(len(self.__transactions) * self.__minsup))
M = self.__M
while not finished:
trans_count = 0
for i in transaction_order:
transaction = self.__transactions.get(i)
self.root.increment(transaction)
trans_count += 1
if (i + 1) % M == 0 or (i + 1) == len(self.__transactions):
finished = self.root.update_states(trans_count)
trans_count = 0
if finished:
break
M = self.__M
pass_counter += 1
#self.root.print_node()
#print "passes: %d" % (pass_counter,)
large_sets = self.root.get_large_sets(large_sets)
return large_sets
def get_large_sets(self):
"""
Calculates large sets from transactions. Returns dictionary where the key is arity and the value is list of item sets.
"""
large_sets = {}
counter = self.__count_items_in_transactions(self.__transactions)
minsup_count = self.__get_minsup_count()
L1 = self.__getL1(counter, minsup_count)
self.root = Root(self.__minsup, len(self.__transactions), L1, self.__partial)
self.root.update_states(0)
parts_no = math.ceil(float(len(self.__transactions)) / float(self.__M))
finished = False
pass_counter = 0
if self.__randomize:
transaction_order = random.sample(xrange(len(self.__transactions)), len(self.__transactions))
else:
transaction_order = xrange(len(self.__transactions))
if self.__partial:
M = self.__M
else:
M = int(math.ceil(len(self.__transactions) * self.__minsup))
M = self.__M
while not finished:
trans_count = 0
for i in transaction_order:
transaction = self.__transactions.get(i)
self.root.increment(transaction)
trans_count += 1
if (i+1) % M == 0 or (i+1) == len(self.__transactions):
finished = self.root.update_states(trans_count)
trans_count = 0
if finished:
break
M = self.__M
pass_counter += 1
#self.root.print_node()
#print "passes: %d" % (pass_counter,)
large_sets = self.root.get_large_sets(large_sets)
return large_sets
class Dic(object):
def __init__(self, transactions, minsup, M, randomize, partial):
self.__minsup = minsup
self.__transactions = transactions
if M == 0:
self.__M = int(math.ceil(len(transactions) * minsup))
elif M < 0:
self.__M = -M * int(math.ceil(len(transactions) * minsup))
elif M > 0:
self.__M = M
self.__minsup_count = self.__get_minsup_count()
self.__randomize = randomize
self.__partial = partial
def get_large_sets(self):
"""
Calculates large sets from transactions. Returns dictionary where the key is arity and the value is list of item sets.
"""
large_sets = {}
counter = self.__count_items_in_transactions(self.__transactions)
minsup_count = self.__get_minsup_count()
L1 = self.__getL1(counter, minsup_count)
self.root = Root(self.__minsup, len(self.__transactions), L1, self.__partial)
self.root.update_states(0)
parts_no = math.ceil(float(len(self.__transactions)) / float(self.__M))
finished = False
pass_counter = 0
if self.__randomize:
transaction_order = random.sample(xrange(len(self.__transactions)), len(self.__transactions))
else:
transaction_order = xrange(len(self.__transactions))
if self.__partial:
M = self.__M
else:
M = int(math.ceil(len(self.__transactions) * self.__minsup))
M = self.__M
while not finished:
trans_count = 0
for i in transaction_order:
transaction = self.__transactions.get(i)
self.root.increment(transaction)
trans_count += 1
if (i+1) % M == 0 or (i+1) == len(self.__transactions):
finished = self.root.update_states(trans_count)
trans_count = 0
if finished:
break
M = self.__M
pass_counter += 1
#self.root.print_node()
#print "passes: %d" % (pass_counter,)
large_sets = self.root.get_large_sets(large_sets)
return large_sets
def get_counter(self):
return DicCounter(self.root)
def get_large_sets_and_counter(self):
return self.get_large_sets(), self.get_counter()
@staticmethod
def __count_items_in_transactions(transactions):
"""
Counts items and returns dictionary with them as keys and their count as value.
"""
counter = defaultdict(int)
for transaction in transactions:
for item in transaction:
counter[(item,)] += 1
return counter
def __get_minsup_count(self):
"""
Calculates and returns miniumum support given in number of transactions.
"""
return int(len(self.__transactions) * self.__minsup)
@staticmethod
def __getL1(counter, minsup_count):
"""
Calculates and returns first large set.
"""
L1 = {}
for k, v in counter.iteritems():
if v >= minsup_count:
L1[k] = v
return L1
class Dic(object):
def __init__(self, transactions, minsup, M, randomize, partial):
self.__minsup = minsup
self.__transactions = transactions
if M == 0:
self.__M = int(math.ceil(len(transactions) * minsup))
elif M < 0:
self.__M = -M * int(math.ceil(len(transactions) * minsup))
elif M > 0:
self.__M = M
self.__minsup_count = self.__get_minsup_count()
self.__randomize = randomize
self.__partial = partial
def get_large_sets(self):
"""
Calculates large sets from transactions. Returns dictionary where the key is arity and the value is list of item sets.
"""
large_sets = {}
counter = self.__count_items_in_transactions(self.__transactions)
minsup_count = self.__get_minsup_count()
L1 = self.__getL1(counter, minsup_count)
self.root = Root(self.__minsup, len(self.__transactions), L1,
self.__partial)
self.root.update_states(0)
parts_no = math.ceil(float(len(self.__transactions)) / float(self.__M))
finished = False
pass_counter = 0
if self.__randomize:
transaction_order = random.sample(xrange(len(self.__transactions)),
len(self.__transactions))
else:
transaction_order = xrange(len(self.__transactions))
if self.__partial:
M = self.__M
else:
M = int(math.ceil(len(self.__transactions) * self.__minsup))
M = self.__M
while not finished:
trans_count = 0
for i in transaction_order:
transaction = self.__transactions.get(i)
self.root.increment(transaction)
trans_count += 1
if (i + 1) % M == 0 or (i + 1) == len(self.__transactions):
finished = self.root.update_states(trans_count)
trans_count = 0
if finished:
break
M = self.__M
pass_counter += 1
#self.root.print_node()
#print "passes: %d" % (pass_counter,)
large_sets = self.root.get_large_sets(large_sets)
return large_sets
def get_counter(self):
return DicCounter(self.root)
def get_large_sets_and_counter(self):
return self.get_large_sets(), self.get_counter()
@staticmethod
def __count_items_in_transactions(transactions):
"""
Counts items and returns dictionary with them as keys and their count as value.
"""
counter = defaultdict(int)
for transaction in transactions:
for item in transaction:
counter[(item, )] += 1
return counter
def __get_minsup_count(self):
"""
Calculates and returns miniumum support given in number of transactions.
"""
return int(len(self.__transactions) * self.__minsup)
@staticmethod
def __getL1(counter, minsup_count):
"""
Calculates and returns first large set.
"""
L1 = {}
for k, v in counter.iteritems():
if v >= minsup_count:
L1[k] = v
return L1
