Weekly pattern

match ← All logistic growth benchmarks

A custom weekly seasonality on the basic curve.

Input

benchmark/inputs/v2/logistic-weekly-96.json: 96 history rows and 120 forecast rows.

Output

Largest difference: 8.7e-8 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-01-01	14.1889	14.1889	5.2e-8
2020-01-02	15.3462	15.3462	7.8e-8
2020-01-03	16.0484	16.0484	8.5e-8
2020-01-04	16.1800	16.1800	6.6e-8
2020-01-05	16.0674	16.0674	3.6e-8
2020-01-06	16.2332	16.2332	2.0e-8
2020-01-07	17.0127	17.0127	2.6e-8
2020-01-08	18.2694	18.2694	4.7e-8
2020-01-09	19.5475	19.5475	7.3e-8
2020-01-10	20.3719	20.3719	7.8e-8
2020-01-11	20.6271	20.6271	5.9e-8
2020-01-12	20.6392	20.6392	2.8e-8
2020-01-13	20.9308	20.9308	1.1e-8
2020-01-14	21.8369	21.8369	1.6e-8
2020-01-15	23.2209	23.2209	3.6e-8
2020-01-16	24.6267	24.6267	6.1e-8
2020-01-17	25.5792	25.5792	6.6e-8
2020-01-18	25.9624	25.9624	4.5e-8
2020-01-19	26.1023	26.1023	1.3e-8
2020-01-20	26.5213	26.5213	4.6e-9
2020-01-21	27.5541	27.5541	4.3e-10
2020-01-22	29.0639	29.0639	1.8e-8
2020-01-23	30.5943	30.5943	4.2e-8
2020-01-24	31.6698	31.6698	4.6e-8
2020-01-25	32.1744	32.1744	2.5e-8
2020-01-26	32.4336	32.4336	8.4e-9
2020-01-27	32.9696	32.9696	2.7e-8
2020-01-28	34.1168	34.1168	2.4e-8
2020-01-29	35.7380	35.7380	6.3e-9
2020-01-30	37.3766	37.3766	1.7e-8
2020-01-31	38.5569	38.5569	2.0e-8
2020-02-01	39.1625	39.1625	2.1e-9
2020-02-02	39.5186	39.5186	3.6e-8
2020-02-03	40.1471	40.1471	5.5e-8
2020-02-04	41.3823	41.3823	5.2e-8
2020-02-05	43.0867	43.0867	3.5e-8
2020-02-06	44.8033	44.8033	1.3e-8
2020-02-07	46.0562	46.0562	1.0e-8
2020-02-08	46.7288	46.7288	3.2e-8
2020-02-09	47.1467	47.1467	5.9e-8
2020-02-10	47.8309	47.8309	7.1e-8
2020-02-11	49.1157	49.1157	6.1e-8
2020-02-12	50.8634	50.8634	3.6e-8
2020-02-13	52.6169	52.6169	5.7e-9
2020-02-14	53.9002	53.9002	4.6e-9
2020-02-15	54.5967	54.5967	1.0e-8
2020-02-16	55.0313	55.0313	3.6e-8
2020-02-17	55.7256	55.7256	4.8e-8
2020-02-18	57.0137	57.0137	3.8e-8
2020-02-19	58.7581	58.7581	1.3e-8
2020-02-20	60.5016	60.5016	1.7e-8
2020-02-21	61.7683	61.7683	2.8e-8
2020-02-22	62.4415	62.4415	1.3e-8
2020-02-23	62.8466	62.8466	1.3e-8
2020-02-24	63.5049	63.5049	2.5e-8
2020-02-25	64.7510	64.7510	1.5e-8
2020-02-26	66.4472	66.4472	9.0e-9
2020-02-27	68.1367	68.1367	3.9e-8
2020-02-28	69.3438	69.3438	4.9e-8
2020-02-29	69.9520	69.9520	3.4e-8
2020-03-01	70.2868	70.2868	6.9e-9
2020-03-02	70.8700	70.8700	5.6e-9
2020-03-03	72.0361	72.0361	3.8e-9
2020-03-04	73.6479	73.6479	2.8e-8
2020-03-05	75.2489	75.2489	5.7e-8
2020-03-06	76.3636	76.3636	6.6e-8
2020-03-07	76.8757	76.8757	5.0e-8
2020-03-08	77.1111	77.1111	2.3e-8
2020-03-09	77.5918	77.5918	9.4e-9
2020-03-10	78.6526	78.6526	1.8e-8
