Minor code standadization on models; Improvements and bugfixes on benchmarks

pyFTS/benchmarks/Measures.py

@@ -205,14 +205,35 @@ def pinball_mean(tau, targets, forecasts):
     :param forecasts: list of prediction intervals
     :return: float, the pinball loss mean for tau quantile
     """
-    try:
-        if tau <= 0.5:
-            preds = [pinball(tau, targets[i], forecasts[i][0]) for i in np.arange(0, len(forecasts))]
-        else:
-            preds = [pinball(tau, targets[i], forecasts[i][1]) for i in np.arange(0, len(forecasts))]
-        return np.nanmean(preds)
-    except Exception as ex:
-        print(ex)
+    if tau <= 0.5:
+        preds = [pinball(tau, targets[i], forecasts[i][0]) for i in np.arange(0, len(forecasts))]
+    else:
+        preds = [pinball(tau, targets[i], forecasts[i][1]) for i in np.arange(0, len(forecasts))]
+    return np.nanmean(preds)
+
+
+def winkler_score(tau, target, forecast):
+    '''R. L. Winkler, A Decision-Theoretic Approach to Interval Estimation, J. Am. Stat. Assoc. 67 (337) (1972) 187–191. doi:10.2307/2284720. '''
+    delta = forecast[1] - forecast[0]
+    if forecast[0] < target and target < forecast[1]:
+        return delta
+    elif forecast[0] > target:
+        return delta + 2*(forecast[0] - target)/tau
+    elif forecast[1] < target:
+        return delta + 2*(target - forecast[1])/tau
+
+
+def winkler_mean(tau, targets, forecasts):
+    """
+    Mean Winkler score value of the forecast for a given tau-quantile of the targets
+    :param tau: quantile value in the range (0,1)
+    :param targets: list of target values
+    :param forecasts: list of prediction intervals
+    :return: float, the Winkler score mean for tau quantile
+    """
+    preds = [winkler_score(tau, targets[i], forecasts[i]) for i in np.arange(0, len(forecasts))]
+
+    return np.nanmean(preds)
 
 
 def brier_score(targets, densities):
@@ -348,6 +369,8 @@ def get_interval_statistics(data, model, **kwargs):
         ret.append(round(pinball_mean(0.25, data[model.order:], forecasts[:-1]), 2))
         ret.append(round(pinball_mean(0.75, data[model.order:], forecasts[:-1]), 2))
         ret.append(round(pinball_mean(0.95, data[model.order:], forecasts[:-1]), 2))
+        ret.append(round(winkler_mean(0.05, data[model.order:], forecasts[:-1]), 2))
+        ret.append(round(winkler_mean(0.25, data[model.order:], forecasts[:-1]), 2))
     else:
         forecasts = []
         for k in np.arange(model.order, len(data) - steps_ahead):
@@ -363,6 +386,8 @@ def get_interval_statistics(data, model, **kwargs):
         ret.append(round(pinball_mean(0.25, data[start:], forecasts), 2))
         ret.append(round(pinball_mean(0.75, data[start:], forecasts), 2))
         ret.append(round(pinball_mean(0.95, data[start:], forecasts), 2))
+        ret.append(round(winkler_mean(0.05, data[start:], forecasts), 2))
+        ret.append(round(winkler_mean(0.25, data[start:], forecasts), 2))
     return ret
 
 

pyFTS/benchmarks/Util.py

@@ -18,6 +18,11 @@ from pyFTS.common import Util
 
 def open_benchmark_db(name):
     conn = sqlite3.connect(name)
+
+    #performance optimizations
+    conn.execute("PRAGMA journal_mode = WAL")
+    conn.execute("PRAGMA synchronous = NORMAL")
+
     create_benchmark_tables(conn)
     return conn
 
@@ -31,7 +36,6 @@ def create_benchmark_tables(conn):
                  Scheme text, Partitions int,
                  Size int, Steps int, Method text, Measure text, Value real)''')
 
-    # Save (commit) the changes
     conn.commit()
 
 

pyFTS/benchmarks/arima.py

@@ -13,8 +13,8 @@ class ARIMA(fts.FTS):
     """
     Façade for statsmodels.tsa.arima_model
     """
-    def __init__(self, name, **kwargs):
-        super(ARIMA, self).__init__(1, "ARIMA"+name)
+    def __init__(self, **kwargs):
+        super(ARIMA, self).__init__(**kwargs)
         self.name = "ARIMA"
         self.detail = "Auto Regressive Integrated Moving Average"
         self.is_high_order = True
@@ -44,11 +44,8 @@ class ARIMA(fts.FTS):
 
     def train(self, data, **kwargs):
 
-        self.original_min = np.nanmin(data)
-        self.original_max = np.nanmax(data)
-
-        if kwargs.get('order', None) is not None:
-            order = kwargs.get('order', (1,0,0))
+        if 'order' in kwargs:
+            order = kwargs.pop('order')
             self._decompose_order(order)
 
         if self.indexer is not None:

pyFTS/benchmarks/benchmarks.py

@@ -156,12 +156,12 @@ def sliding_window_benchmarks(data, windowsize, train=0.8, **kwargs):
 
     if models is None:
         for method in methods:
-            mfts = method("")
+            mfts = method()
 
             if mfts.is_high_order:
                 for order in orders:
                     if order >= mfts.min_order:
-                        mfts = method("")
+                        mfts = method()
                         mfts.order = order
                         pool.append(mfts)
             else:
@@ -190,7 +190,7 @@ def sliding_window_benchmarks(data, windowsize, train=0.8, **kwargs):
             for transformation in transformations:
                 for count, model in enumerate(benchmark_methods, start=0):
                     par = benchmark_methods_parameters[count]
-                    mfts = model("", **par)
+                    mfts = model(**par)
                     mfts.append_transformation(transformation)
                     benchmark_pool.append(mfts)
 
