In [1]:
import pandas as pd
import numpy as np
import pyaf.ForecastEngine as autof
import pyaf.Bench.TS_datasets as tsds
import pyaf.Bench.NN3 as tNN3
%matplotlib inline
In [2]:
tester1 = tNN3.cNN_Tester(tsds.load_NN5() , "NN5");
tester1.testSignals('NN5-025')
BENCH_TYPE NN5 OneDataFrameForAllSignals
BENCH_DATA NN5 <pyaf.Bench.TS_datasets.cTimeSeriesDatasetSpec object at 0x7f8876ddf5f8>
TIME : Day N= 789 H= 2 HEAD= ['1996-03-18T00:00:00.000000000' '1996-03-19T00:00:00.000000000'
'1996-03-20T00:00:00.000000000' '1996-03-21T00:00:00.000000000'
'1996-03-22T00:00:00.000000000'] TAIL= ['1998-05-11T00:00:00.000000000' '1998-05-12T00:00:00.000000000'
'1998-05-13T00:00:00.000000000' '1998-05-14T00:00:00.000000000'
'1998-05-15T00:00:00.000000000']
SIGNAL : NN5-025 N= 789 H= 2 HEAD= [ 9.467 11.876 17.956 25.071 19.827] TAIL= [ 17.829 15.59 22.676 32.625 32.129]
NN5-025 Day
0 9.467 1996-03-18
1 11.876 1996-03-19
2 17.956 1996-03-20
3 25.071 1996-03-21
4 19.827 1996-03-22
<class 'pandas.core.frame.DataFrame'>
Int64Index: 789 entries, 0 to 788
Data columns (total 2 columns):
NN5-025 789 non-null float64
Day 789 non-null datetime64[ns]
dtypes: datetime64[ns](1), float64(1)
memory usage: 18.5 KB
None
Transformation Model \
0 _NN5-025 _NN5-025_ConstantTrend_residue_Seasonal_DayOfW...
1 _NN5-025 _NN5-025_ConstantTrend_residue_bestCycle_byL2_...
2 _NN5-025 _NN5-025_LinearTrend_residue_Seasonal_DayOfWee...
3 _NN5-025 _NN5-025_LinearTrend_residue_bestCycle_byL2_re...
4 CumSum_NN5-025 CumSum_NN5-025_LinearTrend_residue_zeroCycle_r...
5 CumSum_NN5-025 CumSum_NN5-025_LinearTrend_residue_Seasonal_Da...
6 Diff_NN5-025 Diff_NN5-025_LinearTrend_residue_Seasonal_DayO...
7 _NN5-025 _NN5-025_ConstantTrend_residue_zeroCycle_resid...
8 _NN5-025 _NN5-025_ConstantTrend_residue_Seasonal_DayOfM...
9 CumSum_NN5-025 CumSum_NN5-025_LinearTrend_residue_bestCycle_b...
10 Diff_NN5-025 Diff_NN5-025_LinearTrend_residue_zeroCycle_res...
11 Diff_NN5-025 Diff_NN5-025_ConstantTrend_residue_Seasonal_Da...
12 CumSum_NN5-025 CumSum_NN5-025_PolyTrend_residue_Seasonal_DayO...
13 CumSum_NN5-025 CumSum_NN5-025_PolyTrend_residue_bestCycle_byL...
14 CumSum_NN5-025 CumSum_NN5-025_LinearTrend_residue_Seasonal_Da...
15 _NN5-025 _NN5-025_LinearTrend_residue_Seasonal_DayOfMon...
16 _NN5-025 _NN5-025_LinearTrend_residue_Seasonal_DayOfWee...
17 _NN5-025 _NN5-025_LinearTrend_residue_bestCycle_byL2_re...
18 _NN5-025 _NN5-025_LinearTrend_residue_zeroCycle_residue...
19 Diff_NN5-025 Diff_NN5-025_LinearTrend_residue_Seasonal_DayO...
20 Diff_NN5-025 Diff_NN5-025_LinearTrend_residue_bestCycle_byL...
21 Diff_NN5-025 Diff_NN5-025_PolyTrend_residue_Seasonal_DayOfM...
22 _NN5-025 _NN5-025_PolyTrend_residue_Seasonal_DayOfWeek_...
