common.Util.plot_rules; CVFTS bugfixes and other nonstationary sets improvements

pyFTS/common/FLR.py

@@ -22,7 +22,7 @@ class FLR(object):
         self.RHS = RHS
 
     def __str__(self):
-        return self.LHS + " -> " + self.RHS
+        return str(self.LHS) + " -> " + str(self.RHS)
 
 
 class IndexedFLR(FLR):

pyFTS/common/FuzzySet.py

@@ -59,8 +59,9 @@ class FuzzySet(object):
 
 
 def set_ordered(fuzzySets):
-    tmp1 = [fuzzySets[k] for k in fuzzySets.keys()]
+    if len(fuzzySets) > 0:
-    return [k.name for k in sorted(tmp1, key=lambda x: x.centroid)]
+        tmp1 = [fuzzySets[k] for k in fuzzySets.keys()]
+        return [k.name for k in sorted(tmp1, key=lambda x: x.centroid)]
 
 
 def fuzzyfy_instance(inst, fuzzySets, ordered_sets=None):

pyFTS/common/Util.py

@@ -8,6 +8,43 @@ import dill
 import numpy as np
 
 
+def plot_rules(model, size=[5, 5], axis=None):
+    if axis is None:
+        fig, axis = plt.subplots(nrows=1, ncols=1, figsize=size)
+
+    for ct, key in enumerate(model.partitioner.ordered_sets):
+        fs = model.sets[key]
+        axis.plot([0, 1, 0], fs.parameters, label=fs.name)
+        axis.axhline(fs.centroid, c="lightgray", alpha=0.5)
+
+    axis.set_xlim([0, len(model.partitioner.ordered_sets)])
+    axis.set_xticks(range(0,len(model.partitioner.ordered_sets)))
+    tmp = ['']
+    tmp.extend(model.partitioner.ordered_sets)
+    axis.set_xticklabels(tmp)
+    axis.set_ylim([model.partitioner.min, model.partitioner.max])
+    axis.set_yticks([model.sets[k].centroid for k in model.partitioner.ordered_sets])
+    axis.set_yticklabels([str(round(model.sets[k].centroid,1)) + " - " + k
+                          for k in model.partitioner.ordered_sets])
+
+    if not model.is_high_order:
+        for ct, key in enumerate(model.partitioner.ordered_sets):
+            if key in model.flrgs:
+                flrg = model.flrgs[key]
+                orig = model.sets[key].centroid
+                axis.plot([ct+1],[orig],'o')
+                for rhs in flrg.RHS:
+                    dest = model.sets[rhs].centroid
+                    axis.arrow(ct+1.1, orig, 0.8, dest - orig, #length_includes_head=True,
+                               head_width=0.1, head_length=0.1, shape='full', overhang=0,
+                               fc='k', ec='k')
+    plt.tight_layout()
+    plt.show()
+    print("fim")
+
+
+
+
 current_milli_time = lambda: int(round(time.time() * 1000))
 
 

pyFTS/common/fts.py

@@ -38,6 +38,7 @@ class FTS(object):
         self.auto_update = False
         self.benchmark_only = False
         self.indexer = None
+        self.uod_clip = kwargs.get("uod_clip", True)
 
     def fuzzy(self, data):
         """
@@ -75,7 +76,8 @@ class FTS(object):
         else:
             ndata = self.apply_transformations(data)
 