@@ -203,6 +203,8 @@ def sliding_window_benchmarks(data, windowsize, train=0.8, **kwargs):
     elif type == 'distribution':
         experiment_method = run_probabilistic
         synthesis_method = process_probabilistic_jobs
+    else:
+        raise ValueError("Type parameter has a unkown value!")
 
     if distributed:
         import dispy, dispy.httpd
@@ -213,28 +215,29 @@ def sliding_window_benchmarks(data, windowsize, train=0.8, **kwargs):
     experiments = 0
     jobs = []
 
+    inc = __pop("inc", 0.1, kwargs)
+
     if progress:
         from tqdm import tqdm
-        progressbar = tqdm(total=len(data), desc="Sliding Window:")
-
-    inc = __pop("inc", 0.1, kwargs)
+        _tdata = len(data) / (windowsize * inc)
+        _tasks = (len(partitioners_models) * len(orders) * len(partitions) * len(transformations) * len(steps_ahead))
+        _tbcmk = len(benchmark_pool)*len(steps_ahead)
+        progressbar = tqdm(total=_tdata*_tasks + _tdata*_tbcmk, desc="Benchmarks:")
 
     file = kwargs.get('file', "benchmarks.db")
 
     conn = bUtil.open_benchmark_db(file)
 
     for ct, train, test in cUtil.sliding_window(data, windowsize, train, inc=inc, **kwargs):
-        experiments += 1
-
-        if progress:
-            progressbar.update(windowsize * inc)
-
         if benchmark_models != False:
             for model in benchmark_pool:
                 for step in steps_ahead:
+
                     kwargs['steps_ahead'] = step
 
                     if not distributed:
+                        if progress:
+                            progressbar.update(1)
                         job = experiment_method(deepcopy(model), None, train, test, **kwargs)
                         synthesis_method(dataset, tag, job, conn)
                     else:
@@ -257,28 +260,17 @@ def sliding_window_benchmarks(data, windowsize, train=0.8, **kwargs):
         else:
             partitioners_pool = partitioners_models
 
-        rng1 = steps_ahead
-        if progress:
-            rng1 = tqdm(steps_ahead, desc="Steps")
+        for step in steps_ahead:
 
-        for step in rng1:
-            rng2 = partitioners_pool
+            for partitioner in partitioners_pool:
 
-            if progress:
-                rng2 = tqdm(partitioners_pool, desc="Partitioners")
-
-            for partitioner in rng2:
-
-                rng3 = enumerate(pool,start=0)
-
-                if progress:
-                    rng3 = enumerate(tqdm(pool, desc="Models"),start=0)
-
-                for _id, model in rng3:
+                for _id, model in enumerate(pool,start=0):
 
                     kwargs['steps_ahead'] = step
 
                     if not distributed:
+                        if progress:
+                            progressbar.update(1)
                         job = experiment_method(deepcopy(model), deepcopy(partitioner), train, test, **kwargs)
                         synthesis_method(dataset, tag, job, conn)
                     else:
@@ -291,12 +283,9 @@ def sliding_window_benchmarks(data, windowsize, train=0.8, **kwargs):
 
     if distributed:
 
-        rng = jobs
-
-        if progress:
-            rng = tqdm(jobs)
-
-        for job in rng:
+        for job in jobs:
+            if progress:
+                progressbar.update(1)
             job()
             if job.status == dispy.DispyJob.Finished and job is not None:
                 tmp = job.result
@@ -424,13 +413,15 @@ def run_interval(mfts, partitioner, train_data, test_data, window_key=None, **kw
     times = _end - _start
 
     _start = time.time()
-    _sharp, _res, _cov, _q05, _q25, _q75, _q95 = Measures.get_interval_statistics(test_data, mfts, **kwargs)
+    #_sharp, _res, _cov, _q05, _q25, _q75, _q95, _w05, _w25
+    metrics = Measures.get_interval_statistics(test_data, mfts, **kwargs)
     _end = time.time()
     times += _end - _start
 
-    ret = {'key': _key, 'obj': mfts, 'sharpness': _sharp, 'resolution': _res, 'coverage': _cov, 'time': times,
-           'Q05': _q05, 'Q25': _q25, 'Q75': _q75, 'Q95': _q95, 'window': window_key,
-           'steps': steps_ahead, 'method': method}
+    ret = {'key': _key, 'obj': mfts, 'sharpness': metrics[0], 'resolution': metrics[1], 'coverage': metrics[2],
+           'time': times,'Q05': metrics[3], 'Q25': metrics[4], 'Q75': metrics[5], 'Q95': metrics[6],
+           'winkler05': metrics[7], 'winkler25': metrics[8],
+           'window': window_key,'steps': steps_ahead, 'method': method}
 
     return ret
 
@@ -543,6 +534,12 @@ def process_interval_jobs(dataset, tag, job, conn):
     Q95 = deepcopy(data)
     Q95.extend(["Q95", job["Q95"]])
     bUtil.insert_benchmark(Q95, conn)
+    W05 = deepcopy(data)
+    W05.extend(["winkler05", job["winkler05"]])
+    bUtil.insert_benchmark(W05, conn)
+    W25 = deepcopy(data)
+    W25.extend(["winkler25", job["winkler25"]])
+    bUtil.insert_benchmark(W25, conn)
 
