No changepoints, absmax scaling

match ← All logistic growth benchmarks

96 daily rows rising toward a ceiling of about 100, with a little noise. Prediction covers the history and 24 more days. The curve's rate never changes. Values are divided by their largest distance from the floor.

Input

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

Output

Largest difference: 3.9e-8 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-01-01	14.1851	14.1851	1.9e-8
2020-01-02	14.7172	14.7172	1.9e-8
2020-01-03	15.2658	15.2658	2.0e-8
2020-01-04	15.8313	15.8313	2.0e-8
2020-01-05	16.4138	16.4138	2.1e-8
2020-01-06	17.0136	17.0136	2.2e-8
2020-01-07	17.6310	17.6310	2.2e-8
2020-01-08	18.2660	18.2660	2.3e-8
2020-01-09	18.9189	18.9189	2.3e-8
2020-01-10	19.5898	19.5898	2.4e-8
2020-01-11	20.2789	20.2789	2.5e-8
2020-01-12	20.9862	20.9862	2.5e-8
2020-01-13	21.7118	21.7118	2.6e-8
2020-01-14	22.4558	22.4558	2.7e-8
2020-01-15	23.2182	23.2182	2.7e-8
2020-01-16	23.9988	23.9988	2.8e-8
2020-01-17	24.7978	24.7978	2.8e-8
2020-01-18	25.6150	25.6150	2.9e-8
2020-01-19	26.4502	26.4502	3.0e-8
2020-01-20	27.3032	27.3032	3.0e-8
2020-01-21	28.1740	28.1740	3.1e-8
2020-01-22	29.0622	29.0622	3.2e-8
2020-01-23	29.9675	29.9675	3.2e-8
2020-01-24	30.8896	30.8896	3.3e-8
2020-01-25	31.8282	31.8282	3.3e-8
2020-01-26	32.7828	32.7828	3.4e-8
2020-01-27	33.7529	33.7529	3.4e-8
2020-01-28	34.7380	34.7380	3.5e-8
2020-01-29	35.7377	35.7377	3.5e-8
2020-01-30	36.7513	36.7513	3.6e-8
2020-01-31	37.7783	37.7783	3.6e-8
2020-02-01	38.8178	38.8178	3.7e-8
2020-02-02	39.8693	39.8693	3.7e-8
2020-02-03	40.9321	40.9321	3.7e-8
2020-02-04	42.0052	42.0052	3.8e-8
2020-02-05	43.0881	43.0881	3.8e-8
2020-02-06	44.1797	44.1797	3.8e-8
2020-02-07	45.2793	45.2793	3.8e-8
2020-02-08	46.3859	46.3859	3.9e-8
2020-02-09	47.4988	47.4988	3.9e-8
2020-02-10	48.6168	48.6168	3.9e-8
2020-02-11	49.7392	49.7392	3.9e-8
2020-02-12	50.8650	50.8650	3.9e-8
2020-02-13	51.9932	51.9932	3.9e-8
2020-02-14	53.1228	53.1228	3.9e-8
2020-02-15	54.2528	54.2528	3.9e-8
2020-02-16	55.3824	55.3824	3.9e-8
2020-02-17	56.5106	56.5106	3.9e-8
2020-02-18	57.6363	57.6363	3.9e-8
2020-02-19	58.7587	58.7587	3.8e-8
2020-02-20	59.8769	59.8769	3.8e-8
2020-02-21	60.9898	60.9898	3.8e-8
2020-02-22	62.0967	62.0967	3.8e-8
2020-02-23	63.1967	63.1967	3.7e-8
2020-02-24	64.2889	64.2889	3.7e-8
2020-02-25	65.3725	65.3725	3.7e-8
2020-02-26	66.4468	66.4468	3.6e-8
2020-02-27	67.5110	67.5110	3.6e-8
2020-02-28	68.5644	68.5644	3.6e-8
2020-02-29	69.6062	69.6062	3.5e-8
2020-03-01	70.6360	70.6360	3.5e-8