-        ndata = np.clip(ndata, self.original_min, self.original_max)
+        if self.uod_clip:
+            ndata = np.clip(ndata, self.original_min, self.original_max)
 
         if 'distributed' in kwargs:
             distributed = kwargs.pop('distributed')

pyFTS/models/nonstationary/common.py

@@ -56,12 +56,13 @@ class FuzzySet(FS.FuzzySet):
     
     def perform_location(self, t, param):
         if self.location is None:
-            return param
+            inc = t
-    
+        else:
-        l = len(self.location)
+
-    
+            l = len(self.location)
-        inc = sum([self.location[k](t + self.location_roots[k], self.location_params[k]) for k in np.arange(0, l)])
+
-    
+            inc = sum([self.location[k](t + self.location_roots[k], self.location_params[k]) for k in np.arange(0, l)])
+
         if self.mf == Membership.gaussmf:
             # changes only the mean parameter
             return [param[0] + inc, param[1]]
@@ -76,12 +77,13 @@ class FuzzySet(FS.FuzzySet):
     
     def perform_width(self, t, param):
         if self.width is None:
-            return param
+            inc = t
-    
+        else:
-        l = len(self.width)
+
-    
+            l = len(self.width)
-        inc = sum([self.width[k](t + self.width_roots[k], self.width_params[k]) for k in np.arange(0, l)])
+
-    
+            inc = sum([self.width[k](t + self.width_roots[k], self.width_params[k]) for k in np.arange(0, l)])
+
         if self.mf == Membership.gaussmf:
             # changes only the variance parameter
             return [param[0], param[1] + inc]

pyFTS/models/nonstationary/cvfts.py

@@ -5,26 +5,20 @@ from pyFTS.common import FLR
 
 
 class ConditionalVarianceFTS(chen.ConventionalFTS):
-    def __init__(self, name, **kwargs):
+    def __init__(self, **kwargs):
-        super(ConditionalVarianceFTS, self).__init__("CVFTS " + name, **kwargs)
+        super(ConditionalVarianceFTS, self).__init__(**kwargs)
         self.name = "Conditional Variance FTS"
+        self.shortname = "CVFTS "
         self.detail = ""
         self.flrgs = {}
-        #self.append_transformation(Transformations.Differential(1))
+        if self.partitioner is not None:
-        if self.partitioner is None:
-            self.min_tx = None
-            self.max_tx = None
-        else:
-            self.min_tx = self.partitioner.min
-            self.max_tx = self.partitioner.max
             self.append_transformation(self.partitioner.transformation)
 
         self.min_stack = [0,0,0]
         self.max_stack = [0,0,0]
+        self.uod_clip = False
 
     def train(self, ndata, **kwargs):
-        self.min_tx = min(ndata)
-        self.max_tx = max(ndata)
 
         tmpdata = common.fuzzySeries(ndata, self.sets, self.partitioner.ordered_sets, method='fuzzy', const_t=0)
         flrs = FLR.generate_non_recurrent_flrs(tmpdata)
@@ -44,10 +38,10 @@ class ConditionalVarianceFTS(chen.ConventionalFTS):
     def perturbation_factors(self, data):
         _max = 0
         _min = 0
-        if data < self.min_tx:
+        if data < self.original_min:
-            _min = data - self.min_tx if data < 0 else self.min_tx - data
+            _min = data - self.original_min if data < 0 else self.original_min - data
-        elif data > self.max_tx:
+        elif data > self.original_max:
-            _max = data - self.max_tx if data > 0 else self.max_tx - data
+            _max = data - self.original_max if data > 0 else self.original_max - data
         self.min_stack.pop(2)
         self.min_stack.insert(0,_min)
         _min = min(self.min_stack)
@@ -96,21 +90,22 @@ class ConditionalVarianceFTS(chen.ConventionalFTS):
 