 
 def process_probabilistic_jobs(dataset, tag,  job, conn):

pyFTS/benchmarks/knn.py

@@ -11,8 +11,8 @@ class KNearestNeighbors(fts.FTS):
     """
     K-Nearest Neighbors
     """
-    def __init__(self, name, **kwargs):
-        super(KNearestNeighbors, self).__init__(1, "kNN"+name)
+    def __init__(self, **kwargs):
+        super(KNearestNeighbors, self).__init__(**kwargs)
         self.name = "kNN"
         self.shortname = "kNN"
         self.detail = "K-Nearest Neighbors"
@@ -23,20 +23,12 @@ class KNearestNeighbors(fts.FTS):
         self.benchmark_only = True
         self.min_order = 1
         self.alpha = kwargs.get("alpha", 0.05)
-        self.order = kwargs.get("order", 1)
         self.lag = None
         self.k = kwargs.get("k", 30)
         self.uod = None
 
     def train(self, data, **kwargs):
-        if kwargs.get('order', None) is not None:
-            self.order = kwargs.get('order', 1)
-
         self.data = np.array(data)
-        self.original_max = max(data)
-        self.original_min = min(data)
-
-        #self.lagdata, = lagmat(data, maxlag=self.order, trim="both", original='sep')
 
 
     def knn(self, sample):

pyFTS/benchmarks/naive.py

@@ -6,8 +6,8 @@ from pyFTS.common import fts
 
 class Naive(fts.FTS):
     """Naïve Forecasting method"""
-    def __init__(self, name, **kwargs):
-        super(Naive, self).__init__(1, "Naive")
+    def __init__(self, **kwargs):
+        super(Naive, self).__init__(order=1, name="Naive",**kwargs)
         self.name = "Naïve Model"
         self.detail = "Naïve Model"
         self.benchmark_only = True

pyFTS/benchmarks/quantreg.py

@@ -11,8 +11,8 @@ from pyFTS.probabilistic import ProbabilityDistribution
 
 class QuantileRegression(fts.FTS):
     """Façade for statsmodels.regression.quantile_regression"""
-    def __init__(self, name, **kwargs):
-        super(QuantileRegression, self).__init__(1, "")
+    def __init__(self, **kwargs):
+        super(QuantileRegression, self).__init__(**kwargs)
         self.name = "QR"
         self.detail = "Quantile Regression"
         self.is_high_order = True
@@ -27,13 +27,8 @@ class QuantileRegression(fts.FTS):
         self.mean_qt = None
         self.lower_qt = None
         self.dist_qt = None
-        self.order = kwargs.get('order', 1)
-        self.shortname = "QAR("+str(self.order)+","+str(self.alpha)+")"
 
     def train(self, data, **kwargs):
-        if 'order' in kwargs:
-            self.order = kwargs.get('order', 1)
-
         if self.indexer is not None and isinstance(data, pd.DataFrame):
             data = self.indexer.get_data(data)
 
@@ -58,9 +53,6 @@ class QuantileRegression(fts.FTS):
                 up_qt = [k for k in uqt.params]
                 self.dist_qt.append([lo_qt, up_qt])
 
-        self.original_min = min(data)
-        self.original_max = max(data)
-
         self.shortname = "QAR(" + str(self.order) + ") - " + str(self.alpha)
 
     def linearmodel(self,data,params):

pyFTS/common/flrg.py

@@ -38,11 +38,11 @@ class FLRG(object):
                 self.key = self.key + n
         return self.key
 
-
     def get_membership(self, data, sets):
         ret = 0.0
         if isinstance(self.LHS, (list, set)):
-            ret = np.nanmin([sets[self.LHS[ct]].membership(dat) for ct, dat in enumerate(data)])
+            if len(self.LHS) == len(data):
+                ret = np.nanmin([sets[self.LHS[ct]].membership(dat) for ct, dat in enumerate(data)])
         else:
             ret = sets[self.LHS].membership(data)
         return ret

pyFTS/common/fts.py

@@ -7,7 +7,7 @@ class FTS(object):
     """
     Fuzzy Time Series object model
     """
-    def __init__(self, order, name, **kwargs):
+    def __init__(self, **kwargs):
         """
         Create a Fuzzy Time Series model
         :param order: model order
@@ -16,10 +16,10 @@ class FTS(object):
         """
         self.sets = {}
         self.flrgs = {}
-        self.order = order
-        self.shortname = name
-        self.name = name
-        self.detail = name
+        self.order = kwargs.get('order',"")
+        self.shortname = kwargs.get('name',"")
+        self.name = kwargs.get('name',"")
+        self.detail = kwargs.get('name',"")
         self.is_high_order = False
         self.min_order = 1
         self.has_seasonality = False
@@ -75,6 +75,8 @@ class FTS(object):
         else:
             ndata = self.apply_transformations(data)
 
+        ndata = np.clip(ndata, self.original_min, self.original_max)
+
         if 'distributed' in kwargs:
             distributed = kwargs.pop('distributed')
         else:
@@ -222,6 +224,24 @@ class FTS(object):
         else:
             data = self.apply_transformations(ndata)
 