23 _NN5-025 _NN5-025_PolyTrend_residue_bestCycle_byL2_resi...
24 _NN5-025 _NN5-025_ConstantTrend_residue_Seasonal_DayOfM...
25 _NN5-025 _NN5-025_ConstantTrend_residue_Seasonal_DayOfW...
26 _NN5-025 _NN5-025_ConstantTrend_residue_bestCycle_byL2_...
27 Diff_NN5-025 Diff_NN5-025_ConstantTrend_residue_zeroCycle_r...
28 _NN5-025 _NN5-025_PolyTrend_residue_Seasonal_DayOfWeek_...
29 _NN5-025 _NN5-025_PolyTrend_residue_bestCycle_byL2_resi...
.. ... ...
98 CumSum_NN5-025 CumSum_NN5-025_ConstantTrend_residue_Seasonal_...
99 RelDiff_NN5-025 RelDiff_NN5-025_ConstantTrend_residue_zeroCycl...
100 RelDiff_NN5-025 RelDiff_NN5-025_ConstantTrend_residue_Seasonal...
101 RelDiff_NN5-025 RelDiff_NN5-025_ConstantTrend_residue_Seasonal...
102 RelDiff_NN5-025 RelDiff_NN5-025_ConstantTrend_residue_bestCycl...
103 RelDiff_NN5-025 RelDiff_NN5-025_LinearTrend_residue_zeroCycle_...
104 RelDiff_NN5-025 RelDiff_NN5-025_PolyTrend_residue_zeroCycle_re...
105 RelDiff_NN5-025 RelDiff_NN5-025_LinearTrend_residue_Seasonal_D...
106 RelDiff_NN5-025 RelDiff_NN5-025_LinearTrend_residue_Seasonal_D...
107 RelDiff_NN5-025 RelDiff_NN5-025_PolyTrend_residue_Seasonal_Day...
108 RelDiff_NN5-025 RelDiff_NN5-025_PolyTrend_residue_Seasonal_Day...
109 RelDiff_NN5-025 RelDiff_NN5-025_LinearTrend_residue_bestCycle_...
110 RelDiff_NN5-025 RelDiff_NN5-025_PolyTrend_residue_bestCycle_by...
111 RelDiff_NN5-025 RelDiff_NN5-025_ConstantTrend_residue_zeroCycl...
112 RelDiff_NN5-025 RelDiff_NN5-025_ConstantTrend_residue_Seasonal...
113 RelDiff_NN5-025 RelDiff_NN5-025_ConstantTrend_residue_Seasonal...
114 RelDiff_NN5-025 RelDiff_NN5-025_ConstantTrend_residue_bestCycl...
115 RelDiff_NN5-025 RelDiff_NN5-025_LinearTrend_residue_zeroCycle_...
116 RelDiff_NN5-025 RelDiff_NN5-025_PolyTrend_residue_zeroCycle_re...
117 RelDiff_NN5-025 RelDiff_NN5-025_LinearTrend_residue_Seasonal_D...
118 RelDiff_NN5-025 RelDiff_NN5-025_LinearTrend_residue_Seasonal_D...
119 RelDiff_NN5-025 RelDiff_NN5-025_PolyTrend_residue_Seasonal_Day...
120 RelDiff_NN5-025 RelDiff_NN5-025_PolyTrend_residue_Seasonal_Day...
121 RelDiff_NN5-025 RelDiff_NN5-025_LinearTrend_residue_bestCycle_...
122 RelDiff_NN5-025 RelDiff_NN5-025_PolyTrend_residue_bestCycle_by...
123 RelDiff_NN5-025 RelDiff_NN5-025_Lag1Trend_residue_zeroCycle_re...
124 RelDiff_NN5-025 RelDiff_NN5-025_Lag1Trend_residue_Seasonal_Day...
125 RelDiff_NN5-025 RelDiff_NN5-025_Lag1Trend_residue_Seasonal_Day...
126 RelDiff_NN5-025 RelDiff_NN5-025_Lag1Trend_residue_bestCycle_by...
127 CumSum_NN5-025 CumSum_NN5-025_ConstantTrend_residue_Seasonal_...
Complexity FitCount FitL2 FitMAPE ForecastCount ForecastL2 \
0 4 629 5.022606 0.2931 158 7.573948
1 8 629 5.022606 0.2931 158 7.573948
2 20 629 4.970688 0.2912 158 7.018754
3 24 629 4.970688 0.2912 158 7.018754
4 48 629 6.877336 0.4130 158 9.032353
5 52 629 5.392996 0.3005 158 7.934865
6 52 629 6.928297 0.4320 158 9.924127
7 0 629 6.658265 0.4231 158 8.818396
8 4 629 6.573369 0.4200 158 8.887663
9 56 629 6.933389 0.4190 158 9.106485