             if len(affected_sets) == 1:
                 ix = affected_sets[0][0]
-                aset = self.sets[ix]
+                aset = self.partitioner.ordered_sets[ix]
-                if aset.name in self.flrgs:
+                if aset in self.flrgs:
-                    tmp.append(self.flrgs[aset.name].get_midpoint(perturb[ix]))
+                    tmp.append(self.flrgs[aset].get_midpoint(perturb[ix]))
                 else:
-                    print('naive')
+                    fuzzy_set = self.sets[aset]
-                    tmp.append(aset.get_midpoint(perturb[ix]))
+                    tmp.append(fuzzy_set.get_midpoint(perturb[ix]))
             else:
                 for aset in affected_sets:
                     ix = aset[0]
-                    fs = self.sets[ix]
+                    fs = self.partitioner.ordered_sets[ix]
                     tdisp = perturb[ix]
-                    if fs.name in self.flrgs:
+                    if fs in self.flrgs:
-                        tmp.append(self.flrgs[fs.name].get_midpoint(tdisp) * aset[1])
+                        tmp.append(self.flrgs[fs].get_midpoint(tdisp) * aset[1])
                     else:
-                        tmp.append(fs.get_midpoint(tdisp) * aset[1])
+                        fuzzy_set = self.sets[fs]
+                        tmp.append(fuzzy_set.get_midpoint(tdisp) * aset[1])
 
             pto = sum(tmp)
 
@@ -137,24 +132,26 @@ class ConditionalVarianceFTS(chen.ConventionalFTS):
 
             if len(affected_sets) == 1:
                 ix = affected_sets[0][0]
-                aset = self.sets[ix]
+                aset = self.partitioner.ordered_sets[ix]
-                if aset.name in self.flrgs:
+                if aset in self.flrgs:
-                    lower.append(self.flrgs[aset.name].get_lower(perturb[ix]))
+                    lower.append(self.flrgs[aset].get_lower(perturb[ix]))
-                    upper.append(self.flrgs[aset.name].get_upper(perturb[ix]))
+                    upper.append(self.flrgs[aset].get_upper(perturb[ix]))
                 else:
-                    lower.append(aset.get_lower(perturb[ix]))
+                    fuzzy_set = self.sets[aset]
-                    upper.append(aset.get_upper(perturb[ix]))
+                    lower.append(fuzzy_set.get_lower(perturb[ix]))
+                    upper.append(fuzzy_set.get_upper(perturb[ix]))
             else:
                 for aset in affected_sets:
                     ix = aset[0]
-                    fs = self.sets[ix]
+                    fs = self.partitioner.ordered_sets[ix]
                     tdisp = perturb[ix]
-                    if fs.name in self.flrgs:
+                    if fs in self.flrgs:
-                        lower.append(self.flrgs[fs.name].get_lower(tdisp) * aset[1])
+                        lower.append(self.flrgs[fs].get_lower(tdisp) * aset[1])
-                        upper.append(self.flrgs[fs.name].get_upper(tdisp) * aset[1])
+                        upper.append(self.flrgs[fs].get_upper(tdisp) * aset[1])
                     else:
-                        lower.append(fs.get_lower(tdisp) * aset[1])
+                        fuzzy_set = self.sets[fs]
-                        upper.append(fs.get_upper(tdisp) * aset[1])
+                        lower.append(fuzzy_set.get_lower(tdisp) * aset[1])
+                        upper.append(fuzzy_set.get_upper(tdisp) * aset[1])
 
             itvl = [sum(lower), sum(upper)]
 

pyFTS/models/nonstationary/partitioners.py

@@ -109,22 +109,26 @@ class PolynomialNonStationaryPartitioner(partitioner.Partitioner):
         pass
 
 
-class ConstantNonStationaryPartitioner(partitioner.Partitioner):
+class SimpleNonStationaryPartitioner(partitioner.Partitioner):
     """
     Non Stationary Universe of Discourse Partitioner
     """
 
     def __init__(self, data, part, **kwargs):
         """"""
-        super(ConstantNonStationaryPartitioner, self).__init__(name=part.name, data=data, npart=part.partitions,
+        super(SimpleNonStationaryPartitioner, self).__init__(name=part.name, data=data, npart=part.partitions,
-                                                                 func=part.membership_function, names=part.setnames,
+                                                             func=part.membership_function, names=part.setnames,
-                                                                 prefix=part.prefix, transformation=part.transformation,
+                                                             prefix=part.prefix, transformation=part.transformation,
-                                                                 indexer=part.indexer)
+                                                             indexer=part.indexer)#, preprocess=False)
-
-        self.sets = {}
 
         for key in part.sets.keys():
             set = part.sets[key]
             tmp = common.FuzzySet(set.name, set.mf, set.parameters, **kwargs)
+            tmp.centroid = set.centroid
 
             self.sets[key] =tmp
+
+        self.ordered_sets = stationary_fs.set_ordered(self.sets)
+
+    def build(self, data):
+        return {}

pyFTS/models/nonstationary/util.py

@@ -54,23 +54,23 @@ def plot_sets(partitioner, start=0, end=10, step=1, tam=[5, 5], colors=None,
     Util.show_and_save_image(fig, file, save)
 
 
-def plot_sets_conditional(model, data, start=0, end=10, step=1, tam=[5, 5], colors=None,
+def plot_sets_conditional(model, data, step=1, size=[5, 5], colors=None,
                           save=False, file=None, axes=None):
-
+    range = np.arange(0, len(data), step)
-    range = np.arange(start,end,step)
     ticks = []
     if axes is None:
-        fig, axes = plt.subplots(nrows=1, ncols=1, figsize=tam)
+        fig, axes = plt.subplots(nrows=1, ncols=1, figsize=size)
 
-    for ct, key in enumerate(model.partitioner.ordered_sets):
+    for t in range:
-        set = model.partitioner.sets[key]
+        perturb = model.perturbation_factors(data[t])
-        for t in range:
+
-            tdisp = model.perturbation_factors(data[t])
+        for ct, key in enumerate(model.partitioner.ordered_sets):
-            set.perturbate_parameters(tdisp[ct])
+            set = model.partitioner.sets[key]
-            param = set.perturbated_parameters[str(tdisp[ct])]
+            set.perturbate_parameters(perturb[ct])
+            param = set.perturbated_parameters[str(perturb[ct])]
 
             if set.mf == Membership.trimf:
-                if t == start:
+                if t == 0:
                     line = axes.plot([t, t+1, t], param, label=set.name)
                     set.metadata['color'] = line[0].get_color()
                 else:
@@ -86,7 +86,7 @@ def plot_sets_conditional(model, data, start=0, end=10, step=1, tam=[5, 5], colo
     lgd = axes.legend(handles0, labels0, loc=2, bbox_to_anchor=(1, 1))
 
     if data is not None:
-        axes.plot(np.arange(start, start + len(data), 1), data,c="black")
+        axes.plot(np.arange(0, len(data), 1), data,c="black")
 
     plt.tight_layout()
 

pyFTS/tests/general.py

@@ -13,20 +13,27 @@ tdiff = Transformations.Differential(1)
 
 from pyFTS.data import TAIEX, SP500, NASDAQ
 
-#dataset = TAIEX.get_data()
+dataset = TAIEX.get_data()
-dataset = SP500.get_data()[11500:16000]
+#dataset = SP500.get_data()[11500:16000]
 #dataset = NASDAQ.get_data()
 #print(len(dataset))
-'''
+
 from pyFTS.partitioners import Grid, Util as pUtil
-partitioner = Grid.GridPartitioner(data=dataset[:800], npart=10) #, transformation=tdiff)
+partitioner = Grid.GridPartitioner(data=dataset[:800], npart=10, transformation=tdiff)
-'''
+
+
+from pyFTS.common import Util as cUtil
 from pyFTS.benchmarks import benchmarks as bchmk, Util as bUtil, Measures, knn, quantreg, arima, naive
 
-
+from pyFTS.models import pwfts, song, chen, ifts, hofts
-from pyFTS.models import pwfts, song, ifts, hofts
 from pyFTS.models.ensemble import ensemble
 