+        self.original_min = np.nanmin(data)
+        self.original_max = np.nanmax(data)
+
+        if 'sets' in kwargs:
+            self.sets = kwargs.pop('sets')
+
+        if 'partitioner' in kwargs:
+            self.partitioner = kwargs.pop('partitioner')
+
+        if (self.sets is None or len(self.sets) == 0) and not self.benchmark_only:
+            if self.partitioner is not None:
+                self.sets = self.partitioner.sets
+            else:
+                raise Exception("Fuzzy sets were not provided for the model. Use 'sets' parameter or 'partitioner'. ")
+
+        if 'order' in kwargs:
+            self.order = kwargs.pop('order')
+
         dump = kwargs.get('dump', None)
 
         num_batches = kwargs.get('num_batches', None)

pyFTS/data/NASDAQ.py

@@ -3,7 +3,7 @@ import pandas as pd
 import numpy as np
 
 
-def get_data(field):
+def get_data(field="avg"):
     """
     Get a simple univariate time series data.
     :param field: the dataset field name to extract
@@ -21,6 +21,6 @@ def get_dataframe():
     """
     dat = common.get_dataframe('NASDAQ.csv.bz2',
                                'https://github.com/petroniocandido/pyFTS/raw/8f20f3634aa6a8f58083bdcd1bbf93795e6ed767/pyFTS/data/NASDAQ.csv.bz2',
-                               sep=";", compression='bz2')
+                               sep=",", compression='bz2')
     return dat
 

pyFTS/models/chen.py

@@ -33,10 +33,11 @@ class ConventionalFLRG(flrg.FLRG):
 
 class ConventionalFTS(fts.FTS):
     """Conventional Fuzzy Time Series"""
-    def __init__(self, name, **kwargs):
-        super(ConventionalFTS, self).__init__(1, "CFTS " + name, **kwargs)
+    def __init__(self, **kwargs):
+        super(ConventionalFTS, self).__init__(order=1, **kwargs)
         self.name = "Conventional FTS"
         self.detail = "Chen"
+        self.shortname = "CFTS"
         self.flrgs = {}
 
     def generate_flrg(self, flrs):
@@ -48,10 +49,6 @@ class ConventionalFTS(fts.FTS):
                 self.flrgs[flr.LHS].append_rhs(flr.RHS)
 
     def train(self, data, **kwargs):
-        if kwargs.get('sets', None) is not None:
-            self.sets = kwargs.get('sets', None)
-        else:
-            self.sets = self.partitioner.sets
 
         tmpdata = FuzzySet.fuzzyfy_series(data, self.sets, method='maximum')
         flrs = FLR.generate_non_recurrent_flrs(tmpdata)

pyFTS/models/cheng.py

@@ -45,9 +45,9 @@ class TrendWeightedFLRG(yu.WeightedFLRG):
 
 class TrendWeightedFTS(yu.WeightedFTS):
     """First Order Trend Weighted Fuzzy Time Series"""
-    def __init__(self, name, **kwargs):
-        super(TrendWeightedFTS, self).__init__("", **kwargs)
-        self.shortname = "TWFTS " + name
+    def __init__(self, **kwargs):
+        super(TrendWeightedFTS, self).__init__(**kwargs)
+        self.shortname = "TWFTS"
         self.name = "Trend Weighted FTS"
         self.detail = "Cheng"
         self.is_high_order = False

pyFTS/models/ensemble/ensemble.py

@@ -17,9 +17,9 @@ def sampler(data, quantiles):
 
 
 class EnsembleFTS(fts.FTS):
-    def __init__(self, name, **kwargs):
-        super(EnsembleFTS, self).__init__(1, "Ensemble FTS", **kwargs)
-        self.shortname = "Ensemble FTS " + name
+    def __init__(self, **kwargs):
+        super(EnsembleFTS, self).__init__(**kwargs)
+        self.shortname = "Ensemble FTS"
         self.name = "Ensemble FTS"
         self.flrgs = {}
         self.has_point_forecasting = True
@@ -29,7 +29,6 @@ class EnsembleFTS(fts.FTS):
         self.models = []
         self.parameters = []
         self.alpha = kwargs.get("alpha", 0.05)
-        self.order = 1
         self.point_method = kwargs.get('point_method', 'mean')
         self.interval_method = kwargs.get('interval_method', 'quantile')
 
@@ -39,8 +38,7 @@ class EnsembleFTS(fts.FTS):
             self.order = model.order
 
     def train(self, data, **kwargs):
-        self.original_max = max(data)
-        self.original_min = min(data)
+        pass
 
     def get_models_forecasts(self,data):
         tmp = []
@@ -246,8 +244,8 @@ class EnsembleFTS(fts.FTS):
 
 
 class AllMethodEnsembleFTS(EnsembleFTS):
-    def __init__(self, name, **kwargs):
-        super(AllMethodEnsembleFTS, self).__init__(name="Ensemble FTS"+name, **kwargs)
+    def __init__(self, **kwargs):
+        super(AllMethodEnsembleFTS, self).__init__(**kwargs)
         self.min_order = 3
         self.shortname ="Ensemble FTS"
 
@@ -256,26 +254,22 @@ class AllMethodEnsembleFTS(EnsembleFTS):
             model.append_transformation(t)
 
     def train(self, data, **kwargs):
-        self.original_max = max(data)
-        self.original_min = min(data)
-
-        order = kwargs.get('order',2)
-
         fo_methods = [song.ConventionalFTS, chen.ConventionalFTS, yu.WeightedFTS, cheng.TrendWeightedFTS,
                       sadaei.ExponentialyWeightedFTS, ismailefendi.ImprovedWeightedFTS]
 
         ho_methods = [hofts.HighOrderFTS, hwang.HighOrderFTS]
 
         for method in fo_methods:
-            model = method("", partitioner=self.partitioner)
+            model = method(partitioner=self.partitioner)
             self.set_transformations(model)
             model.fit(data, **kwargs)
             self.append_model(model)
 