10 48 629 7.084834 0.4570 158 8.909653
11 36 629 7.585191 0.3767 158 8.656637
12 52 629 16.576367 0.3930 158 6.834463
13 56 629 16.576367 0.3930 158 6.834463
14 52 629 7.306737 0.4443 158 9.713018
15 20 629 6.532302 0.4203 158 8.419405
16 217 629 3.937904 0.2449 158 7.645109
17 221 629 3.937904 0.2449 158 7.645109
18 16 629 6.617517 0.4235 158 8.359341
19 52 629 6.773523 0.5122 158 6.873718
20 56 629 6.773523 0.5122 158 6.873718
21 52 629 6.848235 0.4138 158 8.432172
22 20 629 4.944571 0.2855 158 6.882791
23 24 629 4.944571 0.2855 158 6.882791
24 201 629 4.071843 0.2656 158 7.493268
25 201 629 3.909158 0.2447 158 7.411272
26 205 629 3.909158 0.2447 158 7.411272
27 32 629 7.500825 0.3890 158 8.231308
28 217 629 3.925036 0.2443 158 7.392021
29 221 629 3.925036 0.2443 158 7.392021
.. ... ... ... ... ... ...
98 36 629 214.830337 3.0047 158 79.982365
99 32 629 37.519985 2.8167 158 37.732646
100 36 629 37.452915 2.8096 158 37.732646
101 36 629 35.865084 2.6568 158 37.732646
102 40 629 35.865084 2.6568 158 37.732646
103 48 629 37.105786 2.7749 158 37.732646
104 48 629 36.903691 2.7539 158 37.732646
105 52 629 37.044432 2.7679 158 37.732646
106 52 629 34.855346 2.5555 158 37.732646
107 52 629 36.823037 2.7457 158 37.732646
108 52 629 34.463433 2.5122 158 37.732646
109 56 629 34.855346 2.5555 158 37.732646
110 56 629 34.463433 2.5122 158 37.732646
111 229 629 37.687394 2.8325 158 37.732646
112 233 629 36.185949 2.6678 158 37.732646
113 233 629 34.796263 2.5524 158 37.732646
114 237 629 34.796263 2.5524 158 37.732646
115 245 629 37.193370 2.7841 158 37.732646
116 245 629 36.889550 2.7537 158 37.732646
117 249 629 35.931437 2.6423 158 37.732646
118 249 629 33.895066 2.4603 158 37.732646
119 249 629 35.828164 2.6313 158 37.732646
120 249 629 34.091700 2.4806 158 37.732646
121 253 629 33.895066 2.4603 158 37.732646
122 253 629 34.091700 2.4806 158 37.732646
123 261 629 33.909216 2.5161 158 37.732646
124 265 629 34.957567 2.6181 158 37.732646
125 265 629 30.142738 2.1049 158 37.732646
126 269 629 30.142738 2.1049 158 37.732646
127 233 629 49.146787 2.2931 158 59.424438
ForecastMAPE TestCount TestL2 TestMAPE
0 0.4959 2 9.470316 0.2865
1 0.4959 2 9.470316 0.2865
2 0.4990 2 7.643239 0.2285
3 0.4990 2 7.643239 0.2285
4 0.5000 2 15.985112 0.4936
5 0.5024 2 9.470316 0.2865
6 0.5039 2 19.921159 0.6150
7 0.5103 2 15.450946 0.4771
8 0.5134 2 15.520670 0.4790
9 0.5145 2 15.295947 0.4723
10 0.5148 2 17.419444 0.5379
11 0.5225 2 14.275721 0.4405
12 0.5427 2 5.280554 0.1517
13 0.5427 2 5.280554 0.1517
14 0.5442 2 15.520670 0.4790
15 0.5475 2 13.652146 0.4212
16 0.5501 2 6.442956 0.1750
17 0.5501 2 6.442956 0.1750
18 0.5526 2 13.568739 0.4190
19 0.5553 2 7.626849 0.2278
20 0.5553 2 7.626849 0.2278
21 0.5607 2 10.406847 0.3208
22 0.5639 2 3.812120 0.1013
23 0.5639 2 3.812120 0.1013
24 0.5693 2 7.867496 0.1881
25 0.5784 2 5.615530 0.1454
26 0.5784 2 5.615530 0.1454
27 0.5848 2 11.770077 0.3634
28 0.5848 2 3.539562 0.0819
29 0.5848 2 3.539562 0.0819
.. ... ... ... ...
