Logo and icon

pyFTS/benchmarks/Util.py

@@ -47,8 +47,8 @@ def find_best(dataframe, criteria, ascending):
 
 
 def point_dataframe_synthetic_columns():
-    return ["Model", "Order", "Scheme", "Partitions", "Size", "RMSEAVG", "RMSESTD", "SMAPEAVG", "SMAPESTD", "UAVG",
+    return ["Model", "Order", "Scheme", "Partitions", "Size", "Steps", "Method", "RMSEAVG", "RMSESTD",
-            "USTD", "TIMEAVG", "TIMESTD"]
+            "SMAPEAVG", "SMAPESTD", "UAVG","USTD", "TIMEAVG", "TIMESTD"]
 
 
 def point_dataframe_analytic_columns(experiments):
@@ -58,11 +58,13 @@ def point_dataframe_analytic_columns(experiments):
     columns.insert(2, "Scheme")
     columns.insert(3, "Partitions")
     columns.insert(4, "Size")
-    columns.insert(5, "Measure")
+    columns.insert(5, "Steps")
+    columns.insert(6, "Method")
+    columns.insert(7, "Measure")
     return columns
 
 
-def save_dataframe_point(experiments, file, objs, rmse, save, synthetic, smape, times, u):
+def save_dataframe_point(experiments, file, objs, rmse, save, synthetic, smape, times, u, steps, method):
     """
     Create a dataframe to store the benchmark results
     :param experiments: dictionary with the execution results
@@ -90,6 +92,8 @@ def save_dataframe_point(experiments, file, objs, rmse, save, synthetic, smape,
                     mod.append(mfts.partitioner.name)
                     mod.append(mfts.partitioner.partitions)
                     mod.append(len(mfts))
+                    mod.append(steps)
+                    mod.append(method)
                 else:
                     mod.append('-')
                     mod.append('-')
@@ -122,17 +126,17 @@ def save_dataframe_point(experiments, file, objs, rmse, save, synthetic, smape,
                     s = '-'
                     p = '-'
                     l = '-'
-                print([n, o, s, p, l])
+                print([n, o, s, p, l, steps, method])
-                tmp = [n, o, s, p, l, 'RMSE']
+                tmp = [n, o, s, p, l, steps, method, 'RMSE']
                 tmp.extend(rmse[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'SMAPE']
+                tmp = [n, o, s, p, l, steps, method, 'SMAPE']
                 tmp.extend(smape[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'U']
+                tmp = [n, o, s, p, l, steps, method, 'U']
                 tmp.extend(u[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'TIME']
+                tmp = [n, o, s, p, l, steps, method, 'TIME']
                 tmp.extend(times[k])
                 ret.append(deepcopy(tmp))
             except Exception as ex:
@@ -401,7 +405,8 @@ def check_ignore_list(b, ignore):
     return flag
 
 
-def save_dataframe_interval(coverage, experiments, file, objs, resolution, save, sharpness, synthetic, times, q05, q25, q75, q95):
+def save_dataframe_interval(coverage, experiments, file, objs, resolution, save, sharpness, synthetic, times,
+                            q05, q25, q75, q95, steps, method):
     ret = []
     if synthetic:
         for k in sorted(objs.keys()):
@@ -409,14 +414,19 @@ def save_dataframe_interval(coverage, experiments, file, objs, resolution, save,
             mfts = objs[k]
             mod.append(mfts.shortname)
             mod.append(mfts.order)
+            l = len(mfts)
             if not mfts.benchmark_only:
                 mod.append(mfts.partitioner.name)
                 mod.append(mfts.partitioner.partitions)
-                l = len(mfts)
+                mod.append(l)
+                mod.append(steps)
+                mod.append(method)
             else:
                 mod.append('-')
                 mod.append('-')
-                l = '-'
+                mod.append('-')
+                mod.append(steps)
+                mod.append(method)
             mod.append(round(np.nanmean(sharpness[k]), 2))
             mod.append(round(np.nanstd(sharpness[k]), 2))
             mod.append(round(np.nanmean(resolution[k]), 2))
@@ -452,28 +462,28 @@ def save_dataframe_interval(coverage, experiments, file, objs, resolution, save,
                     p = '-'
                     l = '-'
 
-                tmp = [n, o, s, p, l, 'Sharpness']
+                tmp = [n, o, s, p, l, steps, method, 'Sharpness']
                 tmp.extend(sharpness[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'Resolution']
+                tmp = [n, o, s, p, l, steps, method, 'Resolution']
                 tmp.extend(resolution[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'Coverage']
+                tmp = [n, o, s, p, l, steps, method, 'Coverage']
                 tmp.extend(coverage[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'TIME']
+                tmp = [n, o, s, p, l, steps, method, 'TIME']
                 tmp.extend(times[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'Q05']
+                tmp = [n, o, s, p, l, steps, method, 'Q05']
                 tmp.extend(q05[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'Q25']
+                tmp = [n, o, s, p, l, steps, method, 'Q25']
                 tmp.extend(q25[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'Q75']
+                tmp = [n, o, s, p, l, steps, method, 'Q75']
                 tmp.extend(q75[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'Q95']
+                tmp = [n, o, s, p, l, steps, method, 'Q95']
                 tmp.extend(q95[k])
                 ret.append(deepcopy(tmp))
             except Exception as ex:
@@ -492,18 +502,19 @@ def interval_dataframe_analytic_columns(experiments):
     columns.insert(2, "Scheme")
     columns.insert(3, "Partitions")
     columns.insert(4, "Size")
-    columns.insert(5, "Measure")
+    columns.insert(5, "Steps")
+    columns.insert(6, "Method")
+    columns.insert(7, "Measure")
     return columns
 
 
 
 def interval_dataframe_synthetic_columns():
-    columns = ["Model", "Order", "Scheme", "Partitions", "SHARPAVG", "SHARPSTD", "RESAVG", "RESSTD", "COVAVG",
+    columns = ["Model", "Order", "Scheme", "Partitions","SIZE", "Steps","Method" "SHARPAVG", "SHARPSTD", "RESAVG", "RESSTD", "COVAVG",
                "COVSTD", "TIMEAVG", "TIMESTD", "Q05AVG", "Q05STD", "Q25AVG", "Q25STD", "Q75AVG", "Q75STD", "Q95AVG", "Q95STD"]
     return columns
 
 
-
 def cast_dataframe_to_synthetic_interval(infile, outfile, experiments):
     columns = interval_dataframe_analytic_columns(experiments)
     dat = pd.read_csv(infile, sep=";", usecols=columns)
@@ -863,7 +874,7 @@ def plot_dataframe_interval_pinball(file_synthetic, file_analytic, experiments,
     Util.show_and_save_image(fig, file, save)
 
 
-def save_dataframe_probabilistic(experiments, file, objs, crps, times, save, synthetic):
+def save_dataframe_probabilistic(experiments, file, objs, crps, times, save, synthetic, steps, method):
     """
     Save benchmark results for m-step ahead probabilistic forecasters 
     :param experiments: 
@@ -893,11 +904,15 @@ def save_dataframe_probabilistic(experiments, file, objs, crps, times, save, syn
                         if not mfts.benchmark_only:
                             mod.append(mfts.partitioner.name)
                             mod.append(mfts.partitioner.partitions)
-                            l = len(mfts)
+                            mod.append(len(mfts))
+                            mod.append(steps)
+                            mod.append(method)
                         else:
                             mod.append('-')
                             mod.append('-')
-                            l = '-'
+                            mod.append('-')
+                            mod.append(steps)
+                            mod.append(method)
                         mod.append(np.round(np.nanmean(crps[k]), 2))
                         mod.append(np.round(np.nanstd(crps[k]), 2))
                         mod.append(l)
@@ -925,10 +940,10 @@ def save_dataframe_probabilistic(experiments, file, objs, crps, times, save, syn
                     s = '-'
                     p = '-'
                     l = '-'
-                tmp = [n, o, s, p, l, 'CRPS']
+                tmp = [n, o, s, p, l, steps, method, 'CRPS']
                 tmp.extend(crps[k])
                 ret.append(deepcopy(tmp))
-                tmp = [n, o, s, p, l, 'TIME']
+                tmp = [n, o, s, p, l, steps, method, 'TIME']
                 tmp.extend(times[k])
                 ret.append(deepcopy(tmp))
             except Exception as ex:
@@ -940,7 +955,6 @@ def save_dataframe_probabilistic(experiments, file, objs, crps, times, save, syn
     return dat
 
 
-
 def probabilistic_dataframe_analytic_columns(experiments):
     columns = [str(k) for k in np.arange(0, experiments)]
     columns.insert(0, "Model")
@@ -948,12 +962,14 @@ def probabilistic_dataframe_analytic_columns(experiments):
     columns.insert(2, "Scheme")
     columns.insert(3, "Partitions")
     columns.insert(4, "Size")
-    columns.insert(5, "Measure")
+    columns.insert(5, "Steps")
+    columns.insert(6, "Method")
+    columns.insert(7, "Measure")
     return columns
 
 
 def probabilistic_dataframe_synthetic_columns():
-    columns = ["Model", "Order", "Scheme", "Partitions", "CRPSAVG", "CRPSSTD",
+    columns = ["Model", "Order", "Scheme", "Partitions","Size", "Steps", "Method", "CRPSAVG", "CRPSSTD",
                "TIMEAVG", "TIMESTD"]
     return columns
 

pyFTS/benchmarks/benchmarks.py

@@ -293,6 +293,9 @@ def run_point(mfts, partitioner, train_data, test_data, window_key=None, **kwarg
     transformation = kwargs.get('transformation', None)
     indexer = kwargs.get('indexer', None)
 