         for method in ho_methods:
-            for o in np.arange(1, order+1):
-                model = method("", partitioner=self.partitioner)
+            for o in np.arange(1, self.order+1):
+                model = method(partitioner=self.partitioner)
                 if model.min_order >= o:
+                    model.order = o
                     self.set_transformations(model)
                     model.fit(data, **kwargs)
                     self.append_model(model)

pyFTS/models/hofts.py

@@ -38,13 +38,11 @@ class HighOrderFLRG(flrg.FLRG):
 
 class HighOrderFTS(fts.FTS):
     """Conventional High Order Fuzzy Time Series"""
-    def __init__(self, name, **kwargs):
-        super(HighOrderFTS, self).__init__(1, name="HOFTS" + name, **kwargs)
+    def __init__(self, **kwargs):
+        super(HighOrderFTS, self).__init__(**kwargs)
         self.name = "High Order FTS"
-        self.shortname = "HOFTS" + name
+        self.shortname = "HOFTS"
         self.detail = "Chen"
-        self.order = kwargs.get('order',1)
-        self.setsDict = {}
         self.is_high_order = True
         self.min_order = 2
 
@@ -94,13 +92,6 @@ class HighOrderFTS(fts.FTS):
 
     def train(self, data, **kwargs):
 
-        self.order = kwargs.get('order',2)
-
-        if kwargs.get('sets', None) is not None:
-            self.sets = kwargs.get('sets', None)
-        else:
-            self.sets = self.partitioner.sets
-
         self.generate_flrg(data)
 
     def forecast(self, ndata, **kwargs):
@@ -115,10 +106,12 @@ class HighOrderFTS(fts.FTS):
         for k in np.arange(self.order, l+1):
             flrgs = self.generate_lhs_flrg(ndata[k - self.order: k])
 
+            tmp = []
             for flrg in flrgs:
-                tmp = []
+
                 if flrg.get_key() not in self.flrgs:
-                    tmp.append(self.sets[flrg.LHS[-1]].centroid)
+                    if len(flrg.LHS) > 0:
+                        tmp.append(self.sets[flrg.LHS[-1]].centroid)
                 else:
                     flrg = self.flrgs[flrg.get_key()]
                     tmp.append(flrg.get_midpoint(self.sets))

pyFTS/models/hwang.py

@@ -10,12 +10,12 @@ from pyFTS.common import FuzzySet, FLR, Transformations, fts
 
 
 class HighOrderFTS(fts.FTS):
-    def __init__(self, name, **kwargs):
-        super(HighOrderFTS, self).__init__(1, name, **kwargs)
+    def __init__(self, **kwargs):
+        super(HighOrderFTS, self).__init__(**kwargs)
         self.is_high_order = True
         self.min_order = 2
         self.name = "Hwang High Order FTS"
-        self.shortname = "Hwang" + name
+        self.shortname = "Hwang"
         self.detail = "Hwang"
 
     def forecast(self, ndata, **kwargs):
@@ -57,9 +57,4 @@ class HighOrderFTS(fts.FTS):
         return ret
 
     def train(self, data, **kwargs):
-        if kwargs.get('sets', None) is not None:
-            self.sets = kwargs.get('sets', None)
-        else:
-            self.sets = self.partitioner.sets
-
-        self.order = kwargs.get('order', 2)
+        pass

pyFTS/models/ifts.py

@@ -17,9 +17,9 @@ class IntervalFTS(hofts.HighOrderFTS):
     """
     High Order Interval Fuzzy Time Series
     """
-    def __init__(self, name, **kwargs):
-        super(IntervalFTS, self).__init__(name="IFTS " + name, **kwargs)
-        self.shortname = "IFTS " + name
+    def __init__(self, **kwargs):
+        super(IntervalFTS, self).__init__(**kwargs)
+        self.shortname = "IFTS"
         self.name = "Interval FTS"
         self.detail = "Silva, P.; Guimarães, F.; Sadaei, H. (2016)"
         self.flrgs = {}
@@ -29,19 +29,23 @@ class IntervalFTS(hofts.HighOrderFTS):
         self.min_order = 1
 
     def get_upper(self, flrg):
-        if flrg.get_key() in self.flrgs:
-            tmp = self.flrgs[flrg.get_key()]
-            ret = tmp.get_upper(self.sets)
-        else:
-            ret = self.sets[flrg.LHS[-1]].upper
+        ret = np.nan
+        if len(flrg.LHS) > 0:
+            if flrg.get_key() in self.flrgs:
+                tmp = self.flrgs[flrg.get_key()]
+                ret = tmp.get_upper(self.sets)
+            else:
+                ret = self.sets[flrg.LHS[-1]].upper
         return ret
 
     def get_lower(self, flrg):
-        if flrg.get_key() in self.flrgs:
-            tmp = self.flrgs[flrg.get_key()]
-            ret = tmp.get_lower(self.sets)
-        else:
-            ret = self.sets[flrg.LHS[-1]].lower
+        ret = np.nan
+        if len(flrg.LHS) > 0:
+            if flrg.get_key() in self.flrgs:
+                tmp = self.flrgs[flrg.get_key()]
+                ret = tmp.get_lower(self.sets)
+            else:
+                ret = self.sets[flrg.LHS[-1]].lower
         return ret
 
     def get_sequence_membership(self, data, fuzzySets):
@@ -69,11 +73,12 @@ class IntervalFTS(hofts.HighOrderFTS):
             affected_flrgs_memberships = []
 
             for flrg in flrgs:
-                # achar o os bounds de cada FLRG, ponderados pela pertinência
-                mv = flrg.get_membership(sample, self.sets)
-                up.append(mv * self.get_upper(flrg))
-                lo.append(mv * self.get_lower(flrg))
-                affected_flrgs_memberships.append(mv)
+                if len(flrg.LHS) > 0:
+                    # achar o os bounds de cada FLRG, ponderados pela pertinência
+                    mv = flrg.get_membership(sample, self.sets)
+                    up.append(mv * self.get_upper(flrg))
+                    lo.append(mv * self.get_lower(flrg))
+                    affected_flrgs_memberships.append(mv)
 
             # gerar o intervalo
             norm = sum(affected_flrgs_memberships)

pyFTS/models/ismailefendi.py

@@ -47,8 +47,8 @@ class ImprovedWeightedFLRG(flrg.FLRG):
 
 class ImprovedWeightedFTS(fts.FTS):
     """First Order Improved Weighted Fuzzy Time Series"""
-    def __init__(self, name, **kwargs):
-        super(ImprovedWeightedFTS, self).__init__(1, "IWFTS " + name, **kwargs)
+    def __init__(self, **kwargs):
+        super(ImprovedWeightedFTS, self).__init__(order=1, name="IWFTS", **kwargs)
         self.name = "Improved Weighted FTS"
         self.detail = "Ismail & Efendi"
 
@@ -61,10 +61,6 @@ class ImprovedWeightedFTS(fts.FTS):
                 self.flrgs[flr.LHS].append_rhs(flr.RHS)
 
     def train(self, ndata, **kwargs):
-        if kwargs.get('sets', None) is not None:
-            self.sets = kwargs.get('sets', None)
-        else:
-            self.sets = self.partitioner.sets
 
         tmpdata = FuzzySet.fuzzyfy_series(ndata, self.sets, method='maximum')
         flrs = FLR.generate_recurrent_flrs(tmpdata)
@@ -73,7 +69,10 @@ class ImprovedWeightedFTS(fts.FTS):
     def forecast(self, ndata, **kwargs):
         l = 1
 
-        ordered_sets = FuzzySet.set_ordered(self.sets)
+        if self.partitioner is not None:
+            ordered_sets = self.partitioner.ordered_sets
+        else:
+            ordered_sets = FuzzySet.set_ordered(self.sets)
 
         ndata = np.array(ndata)
         l = len(ndata)

pyFTS/models/multivariate/mvfts.py

@@ -11,7 +11,7 @@ class MVFTS(fts.FTS):
     Multivariate extension of Chen's ConventionalFTS method
     """
     def __init__(self, name, **kwargs):
-        super(MVFTS, self).__init__(1, name, **kwargs)
+        super(MVFTS, self).__init__(order=1, name=name, **kwargs)
         self.explanatory_variables = []
         self.target_variable = None
         self.flrgs = {}
@@ -91,8 +91,6 @@ class MVFTS(fts.FTS):
 
         ndata = self.apply_transformations(data)
 
-        self.order = kwargs.get('order',1)
-
         flrs = self.generate_flrs(ndata)
         self.generate_flrg(flrs)
 

pyFTS/models/pwfts.py

@@ -94,9 +94,9 @@ class ProbabilisticWeightedFLRG(hofts.HighOrderFLRG):
 
 class ProbabilisticWeightedFTS(ifts.IntervalFTS):
     """High Order Probabilistic Weighted Fuzzy Time Series"""
-    def __init__(self, name, **kwargs):
-        super(ProbabilisticWeightedFTS, self).__init__(name=name, **kwargs)
-        self.shortname = "PWFTS " + name
+    def __init__(self, **kwargs):
+        super(ProbabilisticWeightedFTS, self).__init__(**kwargs)
+        self.shortname = "PWFTS"
         self.name = "Probabilistic FTS"
         self.detail = "Silva, P.; Guimarães, F.; Sadaei, H."
         self.flrgs = {}
@@ -108,22 +108,10 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
         self.min_order = 1
         self.auto_update = kwargs.get('update',False)
 
-
     def train(self, data, **kwargs):
 
-        data = self.apply_transformations(data, updateUoD=True)
-
         parameters = kwargs.get('parameters','fuzzy')
 
-        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.original_min = self.partitioner.min
-            self.original_max = self.partitioner.max
-        else:
-            self.sets = kwargs.get('sets',None)
-
         if parameters == 'monotonic':
             tmpdata = FuzzySet.fuzzyfy_series_old(data, self.sets)
             flrs = FLR.generate_recurrent_flrs(tmpdata)
@@ -237,7 +225,7 @@ class ProbabilisticWeightedFTS(ifts.IntervalFTS):
             # this may be the problem! TEST IT!!!
             ##########################################
             pi = 1 / len(flrg.LHS)
-            ret = sum(np.array([pi * self.setsDict[s].membership(x) for s in flrg.LHS]))
+            ret = sum(np.array([pi * self.sets[s].membership(x) for s in flrg.LHS]))
         return ret
 
     def get_upper(self, flrg):

pyFTS/models/sadaei.py

@@ -51,8 +51,8 @@ class ExponentialyWeightedFLRG(flrg.FLRG):
 
 class ExponentialyWeightedFTS(fts.FTS):
     """First Order Exponentialy Weighted Fuzzy Time Series"""
-    def __init__(self, name, **kwargs):
-        super(ExponentialyWeightedFTS, self).__init__(1, "EWFTS", **kwargs)
+    def __init__(self, **kwargs):
+        super(ExponentialyWeightedFTS, self).__init__(order=1, name="EWFTS", **kwargs)
         self.name = "Exponentialy Weighted FTS"
         self.detail = "Sadaei"
         self.c = kwargs.get('c', default_c)
@@ -66,12 +66,6 @@ class ExponentialyWeightedFTS(fts.FTS):
                 self.flrgs[flr.LHS].append_rhs(flr.RHS)
 