2020-03-11	80.1566	80.1566	4.1e-8
2020-03-12	81.6476	81.6476	7.0e-8
2020-03-13	82.6503	82.6503	7.8e-8
2020-03-14	83.0488	83.0488	6.1e-8
2020-03-15	83.1691	83.1691	3.3e-8
2020-03-16	83.5336	83.5336	1.9e-8
2020-03-17	84.4773	84.4773	2.7e-8
2020-03-18	85.8636	85.8636	4.9e-8
2020-03-19	87.2364	87.2364	7.7e-8
2020-03-20	88.1207	88.1207	8.5e-8
2020-03-21	88.4008	88.4008	6.7e-8
2020-03-22	88.4030	88.4030	3.8e-8
2020-03-23	88.6497	88.6497	2.4e-8
2020-03-24	89.4762	89.4762	3.1e-8
2020-03-25	90.7461	90.7461	5.2e-8
2020-03-26	92.0034	92.0034	8.0e-8
2020-03-27	92.7733	92.7733	8.7e-8
2020-03-28	92.9402	92.9402	6.9e-8
2020-03-29	92.8304	92.8304	3.9e-8
2020-03-30	92.9667	92.9667	2.4e-8
2020-03-31	93.6843	93.6843	3.1e-8
2020-04-01	94.8470	94.8470	5.2e-8
2020-04-02	95.9987	95.9987	7.8e-8
2020-04-03	96.6649	96.6649	8.5e-8
2020-04-04	96.7300	96.7300	6.7e-8
2020-04-05	96.5203	96.5203	3.7e-8
2020-04-06	96.5587	96.5587	2.1e-8
2020-04-07	97.1805	97.1805	2.8e-8
2020-04-08	98.2493	98.2493	4.8e-8
2020-04-09	99.3093	99.3093	7.5e-8
2020-04-10	99.8859	99.8859	8.1e-8
2020-04-11	99.8635	99.8635	6.3e-8
2020-04-12	99.5684	99.5684	3.3e-8
2020-04-13	99.5236	99.5236	1.7e-8
2020-04-14	100.0643	100.0643	2.3e-8
2020-04-15	101.0542	101.0542	4.4e-8
2020-04-16	102.0374	102.0374	7.0e-8
2020-04-17	102.5392	102.5392	7.6e-8
2020-04-18	102.4442	102.4442	5.8e-8
2020-04-19	102.0786	102.0786	2.7e-8
2020-04-20	101.9652	101.9652	1.1e-8
2020-04-21	102.4394	102.4394	1.7e-8
2020-04-22	103.3647	103.3647	3.8e-8
2020-04-23	104.2854	104.2854	6.4e-8
2020-04-24	104.7266	104.7266	7.1e-8
2020-04-25	104.5728	104.5728	5.2e-8
2020-04-26	104.1503	104.1503	2.2e-8
2020-04-27	103.9818	103.9818	5.6e-9
2020-04-28	104.4027	104.4027	1.2e-8
2020-04-29	105.2764	105.2764	3.2e-8

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Fit	15.6 ms	27.5 ms	1.8× faster
Predict	944 µs	3.45 ms	3.7× faster
Save model to JSON	27 µs	482 µs	18× faster
Load model from JSON	166 µs	2.55 ms	15× faster
Load and predict in a new process	167 ms	1.57 s	9.4× faster

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "logistic-python-reconciliation-v1"
  },
  "configuration": {
    "growth": "logistic",
    "scaling": "absmax",
    "seasonalityMode": "additive",
    "holidaysMode": "additive",
    "changepoints": {
      "mode": "explicit",
      "timestamps": [
        "2020-02-08T00:00:00.000Z"
      ]
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [
      {
        "name": "weekly-custom",
        "periodDays": 7,
        "fourierOrder": 3,
        "priorScale": 10
      }
    ],
    "events": [],
    "regressors": []
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}