2020-03-02	71.6530	71.6530	3.4e-8
2020-03-03	72.6568	72.6568	3.4e-8
2020-03-04	73.6467	73.6467	3.3e-8
2020-03-05	74.6224	74.6224	3.3e-8
2020-03-06	75.5833	75.5833	3.2e-8
2020-03-07	76.5291	76.5291	3.1e-8
2020-03-08	77.4595	77.4595	3.1e-8
2020-03-09	78.3741	78.3741	3.0e-8
2020-03-10	79.2725	79.2725	3.0e-8
2020-03-11	80.1547	80.1547	2.9e-8
2020-03-12	81.0204	81.0204	2.8e-8
2020-03-13	81.8694	81.8694	2.8e-8
2020-03-14	82.7016	82.7016	2.7e-8
2020-03-15	83.5170	83.5170	2.6e-8
2020-03-16	84.3154	84.3154	2.6e-8
2020-03-17	85.0968	85.0968	2.5e-8
2020-03-18	85.8612	85.8612	2.5e-8
2020-03-19	86.6087	86.6087	2.4e-8
2020-03-20	87.3394	87.3394	2.3e-8
2020-03-21	88.0533	88.0533	2.3e-8
2020-03-22	88.7505	88.7505	2.2e-8
2020-03-23	89.4311	89.4311	2.1e-8
2020-03-24	90.0954	90.0954	2.1e-8
2020-03-25	90.7435	90.7435	2.0e-8
2020-03-26	91.3755	91.3755	2.0e-8
2020-03-27	91.9918	91.9918	1.9e-8
2020-03-28	92.5924	92.5924	1.9e-8
2020-03-29	93.1777	93.1777	1.8e-8
2020-03-30	93.7480	93.7480	1.7e-8
2020-03-31	94.3034	94.3034	1.7e-8
2020-04-01	94.8442	94.8442	1.6e-8
2020-04-02	95.3708	95.3708	1.6e-8
2020-04-03	95.8833	95.8833	1.5e-8
2020-04-04	96.3822	96.3822	1.5e-8
2020-04-05	96.8676	96.8676	1.4e-8
2020-04-06	97.3400	97.3400	1.4e-8
2020-04-07	97.7995	97.7995	1.3e-8
2020-04-08	98.2466	98.2466	1.3e-8
2020-04-09	98.6814	98.6814	1.3e-8
2020-04-10	99.1044	99.1044	1.2e-8
2020-04-11	99.5157	99.5157	1.2e-8
2020-04-12	99.9159	99.9159	1.1e-8
2020-04-13	100.3050	100.3050	1.1e-8
2020-04-14	100.6835	100.6835	1.1e-8
2020-04-15	101.0516	101.0516	1.0e-8
2020-04-16	101.4096	101.4096	9.8e-9
2020-04-17	101.7578	101.7578	9.5e-9
2020-04-18	102.0966	102.0966	9.1e-9
2020-04-19	102.4261	102.4261	8.8e-9
2020-04-20	102.7468	102.7468	8.5e-9
2020-04-21	103.0587	103.0587	8.2e-9
2020-04-22	103.3623	103.3623	7.9e-9
2020-04-23	103.6578	103.6578	7.6e-9
2020-04-24	103.9454	103.9454	7.3e-9
2020-04-25	104.2254	104.2254	7.0e-9
2020-04-26	104.4980	104.4980	6.8e-9
2020-04-27	104.7635	104.7635	6.5e-9
2020-04-28	105.0222	105.0222	6.3e-9
2020-04-29	105.2742	105.2742	6.0e-9

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Fit	6.78 ms	12.7 ms	1.9× faster
Predict	675 µs	2.85 ms	4.2× faster
Save model to JSON	28 µs	428 µs	15× faster
Load model from JSON	133 µs	2.08 ms	16× faster
Load and predict in a new process	155 ms	1.41 s	9.1× 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": []
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [],
    "events": [],
    "regressors": []
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}