sam/pyFTS
1
0
Fork 0
Code Issues 3 Pull Requests Actions Packages Projects Releases Wiki Activity

Notebooks and method bugfixes and refactoring

Browse Source
This commit is contained in:
Petrônio Cândido 2018-02-27 18:30:20 -03:00
parent 37862c661d
commit 7a2dd7e54c
27 changed files with 4725 additions and 505 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
pyFTS/benchmarks/benchmarks.py
View File

@ -90,7 +90,7 @@ def run_point(mfts, partitioner, train_data, test_data, window_key=None, transfo
mfts.append_transformation(transformation) mfts.append_transformation(transformation)
_start = time.time() _start = time.time()
mfts.train(train_data, partitioner.sets, order=mfts.order) mfts.train(train_data, sets=partitioner.sets, order=mfts.order)
_end = time.time() _end = time.time()
times = _end - _start times = _end - _start
@ -273,7 +273,7 @@ def all_point_forecasters(data_train, data_test, partitions, max_order=3, statis
#print(model) #print(model)
if transformation is not None: if transformation is not None:
model.append_transformation(transformation) model.append_transformation(transformation)
model.train(data_train, data_train_fs.sets, order=model.order) model.train(data_train, sets=data_train_fs.sets, order=model.order)
objs.append(model) objs.append(model)
lcolors.append( colors[count % ncol] ) lcolors.append( colors[count % ncol] )
@ -385,7 +385,7 @@ def interval_sliding_window(data, windowsize, train=0.8, models=None, partitione
mfts.append_transformation(transformation) mfts.append_transformation(transformation)
_start = time.time() _start = time.time()
mfts.train(training, data_train_fs.sets) mfts.train(training, sets=data_train_fs.sets)
_end = time.time() _end = time.time()
_tdiff = _end - _start _tdiff = _end - _start
@ -419,7 +419,7 @@ def interval_sliding_window(data, windowsize, train=0.8, models=None, partitione
mfts.append_transformation(transformation) mfts.append_transformation(transformation)
_start = time.time() _start = time.time()
mfts.train(training, data_train_fs.sets, order=order) mfts.train(training, sets=data_train_fs.sets, order=order)
_end = time.time() _end = time.time()
_tdiff = _end - _start _tdiff = _end - _start
@ -476,7 +476,7 @@ def all_interval_forecasters(data_train, data_test, partitions, max_order=3,save
for count, model in Util.enumerate2(models, start=0, step=2): for count, model in Util.enumerate2(models, start=0, step=2):
if transformation is not None: if transformation is not None:
model.append_transformation(transformation) model.append_transformation(transformation)
model.train(data_train, data_train_fs, order=model.order) model.train(data_train, sets=data_train_fs, order=model.order)
objs.append(model) objs.append(model)
lcolors.append( colors[count % ncol] ) lcolors.append( colors[count % ncol] )
@ -635,7 +635,7 @@ def ahead_sliding_window(data, windowsize, train, steps, models=None, resolution
mfts.append_transformation(transformation) mfts.append_transformation(transformation)
_start = time.time() _start = time.time()
mfts.train(train, data_train_fs.sets) mfts.train(train, sets=data_train_fs.sets)
_end = time.time() _end = time.time()
_tdiff = _end - _start _tdiff = _end - _start
@ -670,7 +670,7 @@ def ahead_sliding_window(data, windowsize, train, steps, models=None, resolution
mfts.append_transformation(transformation) mfts.append_transformation(transformation)
_start = time.time() _start = time.time()
mfts.train(train, data_train_fs.sets, order=order) mfts.train(train, sets=data_train_fs.sets, order=order)
_end = time.time() _end = time.time()
_tdiff = _end - _start _tdiff = _end - _start
@ -705,7 +705,7 @@ def all_ahead_forecasters(data_train, data_test, partitions, start, steps, resol
if not mfts.is_high_order: if not mfts.is_high_order:
if transformation is not None: if transformation is not None:
mfts.append_transformation(transformation) mfts.append_transformation(transformation)
mfts.train(data_train, data_train_fs) mfts.train(data_train, sets=data_train_fs.sets)
objs.append(mfts) objs.append(mfts)
lcolors.append( colors[count % ncol] ) lcolors.append( colors[count % ncol] )
else: else:
@ -714,7 +714,7 @@ def all_ahead_forecasters(data_train, data_test, partitions, start, steps, resol
mfts = model(" n = " + str(order)) mfts = model(" n = " + str(order))
if transformation is not None: if transformation is not None:
mfts.append_transformation(transformation) mfts.append_transformation(transformation)
mfts.train(data_train, data_train_fs, order=order) mfts.train(data_train, sets=data_train_fs.sets, order=order)
objs.append(mfts) objs.append(mfts)
lcolors.append(colors[count % ncol]) lcolors.append(colors[count % ncol])
@ -896,7 +896,7 @@ def SelecaoSimples_MenorRMSE(original, parameters, modelo):
for p in parameters: for p in parameters:
sets = Grid.GridPartitioner(original, p).sets sets = Grid.GridPartitioner(original, p).sets
fts = modelo(str(p) + " particoes") fts = modelo(str(p) + " particoes")
fts.train(original, sets) fts.train(original, sets=sets)
# print(original) # print(original)
forecasted = fts.forecast(original) forecasted = fts.forecast(original)
forecasted.insert(0, original[0]) forecasted.insert(0, original[0])
@ -936,7 +936,7 @@ def SelecaoSimples_MenorRMSE(original, parameters, modelo):
for p in parameters: for p in parameters:
sets = Grid.GridPartitionerTrimf(difffts, p) sets = Grid.GridPartitionerTrimf(difffts, p)
fts = modelo(str(p) + " particoes") fts = modelo(str(p) + " particoes")
fts.train(difffts, sets) fts.train(difffts, sets=sets)
forecasted = fts.forecast(difffts) forecasted = fts.forecast(difffts)
forecasted.insert(0, difffts[0]) forecasted.insert(0, difffts[0])
ax2.plot(forecasted, label=fts.name) ax2.plot(forecasted, label=fts.name)
@ -1050,7 +1050,7 @@ def simpleSearch_RMSE(train, test, model, partitions, orders, save=False, file=N
for oc, o in enumerate(orders, start=0): for oc, o in enumerate(orders, start=0):
fts = model("q = " + str(p) + " n = " + str(o)) fts = model("q = " + str(p) + " n = " + str(o))
fts.append_transformation(transformation) fts.append_transformation(transformation)
fts.train(train, sets, o, parameters=parameters) fts.train(train, sets=sets, order=o, parameters=parameters)
if not intervals: if not intervals:
forecasted = fts.forecast(test) forecasted = fts.forecast(test)
if not fts.has_seasonality: if not fts.has_seasonality:
@ -1128,7 +1128,7 @@ def sliding_window_simple_search(data, windowsize, model, partitions, orders, sa
_error = [] _error = []
for ct, train, test in Util.sliding_window(data, windowsize, 0.8): for ct, train, test in Util.sliding_window(data, windowsize, 0.8):
fts = model("q = " + str(p) + " n = " + str(o)) fts = model("q = " + str(p) + " n = " + str(o))
fts.train(data, sets, o, parameters=parameters) fts.train(data, sets=sets, order=o, parameters=parameters)
if not intervals: if not intervals:
forecasted = fts.forecast(test) forecasted = fts.forecast(test)
if not fts.has_seasonality: if not fts.has_seasonality:
@ -1191,7 +1191,7 @@ def pftsExploreOrderAndPartitions(data,save=False, file=None):
for order in np.arange(1, 6): for order in np.arange(1, 6):
fts = pwfts.ProbabilisticWeightedFTS("") fts = pwfts.ProbabilisticWeightedFTS("")
fts.shortname = "n = " + str(order) fts.shortname = "n = " + str(order)
fts.train(data, data_fs1, order=order) fts.train(data, sets=data_fs1.sets, order=order)
point_forecasts = fts.forecast(data) point_forecasts = fts.forecast(data)
interval_forecasts = fts.forecast_interval(data) interval_forecasts = fts.forecast_interval(data)
lower = [kk[0] for kk in interval_forecasts] lower = [kk[0] for kk in interval_forecasts]
@ -1213,7 +1213,7 @@ def pftsExploreOrderAndPartitions(data,save=False, file=None):
data_fs = Grid.GridPartitioner(data, partitions).sets data_fs = Grid.GridPartitioner(data, partitions).sets
fts = pwfts.ProbabilisticWeightedFTS("") fts = pwfts.ProbabilisticWeightedFTS("")
fts.shortname = "q = " + str(partitions) fts.shortname = "q = " + str(partitions)
fts.train(data, data_fs, 1) fts.train(data, sets=data_fs.sets, order=1)
point_forecasts = fts.forecast(data) point_forecasts = fts.forecast(data)
interval_forecasts = fts.forecast_interval(data) interval_forecasts = fts.forecast_interval(data)
lower = [kk[0] for kk in interval_forecasts] lower = [kk[0] for kk in interval_forecasts]