     def train(self, data, **kwargs):
-        self.c = kwargs.get('parameters', default_c)
-        if kwargs.get('sets', None) is not None:
-            self.sets = kwargs.get('sets', None)
-        else:
-            self.sets = self.partitioner.sets
-
         tmpdata = FuzzySet.fuzzyfy_series(data, self.sets, method='maximum')
         flrs = FLR.generate_recurrent_flrs(tmpdata)
         self.generate_flrg(flrs, self.c)
@@ -79,7 +73,10 @@ class ExponentialyWeightedFTS(fts.FTS):
     def forecast(self, ndata, **kwargs):
         l = 1
 
-        ordered_sets = FuzzySet.set_ordered(self.sets)
+        if self.partitioner is not None:
+            ordered_sets = self.partitioner.ordered_sets
+        else:
+            ordered_sets = FuzzySet.set_ordered(self.sets)
 
         data = np.array(ndata)
 

pyFTS/models/song.py

@@ -10,8 +10,8 @@ from pyFTS.common import FuzzySet, FLR, fts
 
 class ConventionalFTS(fts.FTS):
     """Traditional Fuzzy Time Series"""
-    def __init__(self, name, **kwargs):
-        super(ConventionalFTS, self).__init__(1, "FTS " + name, **kwargs)
+    def __init__(self, **kwargs):
+        super(ConventionalFTS, self).__init__(order=1, name="FTS", **kwargs)
         self.name = "Traditional FTS"
         self.detail = "Song & Chissom"
         if self.sets is not None and self.partitioner is not None:
@@ -49,10 +49,6 @@ class ConventionalFTS(fts.FTS):
 
 
     def train(self, data, **kwargs):
-        if kwargs.get('sets', None) is not None:
-            self.sets = kwargs.get('sets', None)
-        else:
-            self.sets = self.partitioner.sets
 
         tmpdata = FuzzySet.fuzzyfy_series(data, self.sets, method='maximum')
         flrs = FLR.generate_non_recurrent_flrs(tmpdata)
@@ -60,7 +56,10 @@ class ConventionalFTS(fts.FTS):
 
     def forecast(self, ndata, **kwargs):
 
-        ordered_set = FuzzySet.set_ordered(self.sets)
+        if self.partitioner is not None:
+            ordered_sets = self.partitioner.ordered_sets
+        else:
+            ordered_sets = FuzzySet.set_ordered(self.sets)
 
         l = len(ndata)
         npart = len(self.sets)
@@ -75,9 +74,9 @@ class ConventionalFTS(fts.FTS):
             fs = np.ravel(np.argwhere(r == max(r)))
 
             if len(fs) == 1:
-                ret.append(self.sets[ordered_set[fs[0]]].centroid)
+                ret.append(self.sets[ordered_sets[fs[0]]].centroid)
             else:
-                mp = [self.sets[ordered_set[s]].centroid for s in fs]
+                mp = [self.sets[ordered_sets[s]].centroid for s in fs]
 
                 ret.append( sum(mp)/len(mp))
 

pyFTS/models/yu.py

@@ -44,8 +44,8 @@ class WeightedFLRG(flrg.FLRG):
 
 class WeightedFTS(fts.FTS):
     """First Order Weighted Fuzzy Time Series"""
-    def __init__(self, name, **kwargs):
-        super(WeightedFTS, self).__init__(1, "WFTS " + name, **kwargs)
+    def __init__(self, **kwargs):
+        super(WeightedFTS, self).__init__(order=1, name="WFTS", **kwargs)
         self.name = "Weighted FTS"
         self.detail = "Yu"
 
@@ -58,18 +58,16 @@ class WeightedFTS(fts.FTS):
                 self.flrgs[flr.LHS].append_rhs(flr.RHS)
 
     def train(self, ndata, **kwargs):
-        if kwargs.get('sets', None) is not None:
-            self.sets = kwargs.get('sets', None)
-        else:
-            self.sets = self.partitioner.sets
-
         tmpdata = FuzzySet.fuzzyfy_series(ndata, self.sets, method='maximum')
         flrs = FLR.generate_recurrent_flrs(tmpdata)
         self.generate_FLRG(flrs)
 
     def forecast(self, ndata, **kwargs):
 
-        ordered_sets = FuzzySet.set_ordered(self.sets)
+        if self.partitioner is not None:
+            ordered_sets = self.partitioner.ordered_sets
+        else:
+            ordered_sets = FuzzySet.set_ordered(self.sets)
 
         ndata = np.array(ndata)
 

pyFTS/tests/general.py

@@ -11,35 +11,61 @@ from pyFTS.common import Transformations
 
 tdiff = Transformations.Differential(1)
 
-from pyFTS.data import TAIEX, SP500
+from pyFTS.data import TAIEX, SP500, NASDAQ
 
 #dataset = TAIEX.get_data()
 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)
 '''
-from pyFTS.benchmarks import benchmarks as bchmk, Util as bUtil, Measures, knn, quantreg, arima
+from pyFTS.benchmarks import benchmarks as bchmk, Util as bUtil, Measures, knn, quantreg, arima, naive
 
 
 from pyFTS.models import pwfts, song, ifts
 from pyFTS.models.ensemble import ensemble
 
 '''
-#model = knn.KNearestNeighbors("")
+model = knn.KNearestNeighbors(order=3)
 #model = ensemble.AllMethodEnsembleFTS("", partitioner=partitioner)
 #model = arima.ARIMA("", order=(2,0,2))
 #model = quantreg.QuantileRegression("", order=2, dist=True)
-model.append_transformation(tdiff)
+#model.append_transformation(tdiff)
 model.fit(dataset[:800])
-Measures.get_distribution_statistics(dataset[800:1000], model)
+print(Measures.get_distribution_statistics(dataset[800:1000], model))
 #tmp = model.predict(dataset[800:1000], type='distribution')
 #for tmp2 in tmp:
 #    print(tmp2)
 #'''
+#'''
+
+bchmk.sliding_window_benchmarks(dataset, 1000, train=0.8, inc=0.2,
+                                methods=[pwfts.ProbabilisticWeightedFTS],
+                                benchmark_models=False,
+                                transformations=[None],
+                                orders=[1, 2, 3],
+                                partitions=np.arange(10, 90, 5),
+                                progress=False, type="point",
+                                #steps_ahead=[1,2,4,6,8,10],
+                                distributed=True, nodes=['192.168.0.110', '192.168.0.107', '192.168.0.106'],
+                                file="benchmarks.db", dataset="SP500", tag="partitioning")
 
 
+
+bchmk.sliding_window_benchmarks(dataset, 1000, train=0.8, inc=0.2,
+                                methods=[pwfts.ProbabilisticWeightedFTS],
+                                benchmark_models=False,
+                                transformations=[tdiff],
+                                orders=[1, 2, 3],
+                                partitions=np.arange(3, 30, 2),
+                                progress=False, type="point",
+                                #steps_ahead=[1,2,4,6,8,10],
+                                distributed=True, nodes=['192.168.0.110', '192.168.0.107', '192.168.0.106'],
+                                file="benchmarks.db", dataset="SP500", tag="partitioning")
+
+#'''
 '''
 from pyFTS.partitioners import Grid, Util as pUtil
 partitioner = Grid.GridPartitioner(data=dataset[:800], npart=10, transformation=tdiff)
@@ -52,24 +78,76 @@ 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(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)},
+        {'order': (1, 0, 1)},
+        {'order': (2, 0, 1)},
+        {'order': (2, 0, 2)},
+        {},
+    ],[
+        {'order': (1, 0, 0), 'alpha': .05},
+        {'order': (1, 0, 0), 'alpha': .25},
+        {'order': (1, 0, 1), 'alpha': .05},
+        {'order': (1, 0, 1), 'alpha': .25},
+        {'order': (2, 0, 1), 'alpha': .05},
+        {'order': (2, 0, 1), 'alpha': .25},
+        {'order': (2, 0, 2), 'alpha': .05},
+        {'order': (2, 0, 2), 'alpha': .25},
+        {'order': 1, 'alpha': .05},
+        {'order': 1, 'alpha': .25},
+        {'order': 2, 'alpha': .05},
+        {'order': 2, 'alpha': .25}
+    ],[
+        {'order': (1, 0, 0)},
+        {'order': (1, 0, 1)},
+        {'order': (2, 0, 1)},
+        {'order': (2, 0, 2)},
+        {'order': 1, 'dist': True},
+        {'order': 2, 'dist': True},
+        {'order': 1}, {'order': 2}, {'order': 3},
+    ]
+]
+dataset_name = "NASDAQ"
+tag = "comparisons"
 
 from pyFTS.benchmarks import arima, naive, quantreg
 
-bchmk.sliding_window_benchmarks(dataset, 1000, train=0.8, inc=0.2,
-                                methods=[pwfts.ProbabilisticWeightedFTS],
-                                benchmark_models=False,
-                                transformations=[tdiff],
-                                orders=[1,2,3],
-                                partitions=np.arange(3, 50, 2),
-                                progress=False, type='point',
-                                #steps_ahead=[1,4,7,10], #steps_ahead=[1]
-                                distributed=True, nodes=['192.168.0.110', '192.168.0.107','192.168.0.106'],
-                                file="benchmarks.db", dataset="SP500", tag="partitioning")
+for ct, type in enumerate(types):
+
+    bchmk.sliding_window_benchmarks(dataset, 1000, train=0.8, inc=0.2,
+                                    benchmark_models=True,
+                                    benchmark_methods=benchmark_methods[ct],
+                                    benchmark_methods_parameters=benchmark_methods_parameters[ct],
+                                    transformations=[None],
+                                    orders=[1,2,3],
+                                    partitions=np.arange(15, 85, 5),
+                                    progress=False, type=type,
+                                    distributed=True, nodes=['192.168.0.110', '192.168.0.107','192.168.0.106'],
+                                    file="benchmarks.db", dataset=dataset_name, tag=tag)
+
+    bchmk.sliding_window_benchmarks(dataset, 1000, train=0.8, inc=0.2,
+                                    benchmark_models=True,
+                                    benchmark_methods=benchmark_methods[ct],
+                                    benchmark_methods_parameters=benchmark_methods_parameters[ct],
+                                    transformations=[tdiff],
+                                    orders=[1, 2, 3],
+                                    partitiTAIEXons=np.arange(3, 35, 2),
+                                    progress=False, type=type,
+                                    distributed=True, nodes=['192.168.0.110', '192.168.0.107', '192.168.0.106'],
+                                    file="benchmarks.db", dataset=dataset_name, tag=tag)
 
 
-#'''
+'''
 '''
 dat = pd.read_csv('pwfts_taiex_partitioning.csv', sep=';')
 print(bUtil.analytic_tabular_dataframe(dat))