Defaults, three months of daily data

match ← All logistic growth benchmarks

96 days with a weekly pattern. Weekly seasonality and 25 automatic changepoints switch on by default.

Input

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

Output

Largest difference: 3.1e-9 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-01-01	14.1935	14.1935	2.1e-10
2020-01-02	15.3500	15.3500	1.6e-10
2020-01-03	16.0515	16.0515	6.8e-11
2020-01-04	16.1823	16.1823	4.6e-10
2020-01-05	16.0688	16.0688	1.0e-9
2020-01-06	16.2340	16.2340	1.5e-9
2020-01-07	17.0125	17.0125	1.7e-9
2020-01-08	18.2687	18.2687	1.4e-9
2020-01-09	19.5466	19.5466	1.0e-9
2020-01-10	20.3709	20.3709	7.9e-10
2020-01-11	20.6260	20.6260	8.2e-10
2020-01-12	20.6382	20.6382	9.5e-10
2020-01-13	20.9300	20.9300	9.3e-10
2020-01-14	21.8362	21.8362	8.2e-10
2020-01-15	23.2198	23.2198	7.2e-10
2020-01-16	24.6256	24.6256	5.7e-10
2020-01-17	25.5780	25.5780	3.5e-10
2020-01-18	25.9612	25.9612	2.6e-10
2020-01-19	26.1012	26.1012	2.8e-10
2020-01-20	26.5205	26.5205	2.9e-10
2020-01-21	27.5534	27.5534	4.6e-11
2020-01-22	29.0628	29.0628	1.8e-10
2020-01-23	30.5932	30.5932	4.0e-10
2020-01-24	31.6689	31.6689	4.4e-10
2020-01-25	32.1736	32.1736	3.4e-10
2020-01-26	32.4331	32.4331	5.1e-11
2020-01-27	32.9695	32.9695	9.6e-11
2020-01-28	34.1169	34.1169	8.5e-14
2020-01-29	35.7380	35.7380	3.4e-10
2020-01-30	37.3767	37.3767	7.4e-10
2020-01-31	38.5572	38.5572	9.8e-10
2020-02-01	39.1630	39.1630	9.0e-10
2020-02-02	39.5194	39.5194	7.0e-10
2020-02-03	40.1484	40.1484	6.5e-10
2020-02-04	41.3838	41.3838	8.4e-10
2020-02-05	43.0880	43.0880	1.0e-9
2020-02-06	44.8048	44.8048	1.2e-9
2020-02-07	46.0580	46.0580	1.1e-9
2020-02-08	46.7308	46.7308	8.0e-10
2020-02-09	47.1485	47.1485	3.8e-10
2020-02-10	47.8329	47.8329	1.6e-10
2020-02-11	49.1176	49.1176	4.6e-10
2020-02-12	50.8648	50.8648	7.7e-10
2020-02-13	52.6181	52.6181	7.2e-10
2020-02-14	53.9014	53.9014	8.3e-10
2020-02-15	54.5976	54.5976	1.1e-9
2020-02-16	55.0321	55.0321	7.5e-10
2020-02-17	55.7265	55.7265	2.7e-10
2020-02-18	57.0145	57.0145	4.5e-10
2020-02-19	58.7582	58.7582	1.1e-9
2020-02-20	60.5014	60.5014	1.7e-9
2020-02-21	61.7680	61.7680	2.2e-9
2020-02-22	62.4410	62.4410	2.1e-9
2020-02-23	62.8458	62.8458	2.0e-9
2020-02-24	63.5043	63.5043	2.0e-9
2020-02-25	64.7502	64.7502	2.0e-9
2020-02-26	66.4458	66.4458	1.9e-9
2020-02-27	68.1351	68.1351	2.0e-9
2020-02-28	69.3421	69.3421	1.8e-9
2020-02-29	69.9501	69.9501	1.5e-9
