A new method for fuzzy frequency count at PFTS was added but in benchmarks the previous method proved to be most efficient. The new method was keep for future investigations.

benchmarks/ProbabilityDistribution.py

@@ -71,3 +71,12 @@ class ProbabilityDistribution(object):
 
         axis.set_xlabel('Universe of Discourse')
         axis.set_ylabel('Probability')
+
+
+    def __str__(self):
+        head = '|'
+        body = '|'
+        for k in sorted(self.distribution.keys()):
+            head += str(round(k,2)) + '\t|'
+            body += str(round(self.distribution[k]  / self.count,3)) + '\t|'
+        return head + '\n' + body

common/FuzzySet.py

@@ -28,6 +28,14 @@ def fuzzyInstance(inst, fuzzySets):
     return mv
 
 
+def fuzzyInstances(data, fuzzySets):
+    ret = []
+    for inst in data:
+        mv = np.array([fs.membership(inst) for fs in fuzzySets])
+        ret.append(mv)
+    return ret
+
+
 def getMaxMembershipFuzzySet(inst, fuzzySets):
     mv = fuzzyInstance(inst, fuzzySets)
     return fuzzySets[np.argwhere(mv == max(mv))[0, 0]]

pfts.py

@@ -5,7 +5,7 @@ import numpy as np
 import pandas as pd
 import math
 from operator import itemgetter
-from pyFTS.common import FuzzySet, SortedCollection
+from pyFTS.common import FLR, FuzzySet, SortedCollection
 from pyFTS import hofts, ifts, tree
 
 
@@ -22,6 +22,13 @@ class ProbabilisticFLRG(hofts.HighOrderFLRG):
         else:
             self.RHS[c.name] = 1.0
 
+    def appendRHSFuzzy(self, c, mv):
+        self.frequencyCount += mv
+        if c.name in self.RHS:
+            self.RHS[c.name] += mv
+        else:
+            self.RHS[c.name] = mv
+
     def getProbability(self, c):
         return self.RHS[c] / self.frequencyCount
 
@@ -47,6 +54,63 @@ class ProbabilisticFTS(ifts.IntervalFTS):
         self.hasDistributionForecasting = True
         self.isHighOrder = True
 
+    def train(self, data, sets, order=1,parameters=None):
+
+        data = self.doTransformations(data, updateUoD=True)
+
+        self.order = order
+        self.sets = sets
+        for s in self.sets:    self.setsDict[s.name] = s
+        tmpdata = FuzzySet.fuzzySeries(data, sets)
+        flrs = FLR.generateRecurrentFLRs(tmpdata)
+        self.flrgs = self.generateFLRG(flrs)
+        #self.flrgs = self.generateFLRG2(data)
+
+    def generateFLRG2(self, data):
+        flrgs = {}
+        l = len(data)
+        for k in np.arange(self.order, l):
+            if self.dump: print("FLR: " + str(k))
+            flrg = ProbabilisticFLRG(self.order)
+
+            sample = data[k - self.order: k]
+
+            mvs = FuzzySet.fuzzyInstances(sample, self.sets)
+            lags = {}
+
+            for o in np.arange(0, self.order):
+                _sets = [self.sets[kk] for kk in np.arange(0, len(self.sets)) if mvs[o][kk] > 0]
+
+                lags[o] = _sets
+
+            root = tree.FLRGTreeNode(None)
+
+            self.buildTreeWithoutOrder(root, lags, 0)
+
+            # Trace the possible paths
+            for p in root.paths():
+                path = list(reversed(list(filter(None.__ne__, p))))
+
+                lhs_mv = []
+                for c, e in enumerate(path, start=0):
+                    lhs_mv.append(  e.membership( sample[c] )  )
+                    flrg.appendLHS(e)
+
+                if flrg.strLHS() not in flrgs:
+                    flrgs[flrg.strLHS()] = flrg;
+
+                mv = FuzzySet.fuzzyInstance(data[k], self.sets)
+
+                rhs_mv = [mv[kk] for kk in np.arange(0, len(self.sets)) if mv[kk] > 0]
+                _sets = [self.sets[kk] for kk in np.arange(0, len(self.sets)) if mv[kk] > 0]
+
+                for c, e in enumerate(_sets, start=0):
+                    flrgs[flrg.strLHS()].appendRHSFuzzy(e,rhs_mv[c]*max(lhs_mv))
+
+                self.globalFrequency += max(lhs_mv)
+
+        return (flrgs)
+
     def generateFLRG(self, flrs):
         flrgs = {}
         l = len(flrs)

tests/pfts.py

@@ -16,8 +16,8 @@ from pyFTS.benchmarks import benchmarks as bchmk
 from pyFTS.benchmarks import Measures
 from numpy import random
 
-gauss_treino = random.normal(0,1.0,1600)
-gauss_teste = random.normal(0,1.0,400)
+#gauss_treino = random.normal(0,1.0,1600)
+#gauss_teste = random.normal(0,1.0,400)
 
 
 os.chdir("/home/petronio/dados/Dropbox/Doutorado/Disciplinas/AdvancedFuzzyTimeSeriesModels/")
@@ -25,9 +25,9 @@ os.chdir("/home/petronio/dados/Dropbox/Doutorado/Disciplinas/AdvancedFuzzyTimeSe
 #enrollments = pd.read_csv("DataSets/Enrollments.csv", sep=";")
 #enrollments = np.array(enrollments["Enrollments"])
 
-#taiex = pd.read_csv("DataSets/TAIEX.csv", sep=",")
-#taiex_treino = np.array(taiex["avg"][2500:3900])
-#taiex_teste = np.array(taiex["avg"][3901:4500])
+taiex = pd.read_csv("DataSets/TAIEX.csv", sep=",")
+taiex_treino = np.array(taiex["avg"][2500:3900])
+taiex_teste = np.array(taiex["avg"][3901:4500])
 
 #nasdaq = pd.read_csv("DataSets/NASDAQ_IXIC.csv", sep=",")
 #nasdaq_treino = np.array(nasdaq["avg"][0:1600])
@@ -35,16 +35,25 @@ os.chdir("/home/petronio/dados/Dropbox/Doutorado/Disciplinas/AdvancedFuzzyTimeSe
 
 diff = Transformations.Differential(1)
 
-fs = Grid.GridPartitionerTrimf(gauss_treino,10)
+fs = Grid.GridPartitionerTrimf(taiex_treino,10)
 
 #tmp = chen.ConventionalFTS("")
 
-#pfts1 = pfts.ProbabilisticFTS("1")
+pfts1 = pfts.ProbabilisticFTS("1")
 #pfts1.appendTransformation(diff)
-pfts1.train(gauss_treino,fs,1)
-pfts2 = pfts.ProbabilisticFTS("n = 2")
+pfts1.train(taiex_treino,fs,1)
+
+from pyFTS.benchmarks import ProbabilityDistribution as dist
+
+forecasts = pfts1.forecast(taiex_treino)
+
+pmf1 = dist.ProbabilityDistribution("Original",10,[min(taiex_treino),max(taiex_treino)],data=forecasts)
+
+print(pmf1)
+
+#pfts2 = pfts.ProbabilisticFTS("n = 2")
 #pfts2.appendTransformation(diff)
-pfts2.train(gauss_treino,fs,2)
+#pfts2.train(gauss_treino,fs,2)
 
 #pfts3 = pfts.ProbabilisticFTS("n = 3")
 #pfts3.appendTransformation(diff)