+model = chen.ConventionalFTS(partitioner=partitioner)
+model.append_transformation(tdiff)
+model.fit(dataset[:800])
+
+cUtil.plot_rules(model)
+
 '''
 model = knn.KNearestNeighbors(order=3)
 #model = ensemble.AllMethodEnsembleFTS("", partitioner=partitioner)
@@ -78,17 +85,16 @@ print(Measures.get_distribution_statistics(dataset[800:1000], model, steps_ahead
 #for tmp2 in tmp:
 #    print(tmp2)
 '''
-#'''
+'''
 
 types = ['point','interval','distribution']
 benchmark_methods=[[arima.ARIMA for k in range(8)] + [quantreg.QuantileRegression for k in range(4)]]
-'''
 benchmark_methods=[
     [arima.ARIMA for k in range(4)] + [naive.Naive],
     [arima.ARIMA for k in range(8)] + [quantreg.QuantileRegression for k in range(4)],
     [arima.ARIMA for k in range(4)] + [quantreg.QuantileRegression for k in range(2)]
     + [knn.KNearestNeighbors for k in range(3)]
-    ]'''
+    ]
 benchmark_methods_parameters= [
     [
         {'order': (1, 0, 0), 'alpha': .05},
@@ -105,7 +111,7 @@ benchmark_methods_parameters= [
         {'order': 2, 'alpha': .25}
     ]
 ]
-'''benchmark_methods_parameters= [
+benchmark_methods_parameters= [
     [
         {'order': (1, 0, 0)},
         {'order': (1, 0, 1)},
@@ -134,7 +140,7 @@ benchmark_methods_parameters= [
         {'order': 2, 'dist': True},
         {'order': 1}, {'order': 2}, {'order': 3},
     ]
-]'''
+]
 dataset_name = "SP500"
 tag = "ahead2"
 
@@ -169,7 +175,7 @@ for ct, type in enumerate(types):
                                     file="benchmarks.db", dataset=dataset_name, tag=tag)
 
 
-#'''
+'''
 '''
 dat = pd.read_csv('pwfts_taiex_partitioning.csv', sep=';')
 print(bUtil.analytic_tabular_dataframe(dat))

pyFTS/tests/nonstationary.py

@@ -1,13 +1,14 @@
 import os
 import numpy as np
 from pyFTS.common import Membership, Transformations
-from pyFTS.models.nonstationary import common, perturbation, partitioners, util, honsfts, cvfts
+from pyFTS.models.nonstationary import common, perturbation, partitioners, util
-from pyFTS.models.nonstationary import nsfts
+from pyFTS.models.nonstationary import nsfts, cvfts
 from pyFTS.partitioners import Grid
 import matplotlib.pyplot as plt
 from pyFTS.common import Util as cUtil
 import pandas as pd
 
+'''
 from pyFTS.data import artificial
 
 lmv1 = artificial.generate_gaussian_linear(1,0.2,0.2,0.05)
@@ -31,4 +32,36 @@ print(nsfts1.predict(test1))
 
 print(nsfts1)
 
-util.plot_sets(fs1, tam=[10, 5], start=0, end=100, step=2, data=lmv1[:100], window_size=35)
+util.plot_sets(fs1, tam=[10, 5], start=0, end=100, step=2, data=lmv1[:100], window_size=35)
+'''
+
+
+from pyFTS.common import Transformations
+tdiff = Transformations.Differential(1)
+
+
+from pyFTS.common import Util
+
+from pyFTS.data import TAIEX
+
+taiex = TAIEX.get_data()
+taiex_diff = tdiff.apply(taiex)
+
+train = taiex_diff[:600]
+test = taiex_diff[600:1500]
+
+fs_tmp = Grid.GridPartitioner(data=train, npart=20) #, transformation=tdiff)
+
+fs = partitioners.SimpleNonStationaryPartitioner(train, fs_tmp)
+
+print(fs)
+
+model = cvfts.ConditionalVarianceFTS(partitioner=fs)
+model.fit(train)
+
+print(model)
+
+#tmpp4 = model.predict(test, type='point')
+tmp = model.predict(test, type='interval')
+
+#util.plot_sets_conditional(model, test,  step=1, tam=[10, 5])