4
pyFTS/common/fts.py
View File

@ -148,7 +148,7 @@ class FTS(object):
""" """
raise NotImplementedError('This model do not perform multi step ahead distribution forecasts!') raise NotImplementedError('This model do not perform multi step ahead distribution forecasts!')
def train(self, data, sets, order=1, parameters=None): def train(self, data, **kwargs):
""" """
:param data: :param data:
@ -166,7 +166,7 @@ class FTS(object):
:param kwargs: :param kwargs:
:return: :return:
""" """
self.train(data, sets=None) self.train(data, **kwargs)
def append_transformation(self, transformation): def append_transformation(self, transformation):
if transformation is not None: if transformation is not None:

20
pyFTS/data/artificial.py
View File

@ -28,6 +28,7 @@ def generate_gaussian_linear(mu_ini, sigma_ini, mu_inc, sigma_inc, it=100, num=1
sigma += sigma_inc sigma += sigma_inc
return ret return ret
def generate_uniform_linear(min_ini, max_ini, min_inc, max_inc, it=100, num=10, vmin=None, vmax=None): def generate_uniform_linear(min_ini, max_ini, min_inc, max_inc, it=100, num=10, vmin=None, vmax=None):
""" """
Generate data sampled from Uniform distribution, with constant or linear changing bounds Generate data sampled from Uniform distribution, with constant or linear changing bounds
@ -53,4 +54,21 @@ def generate_uniform_linear(min_ini, max_ini, min_inc, max_inc, it=100, num=10,
ret.extend(tmp) ret.extend(tmp)
_min += min_inc _min += min_inc
_max += max_inc _max += max_inc
return ret return ret
def white_noise(n=500):
return np.random.normal(0, 1, n)
def random_walk(n=500, type='gaussian'):
if type == 'gaussian':
tmp = generate_gaussian_linear(0, 1, 0, 0, it=1, num=n)
else:
tmp = generate_uniform_linear(-1, 1, 0, 0, it=1, num=n)
ret = [0]
for i in range(n):
ret.append(tmp[i] + ret[i])
return ret

13
pyFTS/models/chen.py
View File

@ -36,7 +36,7 @@ class ConventionalFTS(fts.FTS):
self.detail = "Chen" self.detail = "Chen"
self.flrgs = {} self.flrgs = {}
def generateFLRG(self, flrs): def generate_flrg(self, flrs):
flrgs = {} flrgs = {}
for flr in flrs: for flr in flrs:
if flr.LHS.name in flrgs: if flr.LHS.name in flrgs:
@ -46,12 +46,13 @@ class ConventionalFTS(fts.FTS):
flrgs[flr.LHS.name].append(flr.RHS) flrgs[flr.LHS.name].append(flr.RHS)
return (flrgs) return (flrgs)
def train(self, data, sets,order=1,parameters=None): def train(self, data, **kwargs):
self.sets = sets if kwargs.get('sets', None) is not None:
self.sets = kwargs.get('sets', None)
ndata = self.apply_transformations(data) ndata = self.apply_transformations(data)
tmpdata = FuzzySet.fuzzyfy_series_old(ndata, sets) tmpdata = FuzzySet.fuzzyfy_series_old(ndata, self.sets)
flrs = FLR.generate_non_recurrent_flrs(tmpdata) flrs = FLR.generate_non_recurrent_flrs(tmpdata)
self.flrgs = self.generateFLRG(flrs) self.flrgs = self.generate_flrg(flrs)
def forecast(self, data, **kwargs): def forecast(self, data, **kwargs):
@ -74,6 +75,6 @@ class ConventionalFTS(fts.FTS):
ret.append(_flrg.get_midpoint()) ret.append(_flrg.get_midpoint())
ret = self.apply_inverse_transformations(ret, params=[data[self.order - 1:]]) ret = self.apply_inverse_transformations(ret, params=[data])
return ret return ret

39
pyFTS/models/cheng.py
View File

@ -16,28 +16,31 @@ class TrendWeightedFLRG(yu.WeightedFLRG):
""" """
def __init__(self, LHS, **kwargs): def __init__(self, LHS, **kwargs):
super(TrendWeightedFLRG, self).__init__(LHS, **kwargs) super(TrendWeightedFLRG, self).__init__(LHS, **kwargs)
self.w = None
def weights(self): def weights(self):
count_nochange = 0.0 if self.w is None:
count_up = 0.0 count_nochange = 0.0
count_down = 0.0 count_up = 0.0
weights = [] count_down = 0.0
weights = []
for c in self.RHS: for c in self.RHS:
tmp = 0 tmp = 0
if self.LHS.centroid == c.centroid: if self.LHS.centroid == c.centroid:
count_nochange += 1.0 count_nochange += 1.0
tmp = count_nochange tmp = count_nochange
elif self.LHS.centroid > c.centroid: elif self.LHS.centroid > c.centroid:
count_down += 1.0 count_down += 1.0
tmp = count_down tmp = count_down
else: else:
count_up += 1.0 count_up += 1.0
tmp = count_up tmp = count_up
weights.append(tmp) weights.append(tmp)
tot = sum(weights) tot = sum(weights)
return np.array([k / tot for k in weights]) self.w = np.array([k / tot for k in weights])
return self.w
class TrendWeightedFTS(yu.WeightedFTS): class TrendWeightedFTS(yu.WeightedFTS):

46
pyFTS/models/hofts.py
View File

@ -16,19 +16,19 @@ class HighOrderFLRG(flrg.FLRG):
self.RHS = {} self.RHS = {}
self.strlhs = "" self.strlhs = ""
def appendRHS(self, c): def append_rhs(self, c):
if c.name not in self.RHS: if c.name not in self.RHS:
self.RHS[c.name] = c self.RHS[c.name] = c
def strLHS(self): def str_lhs(self):
if len(self.strlhs) == 0: if len(self.strlhs) == 0:
for c in self.LHS: for c in self.LHS:
if len(self.strlhs) > 0: if len(self.strlhs) > 0:
self.strlhs += ", " self.strlhs += ", "
self.strlhs = self.strlhs + str(c) self.strlhs = self.strlhs + str(c.name)
return self.strlhs return self.strlhs
def appendLHS(self, c): def append_lhs(self, c):
self.LHS.append(c) self.LHS.append(c)
def __str__(self): def __str__(self):
@ -37,7 +37,7 @@ class HighOrderFLRG(flrg.FLRG):
if len(tmp) > 0: if len(tmp) > 0:
tmp = tmp + "," tmp = tmp + ","
tmp = tmp + c tmp = tmp + c
return self.strLHS() + " -> " + tmp return self.str_lhs() + " -> " + tmp
def __len__(self): def __len__(self):
@ -51,7 +51,7 @@ class HighOrderFTS(fts.FTS):
self.name = "High Order FTS" self.name = "High Order FTS"
self.shortname = "HOFTS" + name self.shortname = "HOFTS" + name
self.detail = "Chen" self.detail = "Chen"
self.order = 1 self.order = kwargs.get('order',1)
self.setsDict = {} self.setsDict = {}
self.is_high_order = True self.is_high_order = True
@ -83,13 +83,13 @@ class HighOrderFTS(fts.FTS):
flrg = HighOrderFLRG(self.order) flrg = HighOrderFLRG(self.order)
for kk in np.arange(k - self.order, k): for kk in np.arange(k - self.order, k):
flrg.appendLHS(flrs[kk].LHS) flrg.append_lhs(flrs[kk].LHS)
if flrg.strLHS() in flrgs: if flrg.str_lhs() in flrgs:
flrgs[flrg.strLHS()].appendRHS(flrs[k].RHS) flrgs[flrg.str_lhs()].append_rhs(flrs[k].RHS)
else: else:
flrgs[flrg.strLHS()] = flrg; flrgs[flrg.str_lhs()] = flrg;
flrgs[flrg.strLHS()].appendRHS(flrs[k].RHS) flrgs[flrg.str_lhs()].append_rhs(flrs[k].RHS)
return (flrgs) return (flrgs)
def generate_flrg(self, data): def generate_flrg(self, data):
@ -118,23 +118,25 @@ class HighOrderFTS(fts.FTS):
flrg = HighOrderFLRG(self.order) flrg = HighOrderFLRG(self.order)
path = list(reversed(list(filter(None.__ne__, p)))) path = list(reversed(list(filter(None.__ne__, p))))
for lhs in enumerate(path, start=0): for lhs in path:
flrg.appendLHS(lhs) flrg.append_lhs(lhs)
if flrg.strLHS() not in flrgs: if flrg.str_lhs() not in flrgs:
flrgs[flrg.strLHS()] = flrg; flrgs[flrg.str_lhs()] = flrg;
for st in rhs: for st in rhs:
flrgs[flrg.strLHS()].appendRHS(st) flrgs[flrg.str_lhs()].append_rhs(st)
return flrgs return flrgs
def train(self, data, sets, order=1,parameters=None): def train(self, data, **kwargs):
data = self.apply_transformations(data, updateUoD=True) data = self.apply_transformations(data, updateUoD=True)
self.order = order self.order = kwargs.get('order',2)
self.sets = sets
if kwargs.get('sets', None) is not None:
self.sets = kwargs.get('sets', None)
for s in self.sets: self.setsDict[s.name] = s for s in self.sets: self.setsDict[s.name] = s
self.flrgs = self.generate_flrg(data) self.flrgs = self.generate_flrg(data)
@ -153,12 +155,12 @@ class HighOrderFTS(fts.FTS):
tmpdata = FuzzySet.fuzzyfy_series_old(ndata[k - self.order: k], self.sets) tmpdata = FuzzySet.fuzzyfy_series_old(ndata[k - self.order: k], self.sets)
tmpflrg = HighOrderFLRG(self.order) tmpflrg = HighOrderFLRG(self.order)
for s in tmpdata: tmpflrg.appendLHS(s) for s in tmpdata: tmpflrg.append_lhs(s)
if tmpflrg.strLHS() not in self.flrgs: if tmpflrg.str_lhs() not in self.flrgs:
ret.append(tmpdata[-1].centroid) ret.append(tmpdata[-1].centroid)
else: else:
flrg = self.flrgs[tmpflrg.strLHS()] flrg = self.flrgs[tmpflrg.str_lhs()]
ret.append(flrg.get_midpoint()) ret.append(flrg.get_midpoint())
ret = self.apply_inverse_transformations(ret, params=[data[self.order - 1:]]) ret = self.apply_inverse_transformations(ret, params=[data[self.order - 1:]])

7
pyFTS/models/hwang.py
View File

@ -50,6 +50,7 @@ class HighOrderFTS(fts.FTS):
return ret return ret
def train(self, data, sets, order=1, parameters=None): def train(self, data, **kwargs):
self.sets = sets if kwargs.get('sets', None) is not None:
self.order = order self.sets = kwargs.get('sets', None)
self.order = kwargs.get('order', 1)

12
pyFTS/models/ifts.py
View File

@ -24,16 +24,16 @@ class IntervalFTS(hofts.HighOrderFTS):
self.is_high_order = True self.is_high_order = True
def get_upper(self, flrg): def get_upper(self, flrg):
if flrg.strLHS() in self.flrgs: if flrg.str_lhs() in self.flrgs:
tmp = self.flrgs[flrg.strLHS()] tmp = self.flrgs[flrg.str_lhs()]
ret = tmp.get_upper() ret = tmp.get_upper()
else: else:
ret = flrg.LHS[-1].upper ret = flrg.LHS[-1].upper
return ret return ret
def get_lower(self, flrg): def get_lower(self, flrg):
if flrg.strLHS() in self.flrgs: if flrg.str_lhs() in self.flrgs:
tmp = self.flrgs[flrg.strLHS()] tmp = self.flrgs[flrg.str_lhs()]
ret = tmp.get_lower() ret = tmp.get_lower()
else: else:
ret = flrg.LHS[-1].lower ret = flrg.LHS[-1].lower
@ -93,7 +93,7 @@ class IntervalFTS(hofts.HighOrderFTS):
for p in root.paths(): for p in root.paths():
path = list(reversed(list(filter(None.__ne__, p)))) path = list(reversed(list(filter(None.__ne__, p))))
flrg = hofts.HighOrderFLRG(self.order) flrg = hofts.HighOrderFLRG(self.order)
for kk in path: flrg.appendLHS(self.sets[kk]) for kk in path: flrg.append_lhs(self.sets[kk])
affected_flrgs.append(flrg) affected_flrgs.append(flrg)
@ -115,7 +115,7 @@ class IntervalFTS(hofts.HighOrderFTS):
for kk in idx: for kk in idx:
flrg = hofts.HighOrderFLRG(self.order) flrg = hofts.HighOrderFLRG(self.order)
flrg.appendLHS(self.sets[kk]) flrg.append_lhs(self.sets[kk])
affected_flrgs.append(flrg) affected_flrgs.append(flrg)
affected_flrgs_memberships.append(mv[kk]) affected_flrgs_memberships.append(mv[kk])

16
pyFTS/models/ismailefendi.py
View File

@ -17,6 +17,7 @@ class ImprovedWeightedFLRG(flrg.FLRG):
self.RHS = {} self.RHS = {}
self.rhs_counts = {} self.rhs_counts = {}
self.count = 0.0 self.count = 0.0
self.w = None
def append(self, c): def append(self, c):
if c.name not in self.RHS: if c.name not in self.RHS:
@ -27,7 +28,9 @@ class ImprovedWeightedFLRG(flrg.FLRG):
self.count += 1.0 self.count += 1.0
def weights(self): def weights(self):
return np.array([self.rhs_counts[c] / self.count for c in self.RHS.keys()]) if self.w is None:
self.w = np.array([self.rhs_counts[c] / self.count for c in self.RHS.keys()])
return self.w
def __str__(self): def __str__(self):
tmp = self.LHS.name + " -> " tmp = self.LHS.name + " -> "
@ -50,7 +53,7 @@ class ImprovedWeightedFTS(fts.FTS):
self.detail = "Ismail & Efendi" self.detail = "Ismail & Efendi"
self.setsDict = {} self.setsDict = {}
def generateFLRG(self, flrs): def generate_flrg(self, flrs):
flrgs = {} flrgs = {}
for flr in flrs: for flr in flrs:
if flr.LHS.name in flrgs: if flr.LHS.name in flrgs:
@ -60,8 +63,9 @@ class ImprovedWeightedFTS(fts.FTS):
flrgs[flr.LHS.name].append(flr.RHS) flrgs[flr.LHS.name].append(flr.RHS)
return (flrgs) return (flrgs)
def train(self, data, sets, order=1, parameters=None): def train(self, data, **kwargs):
self.sets = sets if kwargs.get('sets', None) is not None:
self.sets = kwargs.get('sets', None)
for s in self.sets: self.setsDict[s.name] = s for s in self.sets: self.setsDict[s.name] = s
@ -69,7 +73,7 @@ class ImprovedWeightedFTS(fts.FTS):
tmpdata = FuzzySet.fuzzyfy_series_old(ndata, self.sets) tmpdata = FuzzySet.fuzzyfy_series_old(ndata, self.sets)
flrs = FLR.generate_recurrent_flrs(tmpdata) flrs = FLR.generate_recurrent_flrs(tmpdata)
self.flrgs = self.generateFLRG(flrs) self.flrgs = self.generate_flrg(flrs)
def forecast(self, data, **kwargs): def forecast(self, data, **kwargs):
l = 1 l = 1
@ -95,6 +99,6 @@ class ImprovedWeightedFTS(fts.FTS):
ret.append(mp.dot(flrg.weights())) ret.append(mp.dot(flrg.weights()))
ret = self.apply_inverse_transformations(ret, params=[data[self.order - 1:]]) ret = self.apply_inverse_transformations(ret, params=[data])
return ret return ret