98 1.9911 2 15.520670 0.4790
99 2.9014 2 24.696245 0.7628
100 2.9014 2 24.696245 0.7628
101 2.9014 2 24.696245 0.7628
102 2.9014 2 24.696245 0.7628
103 2.9014 2 24.696245 0.7628
104 2.9014 2 24.696245 0.7628
105 2.9014 2 24.696245 0.7628
106 2.9014 2 24.696245 0.7628
107 2.9014 2 24.696245 0.7628
108 2.9014 2 24.696245 0.7628
109 2.9014 2 24.696245 0.7628
110 2.9014 2 24.696245 0.7628
111 2.9014 2 24.696245 0.7628
112 2.9014 2 24.696245 0.7628
113 2.9014 2 24.696245 0.7628
114 2.9014 2 24.696245 0.7628
115 2.9014 2 24.696245 0.7628
116 2.9014 2 24.696245 0.7628
117 2.9014 2 24.696245 0.7628
118 2.9014 2 24.696245 0.7628
119 2.9014 2 24.696245 0.7628
120 2.9014 2 24.696245 0.7628
121 2.9014 2 24.696245 0.7628
122 2.9014 2 24.696245 0.7628
123 2.9014 2 24.696245 0.7628
124 2.9014 2 24.696245 0.7628
125 2.9014 2 24.696245 0.7628
126 2.9014 2 24.696245 0.7628
127 3.0275 2 33.841962 0.7794
[128 rows x 12 columns]
Transformation Model \
0 _NN5-025 _NN5-025_ConstantTrend_residue_Seasonal_DayOfW...
1 _NN5-025 _NN5-025_ConstantTrend_residue_bestCycle_byL2_...
2 _NN5-025 _NN5-025_LinearTrend_residue_Seasonal_DayOfWee...
3 _NN5-025 _NN5-025_LinearTrend_residue_bestCycle_byL2_re...
4 CumSum_NN5-025 CumSum_NN5-025_LinearTrend_residue_zeroCycle_r...
Complexity FitCount FitL2 FitMAPE ForecastCount ForecastL2 \
0 4 629 5.022606 0.2931 158 7.573948
1 8 629 5.022606 0.2931 158 7.573948
2 20 629 4.970688 0.2912 158 7.018754
3 24 629 4.970688 0.2912 158 7.018754
4 48 629 6.877336 0.4130 158 9.032353
ForecastMAPE TestCount TestL2 TestMAPE
0 0.4959 2 9.470316 0.2865
1 0.4959 2 9.470316 0.2865
2 0.4990 2 7.643239 0.2285
3 0.4990 2 7.643239 0.2285
4 0.5000 2 15.985112 0.4936
2 0 6.25
1 18.41
Name: NN5-025, dtype: float64
2 0 14.380033
1 14.260056
Name: NN5-025_Forecast, dtype: float64
FORECAST_DETAIL_ACTUAL NN5 NN5-025 [ 6.25 18.41]
FORECAST_DETAIL_PREDICTED NN5 NN5-025 [ 14.38003333 14.26005618]
BENCHMARK_PERF_DETAIL NN5 NN5-025 789 2 11.402945041656494 ConstantTrend + Seasonal_DayOfWeek + NoAR 2 0.7631 0.2606
In [ ]:
Content source: antoinecarme/pyaf
Similar notebooks: