- Bugfix on interval forecast of arima

benchmarks/arima.py

@@ -26,7 +26,8 @@ class ARIMA(fts.FTS):
         self.q = 0
         self.benchmark_only = True
         self.min_order = 1
-        self.alpha = (1 - kwargs.get("alpha", 0.90))/2
+        self.alpha = kwargs.get("alpha", 0.05)
+        self.shortname += str(self.alpha)
 
     def train(self, data, sets, order, parameters=None):
         self.p = order[0]
@@ -35,6 +36,8 @@ class ARIMA(fts.FTS):
         self.order = self.p + self.q
         self.shortname = "ARIMA(" + str(self.p) + "," + str(self.d) + "," + str(self.q) + ") - " + str(self.alpha)
 
+        data = self.doTransformations(data, updateUoD=True)
+
         old_fit = self.model_fit
         try:
             self.model =  stats_arima(data, order=(self.p, self.d, self.q))
@@ -85,25 +88,28 @@ class ARIMA(fts.FTS):
 
         sigma = np.sqrt(self.model_fit.sigma2)
 
-        ndata = np.array(self.doTransformations(data))
+        #ndata = np.array(self.doTransformations(data))
 
-        l = len(ndata)
+        l = len(data)
 
         ret = []
 
         for k in np.arange(self.order, l+1):
             tmp = []
 
-            sample = [ndata[i] for i in np.arange(k - self.order, k)]
+            sample = [data[i] for i in np.arange(k - self.order, k)]
 
-            mean = self.forecast(sample)[0]
+            mean = self.forecast(sample)
+
+            if isinstance(mean,(list, np.ndarray)):
+                mean = mean[0]
 
             tmp.append(mean + st.norm.ppf(self.alpha) * sigma)
             tmp.append(mean + st.norm.ppf(1 - self.alpha) * sigma)
 
             ret.append(tmp)
 
-        ret = self.doInverseTransformations(ret, params=[data[self.order - 1:]], interval=True)
+        #ret = self.doInverseTransformations(ret, params=[data[self.order - 1:]], interval=True)
 
         return ret
 
@@ -113,8 +119,6 @@ class ARIMA(fts.FTS):
 
         smoothing = kwargs.get("smoothing",0.2)
 
-        alpha = (1 - kwargs.get("alpha", 0.95))/2
-
         sigma = np.sqrt(self.model_fit.sigma2)
 
         ndata = np.array(self.doTransformations(data))
@@ -130,8 +134,8 @@ class ARIMA(fts.FTS):
 
             hsigma = (1 + k*smoothing)*sigma
 
-            tmp.append(means[k] + st.norm.ppf(alpha) * hsigma)
-            tmp.append(means[k] + st.norm.ppf(1 - alpha) * hsigma)
+            tmp.append(means[k] + st.norm.ppf(self.alpha) * hsigma)
+            tmp.append(means[k] + st.norm.ppf(1 - self.alpha) * hsigma)
 
             ret.append(tmp)
 

benchmarks/distributed_benchmarks.py

@@ -282,7 +282,7 @@ def interval_sliding_window(data, windowsize, train=0.8,  inc=0.1, models=None,
      :return: DataFrame with the results
      """
 
-    alphas = [0.5, 0.25]
+    alphas = [0.05, 0.25]
 
     if benchmark_models is None and models is None:
         benchmark_models = [arima.ARIMA, arima.ARIMA, arima.ARIMA, arima.ARIMA,

benchmarks/quantreg.py

@@ -10,7 +10,7 @@ from pyFTS import fts
 class QuantileRegression(fts.FTS):
     """Façade for statsmodels.regression.quantile_regression"""
     def __init__(self, name, **kwargs):
-        super(QuantileRegression, self).__init__(1, "QR"+name)
+        super(QuantileRegression, self).__init__(1, "")
         self.name = "QR"
         self.detail = "Quantile Regression"
         self.is_high_order = True
@@ -23,11 +23,12 @@ class QuantileRegression(fts.FTS):
         self.upper_qt = None
         self.mean_qt = None
         self.lower_qt = None
+        self.shortname = "QAR("+str(self.order)+","+str(self.alpha)+")"
 
     def train(self, data, sets, order=1, parameters=None):
         self.order = order
 
-        tmp = np.array(self.doTransformations(data))
+        tmp = np.array(self.doTransformations(data, updateUoD=True))
 
         lagdata, ndata = lagmat(tmp, maxlag=order, trim="both", original='sep')
 

tests/general.py

@@ -22,6 +22,8 @@ os.chdir("/home/petronio/dados/Dropbox/Doutorado/Codigos/")
 #enrollments = pd.read_csv("DataSets/Enrollments.csv", sep=";")
 #enrollments = np.array(enrollments["Enrollments"])
 
+diff = Transformations.Differential(1)
+
 """
 DATASETS
 """
@@ -60,25 +62,26 @@ from pyFTS.benchmarks import arima, quantreg, Measures
 
 #Util.plot_dataframe_point("experiments/taiex_point_sintetic.csv","experiments/taiex_point_analitic.csv",11)
 
-#tmp = arima.ARIMA("")
-#tmp.train(taiex[:1600], None, order=(2,0,2))
-#teste = tmp.forecastInterval(taiex[1600:1605])
+tmp = arima.ARIMA("", alpha=0.25)
+#tmp.appendTransformation(diff)
+tmp.train(nasdaq[:1600], None, order=(2,0,2))
+teste = tmp.forecastInterval(nasdaq[1600:1604])
 
 """
 tmp = quantreg.QuantileRegression("", alpha=0.25)
 tmp.train(taiex[:1600], None, order=1)
 teste = tmp.forecastInterval(taiex[1600:1605])
-
-print(taiex[1600:1605])
+"""
+print(nasdaq[1600:1605])
 print(teste)
 
-kk = Measures.get_interval_statistics(taiex[1600:1605], tmp)
+kk = Measures.get_interval_statistics(nasdaq[1600:1605], tmp)
 
 print(kk)
-"""
+
 #bchmk.teste(taiex,['192.168.0.109', '192.168.0.101'])
 
-diff = Transformations.Differential(1)
+
 
 """
 bchmk.point_sliding_window(sonda, 9000, train=0.8, inc=0.4,#models=[yu.WeightedFTS], # #
@@ -95,23 +98,23 @@ bchmk.point_sliding_window(sonda, 9000, train=0.8, inc=0.4, #models=[yu.Weighted
                      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(nasdaq, 2000, train=0.8, inc=0.1,#models=[yu.WeightedFTS], # #
+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(10,200,step=10), #transformation=diff,
-                     dump=True, save=True, file="experiments/nasdaq_interval_analytic.csv",
+                     dump=True, save=True, file="experiments/taiex_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(nasdaq, 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,
+                     partitions= np.arange(3,20,step=2), transformation=diff,
                      dump=True, save=True, file="experiments/nasdaq_interval_analytic_diff.csv",
                      nodes=['192.168.0.103', '192.168.0.106', '192.168.0.108', '192.168.0.109']) #, depends=[hofts, ifts])
 
-#"""
+"""
 
 """
 from pyFTS.partitioners import Grid