+    steps_ahead = kwargs.get('steps_ahead', 1)
+    method = kwargs.get('method', None)
+
     if mfts.benchmark_only:
         _key = mfts.shortname + str(mfts.order if mfts.order is not None else "")
     else:
@@ -313,7 +316,8 @@ def run_point(mfts, partitioner, train_data, test_data, window_key=None, **kwarg
     _end = time.time()
     times += _end - _start
 
-    ret = {'key': _key, 'obj': mfts, 'rmse': _rmse, 'smape': _smape, 'u': _u, 'time': times, 'window': window_key}
+    ret = {'key': _key, 'obj': mfts, 'rmse': _rmse, 'smape': _smape, 'u': _u, 'time': times, 'window': window_key,
+           'steps': steps_ahead, 'method': method}
 
     return ret
 
@@ -346,6 +350,9 @@ def run_interval(mfts, partitioner, train_data, test_data, window_key=None, **kw
     transformation = kwargs.get('transformation', None)
     indexer = kwargs.get('indexer', None)
 
+    steps_ahead = kwargs.get('steps_ahead', 1)
+    method = kwargs.get('method', None)
+
     if mfts.benchmark_only:
         _key = mfts.shortname + str(mfts.order if mfts.order is not None else "") + str(mfts.alpha)
     else:
@@ -367,7 +374,8 @@ def run_interval(mfts, partitioner, train_data, test_data, window_key=None, **kw
     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}
+           'Q05': _q05, 'Q25': _q25, 'Q75': _q75, 'Q95': _q95, 'window': window_key,
+           'steps': steps_ahead, 'method': method}
 
     return ret
 
@@ -403,6 +411,9 @@ def run_probabilistic(mfts, partitioner, train_data, test_data, window_key=None,
     transformation = kwargs.get('transformation', None)
     indexer = kwargs.get('indexer', None)
 
+    steps_ahead = kwargs.get('steps_ahead', 1)
+    method = kwargs.get('method', None)
+
     if mfts.benchmark_only:
         _key = mfts.shortname + str(mfts.order if mfts.order is not None else "") + str(mfts.alpha)
     else:
@@ -429,7 +440,8 @@ def run_probabilistic(mfts, partitioner, train_data, test_data, window_key=None,
         _crps1 = np.nan
         _t1 = np.nan
 
-    ret = {'key': _key, 'obj': mfts, 'CRPS': _crps1, 'time': _t1, 'window': window_key}
+    ret = {'key': _key, 'obj': mfts, 'CRPS': _crps1, 'time': _t1, 'window': window_key,
+           'steps': steps_ahead, 'method': method}
 
     return ret
 
@@ -466,6 +478,8 @@ def process_point_jobs(jobs, experiments, save=False, file=None, sintetic=False)
     smape = {}
     u = {}
     times = {}
+    steps = None
+    method = None
 
     for job in jobs:
         _key = job['key']
@@ -475,6 +489,8 @@ def process_point_jobs(jobs, experiments, save=False, file=None, sintetic=False)
             smape[_key] = []
             u[_key] = []
             times[_key] = []
+            steps[_key] = job['steps']
+            method[_key] = job['method']
         rmse[_key].append(job['rmse'])
         smape[_key].append(job['smape'])
         u[_key].append(job['u'])

pyFTS/common/fts.py

@@ -89,17 +89,17 @@ class FTS(object):
 
             steps_ahead = kwargs.get("steps_ahead", None)
 
-            if type == 'point' and steps_ahead == None:
+            if type == 'point' and (steps_ahead == None or steps_ahead == 1):
                 ret = self.forecast(ndata, **kwargs)
-            elif type == 'point' and steps_ahead != None:
+            elif type == 'point' and steps_ahead > 1:
                 ret = self.forecast_ahead(ndata, steps_ahead, **kwargs)
-            elif type == 'interval' and steps_ahead == None:
+            elif type == 'interval' and (steps_ahead == None or steps_ahead == 1):
                 ret = self.forecast_interval(ndata, **kwargs)
-            elif type == 'interval' and steps_ahead != None:
+            elif type == 'interval' and steps_ahead > 1:
                 ret = self.forecast_ahead_interval(ndata, steps_ahead, **kwargs)
-            elif type == 'distribution' and steps_ahead == None:
+            elif type == 'distribution' and (steps_ahead == None or steps_ahead == 1):
                 ret = self.forecast_distribution(ndata, **kwargs)
-            elif type == 'distribution' and steps_ahead != None:
+            elif type == 'distribution' and steps_ahead > 1:
                 ret = self.forecast_ahead_distribution(ndata, steps_ahead, **kwargs)
             else:
                 raise ValueError('The argument \'type\' has an unknown value.')