28
pyFTS/models/nonstationary/honsfts.py
View File

@ -88,7 +88,7 @@ class HighOrderNonStationaryFTS(hofts.HighOrderFTS):
flrgs[flrg.strLHS()] = flrg; flrgs[flrg.strLHS()] = flrg;
for st in rhs: for st in rhs:
flrgs[flrg.strLHS()].appendRHS(st) flrgs[flrg.strLHS()].append_rhs(st)
# flrgs = sorted(flrgs, key=lambda flrg: flrg.get_midpoint(0, window_size=1)) # flrgs = sorted(flrgs, key=lambda flrg: flrg.get_midpoint(0, window_size=1))
@ -144,7 +144,7 @@ class HighOrderNonStationaryFTS(hofts.HighOrderFTS):
affected_flrgs.append(flrg) affected_flrgs.append(flrg)
# affected_flrgs_memberships.append(flrg.get_membership(sample, disp)) # affected_flrgs_memberships.append(flrg.get_membership(sample, disp))
# print(flrg.strLHS()) # print(flrg.str_lhs())
# the FLRG is here because of the bounds verification # the FLRG is here because of the bounds verification
mv = [] mv = []
@ -196,14 +196,14 @@ class HighOrderNonStationaryFTS(hofts.HighOrderFTS):
tmp.append(common.check_bounds(sample[-1], self.sets, tdisp)) tmp.append(common.check_bounds(sample[-1], self.sets, tdisp))
elif len(affected_flrgs) == 1: elif len(affected_flrgs) == 1:
flrg = affected_flrgs[0] flrg = affected_flrgs[0]
if flrg.strLHS() in self.flrgs: if flrg.str_lhs() in self.flrgs:
tmp.append(self.flrgs[flrg.strLHS()].get_midpoint(tdisp)) tmp.append(self.flrgs[flrg.str_lhs()].get_midpoint(tdisp))
else: else:
tmp.append(flrg.LHS[-1].get_midpoint(tdisp)) tmp.append(flrg.LHS[-1].get_midpoint(tdisp))
else: else:
for ct, aset in enumerate(affected_flrgs): for ct, aset in enumerate(affected_flrgs):
if aset.strLHS() in self.flrgs: if aset.str_lhs() in self.flrgs:
tmp.append(self.flrgs[aset.strLHS()].get_midpoint(tdisp) * tmp.append(self.flrgs[aset.str_lhs()].get_midpoint(tdisp) *
affected_flrgs_memberships[ct]) affected_flrgs_memberships[ct])
else: else:
tmp.append(aset.LHS[-1].get_midpoint(tdisp)* tmp.append(aset.LHS[-1].get_midpoint(tdisp)*
@ -250,19 +250,19 @@ class HighOrderNonStationaryFTS(hofts.HighOrderFTS):
upper.append(aset.get_upper(tdisp)) upper.append(aset.get_upper(tdisp))
elif len(affected_flrgs) == 1: elif len(affected_flrgs) == 1:
_flrg = affected_flrgs[0] _flrg = affected_flrgs[0]
if _flrg.strLHS() in self.flrgs: if _flrg.str_lhs() in self.flrgs:
lower.append(self.flrgs[_flrg.strLHS()].get_lower(tdisp)) lower.append(self.flrgs[_flrg.str_lhs()].get_lower(tdisp))
upper.append(self.flrgs[_flrg.strLHS()].get_upper(tdisp)) upper.append(self.flrgs[_flrg.str_lhs()].get_upper(tdisp))
else: else:
lower.append(_flrg.LHS[-1].get_lower(tdisp)) lower.append(_flrg.LHS[-1].get_lower(tdisp))
upper.append(_flrg.LHS[-1].get_upper(tdisp)) upper.append(_flrg.LHS[-1].get_upper(tdisp))
else: else:
for ct, aset in enumerate(affected_flrgs): for ct, aset in enumerate(affected_flrgs):
if aset.strLHS() in self.flrgs: if aset.str_lhs() in self.flrgs:
lower.append(self.flrgs[aset.strLHS()].get_lower(tdisp) * lower.append(self.flrgs[aset.str_lhs()].get_lower(tdisp) *
affected_flrgs_memberships[ct]) affected_flrgs_memberships[ct])
upper.append(self.flrgs[aset.strLHS()].get_upper(tdisp) * upper.append(self.flrgs[aset.str_lhs()].get_upper(tdisp) *
affected_flrgs_memberships[ct]) affected_flrgs_memberships[ct])
else: else:
lower.append(aset.LHS[-1].get_lower(tdisp) * lower.append(aset.LHS[-1].get_lower(tdisp) *
affected_flrgs_memberships[ct]) affected_flrgs_memberships[ct])

80
pyFTS/models/pwfts.py
View File

@ -19,7 +19,7 @@ class ProbabilisticWeightedFLRG(hofts.HighOrderFLRG):
self.frequency_count = 0.0 self.frequency_count = 0.0
self.Z = None self.Z = None
def appendRHS(self, c): def append_rhs(self, c):
self.frequency_count += 1.0 self.frequency_count += 1.0
if c.name in self.RHS: if c.name in self.RHS:
self.rhs_count[c.name] += 1.0 self.rhs_count[c.name] += 1.0
@ -91,7 +91,7 @@ class ProbabilisticWeightedFLRG(hofts.HighOrderFLRG):
if len(tmp2) > 0: if len(tmp2) > 0:
tmp2 = tmp2 + ", " tmp2 = tmp2 + ", "
tmp2 = tmp2 + "(" + str(round(self.rhs_count[c] / self.frequency_count, 3)) + ")" + c tmp2 = tmp2 + "(" + str(round(self.rhs_count[c] / self.frequency_count, 3)) + ")" + c
return self.strLHS() + " -> " + tmp2 return self.str_lhs() + " -> " + tmp2
class ProbabilisticWeightedFTS(ifts.IntervalFTS): class ProbabilisticWeightedFTS(ifts.IntervalFTS):
@ -111,20 +111,22 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
self.interval_method = kwargs.get('interval_method','extremum') self.interval_method = kwargs.get('interval_method','extremum')
self.alpha = kwargs.get('alpha', 0.05) self.alpha = kwargs.get('alpha', 0.05)
def train(self, data, sets, order=1,parameters='Fuzzy'): def train(self, data, **kwargs):
data = self.apply_transformations(data, updateUoD=True) data = self.apply_transformations(data, updateUoD=True)
self.order = order parameters = kwargs.get('parameters','Fuzzy')
if sets is None and self.partitioner is not None:
self.order = kwargs.get('order',1)
if kwargs.get('sets',None) is None and self.partitioner is not None:
self.sets = self.partitioner.sets self.sets = self.partitioner.sets
self.original_min = self.partitioner.min self.original_min = self.partitioner.min
self.original_max = self.partitioner.max self.original_max = self.partitioner.max
else: else:
self.sets = sets self.sets = kwargs.get('sets',None)
for s in self.sets: self.setsDict[s.name] = s for s in self.sets: self.setsDict[s.name] = s
if parameters == 'Monotonic': if parameters == 'Monotonic':
tmpdata = FuzzySet.fuzzyfy_series_old(data, sets) tmpdata = FuzzySet.fuzzyfy_series_old(data, self.sets)
flrs = FLR.generate_recurrent_flrs(tmpdata) flrs = FLR.generate_recurrent_flrs(tmpdata)
self.flrgs = self.generateFLRG(flrs) self.flrgs = self.generateFLRG(flrs)
else: else:
@ -162,15 +164,15 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
tmp_path = [] tmp_path = []
for c, e in enumerate(path, start=0): for c, e in enumerate(path, start=0):
tmp_path.append( e.membership( sample[c] ) ) tmp_path.append( e.membership( sample[c] ) )
flrg.appendLHS(e) flrg.append_lhs(e)
lhs_mv = np.prod(tmp_path) lhs_mv = np.prod(tmp_path)
if flrg.strLHS() not in flrgs: if flrg.str_lhs() not in flrgs:
flrgs[flrg.strLHS()] = flrg; flrgs[flrg.str_lhs()] = flrg;
for st in idx: for st in idx:
flrgs[flrg.strLHS()].appendRHSFuzzy(self.sets[st], lhs_mv*mv[st]) flrgs[flrg.str_lhs()].appendRHSFuzzy(self.sets[st], lhs_mv * mv[st])
tmp_fq = sum([lhs_mv*kk for kk in mv if kk > 0]) tmp_fq = sum([lhs_mv*kk for kk in mv if kk > 0])
@ -186,14 +188,14 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
flrg = ProbabilisticWeightedFLRG(self.order) flrg = ProbabilisticWeightedFLRG(self.order)
for kk in np.arange(k - self.order, k): for kk in np.arange(k - self.order, k):
flrg.appendLHS(flrs[kk].LHS) flrg.append_lhs(flrs[kk].LHS)
if self.dump: print("LHS: " + str(flrs[kk])) if self.dump: print("LHS: " + str(flrs[kk]))
if flrg.strLHS() in flrgs: if flrg.str_lhs() in flrgs:
flrgs[flrg.strLHS()].appendRHS(flrs[k-1].RHS) flrgs[flrg.str_lhs()].append_rhs(flrs[k - 1].RHS)
else: else:
flrgs[flrg.strLHS()] = flrg flrgs[flrg.str_lhs()] = flrg
flrgs[flrg.strLHS()].appendRHS(flrs[k-1].RHS) flrgs[flrg.str_lhs()].append_rhs(flrs[k - 1].RHS)
if self.dump: print("RHS: " + str(flrs[k-1])) if self.dump: print("RHS: " + str(flrs[k-1]))
self.global_frequency_count += 1 self.global_frequency_count += 1
@ -205,34 +207,34 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
flrg = ProbabilisticWeightedFLRG(self.order) flrg = ProbabilisticWeightedFLRG(self.order)
for k in np.arange(0, self.order): flrg.appendLHS(fzzy[k]) for k in np.arange(0, self.order): flrg.append_lhs(fzzy[k])
if flrg.strLHS() in self.flrgs: if flrg.str_lhs() in self.flrgs:
self.flrgs[flrg.strLHS()].appendRHS(fzzy[self.order]) self.flrgs[flrg.str_lhs()].append_rhs(fzzy[self.order])
else: else:
self.flrgs[flrg.strLHS()] = flrg self.flrgs[flrg.str_lhs()] = flrg
self.flrgs[flrg.strLHS()].appendRHS(fzzy[self.order]) self.flrgs[flrg.str_lhs()].append_rhs(fzzy[self.order])
self.global_frequency_count += 1 self.global_frequency_count += 1
def add_new_PWFLGR(self, flrg): def add_new_PWFLGR(self, flrg):
if flrg.strLHS() not in self.flrgs: if flrg.str_lhs() not in self.flrgs:
tmp = ProbabilisticWeightedFLRG(self.order) tmp = ProbabilisticWeightedFLRG(self.order)
for fs in flrg.LHS: tmp.appendLHS(fs) for fs in flrg.LHS: tmp.append_lhs(fs)
tmp.appendRHS(flrg.LHS[-1]) tmp.append_rhs(flrg.LHS[-1])
self.flrgs[tmp.strLHS()] = tmp; self.flrgs[tmp.str_lhs()] = tmp;
self.global_frequency_count += 1 self.global_frequency_count += 1
def get_flrg_global_probability(self, flrg): def get_flrg_global_probability(self, flrg):
if flrg.strLHS() in self.flrgs: if flrg.str_lhs() in self.flrgs:
return self.flrgs[flrg.strLHS()].frequency_count / self.global_frequency_count return self.flrgs[flrg.str_lhs()].frequency_count / self.global_frequency_count
else: else:
self.add_new_PWFLGR(flrg) self.add_new_PWFLGR(flrg)
return self.get_flrg_global_probability(flrg) return self.get_flrg_global_probability(flrg)
def get_midpoint(self, flrg): def get_midpoint(self, flrg):
if flrg.strLHS() in self.flrgs: if flrg.str_lhs() in self.flrgs:
tmp = self.flrgs[flrg.strLHS()] tmp = self.flrgs[flrg.str_lhs()]
ret = tmp.get_midpoint() #sum(np.array([tmp.get_RHSprobability(s) * self.setsDict[s].centroid for s in tmp.RHS])) ret = tmp.get_midpoint() #sum(np.array([tmp.get_RHSprobability(s) * self.setsDict[s].centroid for s in tmp.RHS]))
else: else:
pi = 1 / len(flrg.LHS) pi = 1 / len(flrg.LHS)
@ -241,8 +243,8 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
def get_conditional_probability(self, x, flrg): def get_conditional_probability(self, x, flrg):
if flrg.strLHS() in self.flrgs: if flrg.str_lhs() in self.flrgs:
_flrg = self.flrgs[flrg.strLHS()] _flrg = self.flrgs[flrg.str_lhs()]
cond = [] cond = []
for s in _flrg.RHS: for s in _flrg.RHS:
_set = self.setsDict[s] _set = self.setsDict[s]
@ -258,8 +260,8 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
return ret return ret
def get_upper(self, flrg): def get_upper(self, flrg):
if flrg.strLHS() in self.flrgs: if flrg.str_lhs() in self.flrgs:
tmp = self.flrgs[flrg.strLHS()] tmp = self.flrgs[flrg.str_lhs()]
ret = tmp.get_upper() ret = tmp.get_upper()
else: else:
pi = 1 / len(flrg.LHS) pi = 1 / len(flrg.LHS)
@ -267,8 +269,8 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
return ret return ret
def get_lower(self, flrg): def get_lower(self, flrg):
if flrg.strLHS() in self.flrgs: if flrg.str_lhs() in self.flrgs:
tmp = self.flrgs[flrg.strLHS()] tmp = self.flrgs[flrg.str_lhs()]
ret = tmp.get_lower() ret = tmp.get_lower()
else: else:
pi = 1 / len(flrg.LHS) pi = 1 / len(flrg.LHS)
@ -324,7 +326,7 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
for p in root.paths(): for p in root.paths():
path = list(reversed(list(filter(None.__ne__, p)))) path = list(reversed(list(filter(None.__ne__, p))))
flrg = hofts.HighOrderFLRG(self.order) flrg = hofts.HighOrderFLRG(self.order)
for kk in path: flrg.appendLHS(self.sets[kk]) for kk in path: flrg.append_lhs(self.sets[kk])
assert len(flrg.LHS) == subset.size, str(subset) + " -> " + str([s.name for s in flrg.LHS]) assert len(flrg.LHS) == subset.size, str(subset) + " -> " + str([s.name for s in flrg.LHS])
@ -350,7 +352,7 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
for kk in idx: for kk in idx:
flrg = hofts.HighOrderFLRG(self.order) flrg = hofts.HighOrderFLRG(self.order)
flrg.appendLHS(self.sets[kk]) flrg.append_lhs(self.sets[kk])
affected_flrgs.append(flrg) affected_flrgs.append(flrg)
affected_flrgs_memberships.append(mv[kk]) affected_flrgs_memberships.append(mv[kk])
@ -446,7 +448,7 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
for p in root.paths(): for p in root.paths():
path = list(reversed(list(filter(None.__ne__, p)))) path = list(reversed(list(filter(None.__ne__, p))))
flrg = hofts.HighOrderFLRG(self.order) flrg = hofts.HighOrderFLRG(self.order)
for kk in path: flrg.appendLHS(self.sets[kk]) for kk in path: flrg.append_lhs(self.sets[kk])
assert len(flrg.LHS) == subset.size, str(subset) + " -> " + str([s.name for s in flrg.LHS]) assert len(flrg.LHS) == subset.size, str(subset) + " -> " + str([s.name for s in flrg.LHS])
@ -473,7 +475,7 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
for kk in idx: for kk in idx:
flrg = hofts.HighOrderFLRG(self.order) flrg = hofts.HighOrderFLRG(self.order)
flrg.appendLHS(self.sets[kk]) flrg.append_lhs(self.sets[kk])
affected_flrgs.append(flrg) affected_flrgs.append(flrg)
affected_flrgs_memberships.append(mv[kk]) affected_flrgs_memberships.append(mv[kk])
for count, flrg in enumerate(affected_flrgs): for count, flrg in enumerate(affected_flrgs):

30
pyFTS/models/sadaei.py
View File

@ -8,6 +8,8 @@ refined exponentially weighted fuzzy time series and an improved harmony search,
import numpy as np import numpy as np
from pyFTS.common import FuzzySet,FLR,fts, flrg from pyFTS.common import FuzzySet,FLR,fts, flrg
default_c = 1.1
class ExponentialyWeightedFLRG(flrg.FLRG): class ExponentialyWeightedFLRG(flrg.FLRG):
"""First Order Exponentialy Weighted Fuzzy Logical Relationship Group""" """First Order Exponentialy Weighted Fuzzy Logical Relationship Group"""
@ -16,16 +18,19 @@ class ExponentialyWeightedFLRG(flrg.FLRG):
self.LHS = LHS self.LHS = LHS
self.RHS = [] self.RHS = []
self.count = 0.0 self.count = 0.0
self.c = kwargs.get("c",2.0) self.c = kwargs.get("c",default_c)
self.w = None
def append(self, c): def append(self, c):
self.RHS.append(c) self.RHS.append(c)
self.count = self.count + 1.0 self.count = self.count + 1.0
def weights(self): def weights(self):
wei = [self.c ** k for k in np.arange(0.0, self.count, 1.0)] if self.w is None:
tot = sum(wei) wei = [self.c ** k for k in np.arange(0.0, self.count, 1.0)]
return np.array([k / tot for k in wei]) tot = sum(wei)
self.w = np.array([k / tot for k in wei])
return self.w
def __str__(self): def __str__(self):
tmp = self.LHS.name + " -> " tmp = self.LHS.name + " -> "
@ -50,9 +55,9 @@ class ExponentialyWeightedFTS(fts.FTS):
super(ExponentialyWeightedFTS, self).__init__(1, "EWFTS", **kwargs) super(ExponentialyWeightedFTS, self).__init__(1, "EWFTS", **kwargs)
self.name = "Exponentialy Weighted FTS" self.name = "Exponentialy Weighted FTS"
self.detail = "Sadaei" self.detail = "Sadaei"
self.c = 1 self.c = kwargs.get('c', default_c)
def generateFLRG(self, flrs, c): def generate_flrg(self, flrs, c):
flrgs = {} flrgs = {}
for flr in flrs: for flr in flrs:
if flr.LHS.name in flrgs: if flr.LHS.name in flrgs:
@ -62,13 +67,14 @@ class ExponentialyWeightedFTS(fts.FTS):
flrgs[flr.LHS.name].append(flr.RHS) flrgs[flr.LHS.name].append(flr.RHS)
return (flrgs) return (flrgs)
def train(self, data, sets,order=1,parameters=1.05): def train(self, data, **kwargs):
self.c = parameters self.c = kwargs.get('parameters', default_c)
self.sets = sets if kwargs.get('sets', None) is not None:
self.sets = kwargs.get('sets', None)
ndata = self.apply_transformations(data) ndata = self.apply_transformations(data)
tmpdata = FuzzySet.fuzzyfy_series_old(ndata, sets) tmpdata = FuzzySet.fuzzyfy_series_old(ndata, self.sets)
flrs = FLR.generate_recurrent_flrs(tmpdata) flrs = FLR.generate_recurrent_flrs(tmpdata)
self.flrgs = self.generateFLRG(flrs, self.c) self.flrgs = self.generate_flrg(flrs, self.c)
def forecast(self, data, **kwargs): def forecast(self, data, **kwargs):
l = 1 l = 1
@ -95,6 +101,6 @@ class ExponentialyWeightedFTS(fts.FTS):
ret.append(mp.dot(flrg.weights())) ret.append(mp.dot(flrg.weights()))
ret = self.apply_inverse_transformations(ret, params=[data[self.order - 1:]]) ret = self.apply_inverse_transformations(ret, params=[data])
return ret return ret

10
pyFTS/models/song.py
View File

@ -37,9 +37,9 @@ class ConventionalFTS(fts.FTS):
return r return r
def train(self, data, sets,order=1,parameters=None): def train(self, data, **kwargs):
if sets != None: if kwargs.get('sets', None) is not None:
self.sets = sets self.sets = kwargs.get('sets', None)
ndata = self.apply_transformations(data) ndata = self.apply_transformations(data)
tmpdata = FuzzySet.fuzzyfy_series_old(ndata, self.sets) tmpdata = FuzzySet.fuzzyfy_series_old(ndata, self.sets)
flrs = FLR.generate_non_recurrent_flrs(tmpdata) flrs = FLR.generate_non_recurrent_flrs(tmpdata)
@ -71,3 +71,7 @@ class ConventionalFTS(fts.FTS):
ret = self.apply_inverse_transformations(ret, params=[data]) ret = self.apply_inverse_transformations(ret, params=[data])
return ret return ret
def __str__(self):
tmp = self.name + ":\n"
return tmp + str(self.R)

9
pyFTS/models/yu.py
View File

@ -56,10 +56,11 @@ class WeightedFTS(fts.FTS):
flrgs[flr.LHS.name].append(flr.RHS) flrgs[flr.LHS.name].append(flr.RHS)
return (flrgs) return (flrgs)
def train(self, data, sets,order=1,parameters=None): def train(self, data, **kwargs):
self.sets = sets if kwargs.get('sets', None) is not None:
self.sets = kwargs.get('sets', None)
ndata = self.apply_transformations(data) ndata = self.apply_transformations(data)
tmpdata = FuzzySet.fuzzyfy_series_old(ndata, sets) tmpdata = FuzzySet.fuzzyfy_series_old(ndata, self.sets)
flrs = FLR.generate_recurrent_flrs(tmpdata) flrs = FLR.generate_recurrent_flrs(tmpdata)
self.flrgs = self.generate_FLRG(flrs) self.flrgs = self.generate_FLRG(flrs)
@ -88,6 +89,6 @@ class WeightedFTS(fts.FTS):
ret.append(mp.dot(flrg.weights())) ret.append(mp.dot(flrg.weights()))
ret = self.apply_inverse_transformations(ret, params=[data[self.order - 1:]]) ret = self.apply_inverse_transformations(ret, params=[data])
return ret return ret

315
pyFTS/notebooks/Benchmarks.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

462
pyFTS/notebooks/Chen - ConventionalFTS.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

462
pyFTS/notebooks/Cheng - TrendWeightedFTS.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

601
pyFTS/notebooks/HighOrderFTS.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

447
pyFTS/notebooks/Hwang - HighOrderFTS.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

460
pyFTS/notebooks/Ismail & Efendi - ImprovedWeightedFTS.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

315
pyFTS/notebooks/Partitioners.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

508
pyFTS/notebooks/Sadaei et Al - ExponentialyWeightedFTS.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

477
pyFTS/notebooks/Song - ConventionalFTS.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

461
pyFTS/notebooks/Yu - WeightedFTS.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

11
pyFTS/partitioners/Grid.py
View File

@ -8,7 +8,16 @@ from pyFTS.partitioners import partitioner
class GridPartitioner(partitioner.Partitioner): class GridPartitioner(partitioner.Partitioner):
"""Even Length Grid Partitioner""" """Even Length Grid Partitioner"""
def __init__(self, data, npart, func = Membership.trimf, transformation=None, indexer=None): def __init__(self, data, npart, func = Membership.trimf, transformation=None, indexer=None):
"""
Even Length Grid Partitioner
:param data: Training data of which the universe of discourse will be extracted. The universe of discourse is the open interval between the minimum and maximum values of the training data.
:param npart: The number of universe of discourse partitions, i.e., the number of fuzzy sets that will be created
:param func: Fuzzy membership function (pyFTS.common.Membership)
:param transformation: data transformation to be applied on data
:param indexer:
"""
super(GridPartitioner, self).__init__("Grid", data, npart, func=func, transformation=transformation, indexer=indexer) super(GridPartitioner, self).__init__("Grid", data, npart, func=func, transformation=transformation, indexer=indexer)
def build(self, data): def build(self, data):
@ -18,7 +27,7 @@ class GridPartitioner(partitioner.Partitioner):
partlen = dlen / self.partitions partlen = dlen / self.partitions
count = 0 count = 0
for c in np.linspace(self.min, self.max, self.partitions): for c in np.arange(self.min, self.max, partlen):
if self.membership_function == Membership.trimf: if self.membership_function == Membership.trimf:
sets.append( sets.append(
FuzzySet.FuzzySet(self.prefix + str(count), Membership.trimf, [c - partlen, c, c + partlen],c)) FuzzySet.FuzzySet(self.prefix + str(count), Membership.trimf, [c - partlen, c, c + partlen],c))

6
pyFTS/partitioners/partitioner.py
View File

@ -12,9 +12,9 @@ class Partitioner(object):
""" """
Universe of Discourse partitioner scheme. Split data on several fuzzy sets Universe of Discourse partitioner scheme. Split data on several fuzzy sets
:param name: partitioner name :param name: partitioner name
:param data: original data to be partitioned :param data: Training data of which the universe of discourse will be extracted. The universe of discourse is the open interval between the minimum and maximum values of the training data.
:param npart: number of partitions :param npart: The number of universe of discourse partitions, i.e., the number of fuzzy sets that will be created
:param func: membership function :param func: Fuzzy membership function (pyFTS.common.Membership)
:param names: list of partitions names. If None is given the partitions will be auto named with prefix :param names: list of partitions names. If None is given the partitions will be auto named with prefix
:param prefix: prefix of auto generated partition names :param prefix: prefix of auto generated partition names
:param transformation: data transformation to be applied on data :param transformation: data transformation to be applied on data

361
pyFTS/tests/general.py
View File

@ -9,9 +9,7 @@ import numpy as np
import pandas as pd import pandas as pd
from pyFTS.common import Transformations from pyFTS.common import Transformations
#from pyFTS.benchmarks import benchmarks as bchmk from pyFTS.benchmarks import benchmarks as bchmk
os.chdir("/home/petronio/dados/Dropbox/Doutorado/Codigos/")
bc = Transformations.BoxCox(0) bc = Transformations.BoxCox(0)
diff = Transformations.Differential(1) diff = Transformations.Differential(1)
@ -21,358 +19,21 @@ diff = Transformations.Differential(1)
DATASETS DATASETS
""" """
#enrollments = pd.read_csv("DataSets/Enrollments.csv", sep=";") from pyFTS.data import Enrollments
#enrollments = np.array(enrollments["Enrollments"])
passengers = pd.read_csv("DataSets/AirPassengers.csv", sep=",") data = Enrollments.get_data()
passengers = np.array(passengers["Passengers"])
#sunspots = pd.read_csv("DataSets/sunspots.csv", sep=",")
#sunspots = np.array(sunspots["SUNACTIVITY"])
#gauss = random.normal(0,1.0,5000)
#gauss_teste = random.normal(0,1.0,400)
#taiexpd = pd.read_csv("DataSets/TAIEX.csv", sep=",")
#taiex = np.array(taiexpd["avg"][:5000])
#del(taiexpd)
#nasdaqpd = pd.read_csv("DataSets/NASDAQ_IXIC.csv", sep=",")
#nasdaq = np.array(nasdaqpd["avg"][0:5000])
#del(nasdaqpd)
#sp500pd = pd.read_csv("DataSets/S&P500.csv", sep=",")
#sp500 = np.array(sp500pd["Avg"][11000:])
#del(sp500pd)
#sondapd = pd.read_csv("DataSets/SONDA_BSB_HOURLY_AVG.csv", sep=";")
#sondapd = sondapd.dropna(axis=0, how='any')
#sonda = np.array(sondapd["glo_avg"])
#del(sondapd)
#bestpd = pd.read_csv("DataSets/BEST_TAVG.csv", sep=";")
#best = np.array(bestpd["Anomaly"])
#del(bestpd)
#print(lag)
#print(a)
#'''
'''
sonda = pd.read_csv("DataSets/SONDA_BSB_15MIN_AVG.csv", sep=";")
sonda['data'] = pd.to_datetime(sonda['data'])
sonda = sonda[:][527041:].dropna()
sonda.index = np.arange(0,len(sonda.index))
sonda_treino = sonda[:105313].dropna()
sonda_teste = sonda[105314:].dropna()
'''
from pyFTS.partitioners import Grid from pyFTS.partitioners import Grid
from pyFTS import song, chen, yu, sadaei, ismailefendi, cheng from pyFTS.models import song, chen, yu, sadaei, ismailefendi, cheng, hofts
train = passengers[:100] train = data
test = passengers[100:] test = data
fs = Grid.GridPartitioner(train, 10, transformation=bc) fs = Grid.GridPartitioner(train, 10) #, transformation=bc)
methods = [song.ConventionalFTS, chen.ConventionalFTS, yu.WeightedFTS, sadaei.ExponentialyWeightedFTS, #tmp = bchmk.simpleSearch_RMSE(train, test, hofts.HighOrderFTS, range(4,12), [2], tam=[10, 5])
ismailefendi.ImprovedWeightedFTS, cheng.TrendWeightedFTS]
#fig, axes = plt.subplots(nrows=1, ncols=1, figsize=[15, 5]) model = hofts.HighOrderFTS("", partitioner=fs)
model.fit(train, order=3)
#axes.plot(test, label="Original") print(model)
for method in methods:
model = method("")
model.append_transformation(bc)
model.train(train, sets=fs.sets)
forecasts = model.forecast(test)
print(forecasts)
#ix_m15 = SeasonalIndexer.DateTimeSeasonalIndexer('data',[SeasonalIndexer.DateTime.minute],[15],'glo_avg', name='m15')
#fs1 = Grid.GridPartitioner(sonda_treino, 50, transformation=diff, indexer=ix_m15)
#ix = cUtil.load_obj("models/sonda_ix_Mhm15.pkl")
#fs = cUtil.load_obj("models/sonda_fs_Entropy40_diff.pkl")
#from pyFTS.models import msfts
#obj = msfts.MultiSeasonalFTS("sonda_msfts_Entropy40_Mhm15", indexer=ix)
#obj.append_transformation(diff)
#obj.train(sonda_treino, fs.sets)
#cUtil.persist_obj(obj, "models/sonda_msfts_Entropy40_Mhm15.pkl")
#ftse = cUtil.load_obj("models/sonda_ensemble_msfts.pkl")
#tmp = ftse.forecast_distribution(sonda_teste[850:860], h=0.5, method="gaussian")
#print(tmp[0])
#'''
'''
from pyFTS.models.seasonal import SeasonalIndexer
indexers = []
for i in ["models/sonda_ix_Mhm15.pkl"]: #, "models/sonda_ix_m15.pkl", "models/sonda_ix_Mh.pkl", ]:
obj = cUtil.load_obj(i)
indexers.append( obj )
print(obj)
partitioners = []
transformations = [""] #, "_diff"]
for max_part in [30, 40, 50, 60, 70, 80, 90]:
for t in transformations:
obj = cUtil.load_obj("models/sonda_fs_grid_" + str(max_part) + t + ".pkl")
partitioners.append( obj )
print(obj)
from pyFTS.ensemble import ensemble, multiseasonal
fts = multiseasonal.SeasonalEnsembleFTS("sonda_msfts_Mhm15")
fts.indexers = indexers
fts.partitioners = partitioners
fts.indexer = indexers[0]
fts.train(sonda_treino, sets=None)
'''
#'''
#ix = cUtil.load_obj("models/sonda_ix_m15.pkl")
#ftse = cUtil.load_obj("models/msfts_Grid40_diff_Mhm15.pkl")
#ftse.indexer = ix
#ftse.update_uod(sonda_treino)
#tmp = ftse.forecast_distribution(sonda_teste,h=1)
#tmp = ftse.forecast(sonda_teste,h=1)
#tmp[5].plot()
#'''
'''
from pyFTS.benchmarks import benchmarks as bchmk
#from pyFTS.benchmarks import distributed_benchmarks as bchmk
#from pyFTS.benchmarks import parallel_benchmarks as bchmk
from pyFTS.benchmarks import Util
from pyFTS.benchmarks import arima, quantreg, Measures
#Util.cast_dataframe_to_synthetic_point("experiments/taiex_point_analitic.csv","experiments/taiex_point_sintetic.csv",11)
#Util.plot_dataframe_point("experiments/taiex_point_sintetic.csv","experiments/taiex_point_analitic.csv",11)
"""
arima100 = arima.ARIMA("", alpha=0.25)
#tmp.append_transformation(diff)
arima100.train(passengers, None, order=(1,0,0))
arima101 = arima.ARIMA("", alpha=0.25)
#tmp.append_transformation(diff)
arima101.train(passengers, None, order=(1,0,1))
arima200 = arima.ARIMA("", alpha=0.25)
#tmp.append_transformation(diff)
arima200.train(passengers, None, order=(2,0,0))
arima201 = arima.ARIMA("", alpha=0.25)
#tmp.append_transformation(diff)
arima201.train(passengers, None, order=(2,0,1))
#tmp = quantreg.QuantileRegression("", alpha=0.25, dist=True)
#tmp.append_transformation(diff)
#tmp.train(sunspots[:150], None, order=1)
#teste = tmp.forecast_ahead_interval(sunspots[150:155], 5)
#teste = tmp.forecast_ahead_distribution(nasdaq[1600:1604], steps=5, resolution=50)
bchmk.plot_compared_series(enrollments,[tmp], ['blue','red'], points=False, intervals=True)
#print(sunspots[150:155])
#print(teste)
#kk = Measures.get_interval_statistics(nasdaq[1600:1605], tmp)
#print(kk)
"""
"""
bchmk.point_sliding_window(sonda, 9000, train=0.8, inc=0.4,#models=[yu.WeightedFTS], # #
partitioners=[Grid.GridPartitioner], #Entropy.EntropyPartitioner], # FCM.FCMPartitioner, ],
partitions= np.arange(10,200,step=10), #transformation=diff,
dump=True, save=True, file="experiments/sondaws_point_analytic.csv",
nodes=['192.168.0.103', '192.168.0.106', '192.168.0.108', '192.168.0.109']) #, depends=[hofts, ifts])
bchmk.point_sliding_window(sonda, 9000, train=0.8, inc=0.4, #models=[yu.WeightedFTS], # #
partitioners=[Grid.GridPartitioner], #Entropy.EntropyPartitioner], # FCM.FCMPartitioner, ],
partitions= np.arange(3,20,step=2), #transformation=diff,
dump=True, save=True, file="experiments/sondaws_point_analytic_diff.csv",
nodes=['192.168.0.103', '192.168.0.106', '192.168.0.108', '192.168.0.109']) #, depends=[hofts, ifts])
bchmk.interval_sliding_window(best, 5000, train=0.8, inc=0.8,#models=[yu.WeightedFTS], # #
partitioners=[Grid.GridPartitioner], #Entropy.EntropyPartitioner], # FCM.FCMPartitioner, ],
partitions= np.arange(10,200,step=10),
dump=True, save=True, file="experiments/best"
"_interval_analytic.csv",
nodes=['192.168.0.103', '192.168.0.106', '192.168.0.108', '192.168.0.109']) #, depends=[hofts, ifts])
bchmk.interval_sliding_window(taiex, 2000, train=0.8, inc=0.1, #models=[yu.WeightedFTS], # #
partitioners=[Grid.GridPartitioner], #Entropy.EntropyPartitioner], # FCM.FCMPartitioner, ],
partitions= np.arange(3,20,step=2), transformation=diff,
dump=True, save=True, file="experiments/taiex_interval_analytic_diff.csv",
nodes=['192.168.0.103', '192.168.0.106', '192.168.0.108', '192.168.0.109']) #, depends=[hofts, ifts])
bchmk.ahead_sliding_window(sonda, 10000, steps=10, resolution=10, train=0.2, inc=0.2,
partitioners=[Grid.GridPartitioner],
partitions= np.arange(10,200,step=10), indexer=ix,
dump=True, save=True, file="experiments/sondawind_ahead_analytic.csv",
nodes=['192.168.0.106', '192.168.0.108', '192.168.0.109']) #, depends=[hofts, ifts])
bchmk.ahead_sliding_window(sonda, 10000, steps=10, resolution=10, train=0.2, inc=0.2,
partitioners=[Grid.GridPartitioner],
partitions= np.arange(3,20,step=2), transformation=diff, indexer=ix,
dump=True, save=True, file="experiments/sondawind_ahead_analytic_diff.csv",
nodes=['192.168.0.106', '192.168.0.108', '192.168.0.109']) #, depends=[hofts, ifts])
from pyFTS import pwfts
from pyFTS.common import Transformations
from pyFTS.partitioners import Grid
#diff = Transformations.Differential(1)
#fs = Grid.GridPartitioner(best, 190) #, transformation=diff)
#model = pwfts.ProbabilisticWeightedFTS("FTS 1")
#model.append_transformation(diff)
#model.train(best[0:1600],fs.sets, order=3)
#bchmk.plot_compared_intervals_ahead(best[1600:1700],[model], ['blue','red'],
# distributions=[True], save=True, file="pictures/best_ahead_forecasts",
# time_from=40, time_to=60, resolution=100)
experiments = [
["experiments/taiex_point_synthetic_diff.csv","experiments/taiex_point_analytic_diff.csv",16],
["experiments/nasdaq_point_synthetic_diff.csv","experiments/nasdaq_point_analytic_diff.csv", 11],
["experiments/sp500_point_synthetic_diff.csv","experiments/sp500_point_analytic_diff.csv", 21],
["experiments/best_point_synthetic_diff.csv","experiments/best_point_analytic_diff.csv", 13],
["experiments/sondasun_point_synthetic_diff.csv","experiments/sondasun_point_analytic_diff.csv", 15],
["experiments/sondawind_point_synthetic_diff.csv","experiments/sondawind_point_analytic_diff.csv", 8],
["experiments/gauss_point_synthetic_diff.csv","experiments/gauss_point_analytic_diff.csv", 16]
]
Util.unified_scaled_point(experiments,tam=[15,8],save=True,file="pictures/unified_experiments_point.png",
ignore=['ARIMA(1,0,0)','ARIMA(2,0,0)','ARIMA(2,0,1)','ARIMA(2,0,2)','QAR(2)'],
replace=[['ARIMA','ARIMA'],['QAR','QAR']])
'''
'''
experiments = [
["experiments/taiex_interval_synthetic.csv","experiments/taiex_interval_analytic.csv",16],
["experiments/nasdaq_interval_synthetic_diff.csv","experiments/nasdaq_interval_analytic_diff.csv",11],
["experiments/sp500_interval_synthetic_diff.csv","experiments/sp500_interval_analytic_diff.csv", 11],
["experiments/best_interval_synthetic_diff.csv","experiments/best_interval_analytic_diff.csv",13],
["experiments/sondasun_interval_synthetic_diff.csv","experiments/sondasun_interval_analytic_diff.csv",8],
["experiments/sondawind_interval_synthetic_diff.csv","experiments/sondawind_interval_analytic_diff.csv",8],
["experiments/gauss_interval_synthetic_diff.csv","experiments/gauss_interval_analytic_diff.csv", 8]
]
Util.unified_scaled_interval(experiments,tam=[15,8],save=True,file="pictures/unified_experiments_interval.png",
ignore=['ARIMA(1,0,0)', 'ARIMA(2,0,0)', 'ARIMA(2,0,1)', 'ARIMA(2,0,2)', 'QAR(2)'],
replace=[['ARIMA(1,0,1) - 0.05', 'ARIMA 0.05'], ['ARIMA(1,0,1) - 0.25', 'ARIMA 0.25'],
['QAR(1) - 0.05', 'QAR 0.05'], ['QAR(1) - 0.25', 'QAR 0.25']])
Util.unified_scaled_interval_pinball(experiments,tam=[15,8],save=True,file="pictures/unified_experiments_interval_pinball.png",
ignore=['ARIMA(1,0,0)', 'ARIMA(2,0,0)', 'ARIMA(2,0,1)', 'ARIMA(2,0,2)', 'QAR(2)'],
replace=[['ARIMA(1,0,1) - 0.05', 'ARIMA 0.05'], ['ARIMA(1,0,1) - 0.25', 'ARIMA 0.25'],
['QAR(1) - 0.05', 'QAR 0.05'], ['QAR(1) - 0.25', 'QAR 0.25']])
'''
'''
experiments = [
["experiments/taiex_ahead_synthetic_diff.csv","experiments/taiex_ahead_analytic_diff.csv",16],
["experiments/nasdaq_ahead_synthetic_diff.csv","experiments/nasdaq_ahead_analytic_diff.csv",11],
["experiments/sp500_ahead_synthetic_diff.csv","experiments/sp500_ahead_analytic_diff.csv", 21],
["experiments/best_ahead_synthetic_diff.csv","experiments/best_ahead_analytic_diff.csv", 24],
["experiments/sondasun_ahead_synthetic_diff.csv","experiments/sondasun_ahead_analytic_diff.csv",13],
["experiments/sondawind_ahead_synthetic_diff.csv","experiments/sondawind_ahead_analytic_diff.csv", 13],
["experiments/gauss_ahead_synthetic_diff.csv","experiments/gauss_ahead_analytic_diff.csv",16]
]
Util.unified_scaled_ahead(experiments,tam=[15,8],save=True,file="pictures/unified_experiments_ahead.png",
ignore=['ARIMA(1,0,0)', 'ARIMA(0,0,1)', 'ARIMA(2,0,0)', 'ARIMA(2,0,1)',
'ARIMA(2,0,2)', 'QAR(2)', 'ARIMA0.05'],
replace=[['ARIMA(1,0,1) - 0.05', 'ARIMA 0.05'], ['ARIMA(1,0,1) - 0.25', 'ARIMA 0.25'],
['QAR(1) - 0.05', 'QAR 0.05'], ['QAR(1) - 0.25', 'QAR 0.25']])
'''
'''
from pyFTS.partitioners import Grid
from pyFTS import sfts
#print(ix.get_season_of_data(best[:2000]))
#print(ix.get_season_by_index(45))
#ix = SeasonalIndexer.LinearSeasonalIndexer([720,24],[False,True,False])
#print(ix.get_season_of_data(sonda[6500:9000])[-20:])
diff = Transformations.Differential(1)
fs = Grid.GridPartitioner(sonda[:9000], 10, transformation=diff)
tmp = sfts.SeasonalFTS("")
tmp.indexer = ix
tmp.append_transformation(diff)
#tmp = pwfts.ProbabilisticWeightedFTS("")
#tmp.append_transformation(diff)
tmp.train(sonda[:9000], fs.sets, order=1)
x = tmp.forecast(sonda[:1610])
#print(taiex[1600:1610])
print(x)
'''