2020-03-01	70.2847	70.2847	1.1e-9
2020-03-02	70.8681	70.8681	7.4e-10
2020-03-03	72.0341	72.0341	6.4e-10
2020-03-04	73.6454	73.6454	5.2e-10
2020-03-05	75.2462	75.2462	2.1e-10
2020-03-06	76.3608	76.3608	2.6e-10
2020-03-07	76.8728	76.8728	8.6e-10
2020-03-08	77.1081	77.1081	1.8e-9
2020-03-09	77.5890	77.5890	2.6e-9
2020-03-10	78.6499	78.6499	3.1e-9
2020-03-11	80.1539	80.1539	2.3e-9
2020-03-12	81.6453	81.6453	1.5e-9
2020-03-13	82.6484	82.6484	9.5e-10
2020-03-14	83.0473	83.0473	7.7e-10
2020-03-15	83.1680	83.1680	7.2e-10
2020-03-16	83.5332	83.5332	5.3e-10
2020-03-17	84.4773	84.4773	3.5e-10
2020-03-18	85.8636	85.8636	1.8e-10
2020-03-19	87.2367	87.2367	4.9e-11
2020-03-20	88.1213	88.1213	1.0e-10
2020-03-21	88.4017	88.4017	1.4e-11
2020-03-22	88.4041	88.4041	2.0e-10
2020-03-23	88.6514	88.6514	2.6e-10
2020-03-24	89.4783	89.4783	9.0e-11
2020-03-25	90.7480	90.7480	6.8e-11
2020-03-26	92.0055	92.0055	2.8e-10
2020-03-27	92.7756	92.7756	3.2e-10
2020-03-28	92.9426	92.9426	2.2e-10
2020-03-29	92.8330	92.8330	2.8e-12
2020-03-30	92.9698	92.9698	8.2e-11
2020-03-31	93.6876	93.6876	8.2e-11
2020-04-01	94.8500	94.8500	2.3e-10
2020-04-02	96.0019	96.0019	4.3e-10
2020-04-03	96.6682	96.6682	4.6e-10
2020-04-04	96.7333	96.7333	3.5e-10
2020-04-05	96.5238	96.5238	1.2e-10
2020-04-06	96.5625	96.5625	2.8e-11
2020-04-07	97.1844	97.1844	1.8e-10
2020-04-08	98.2529	98.2529	3.2e-10
2020-04-09	99.3130	99.3130	5.1e-10
2020-04-10	99.8896	99.8896	5.4e-10
2020-04-11	99.8672	99.8672	4.2e-10
2020-04-12	99.5722	99.5722	1.8e-10
2020-04-13	99.5277	99.5277	8.1e-11
2020-04-14	100.0685	100.0685	2.3e-10
2020-04-15	101.0579	101.0579	3.6e-10
2020-04-16	102.0411	102.0411	5.5e-10
2020-04-17	102.5430	102.5430	5.7e-10
2020-04-18	102.4479	102.4479	4.4e-10
2020-04-19	102.0823	102.0823	1.9e-10
2020-04-20	101.9693	101.9693	9.5e-11
2020-04-21	102.4435	102.4435	2.4e-10
2020-04-22	103.3684	103.3684	3.7e-10
2020-04-23	104.2890	104.2890	5.5e-10
2020-04-24	104.7302	104.7302	5.6e-10
2020-04-25	104.5763	104.5763	4.3e-10
2020-04-26	104.1538	104.1538	1.9e-10
2020-04-27	103.9856	103.9856	8.2e-11
2020-04-28	104.4065	104.4065	2.2e-10
2020-04-29	105.2798	105.2798	3.5e-10

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Fit	102 ms	123 ms	1.2× faster
Predict	1.01 ms	4.50 ms	4.5× faster
Save model to JSON	23 µs	506 µs	22× faster
Load model from JSON	261 µs	2.39 ms	9.2× faster
Load and predict in a new process	158 ms	1.59 s	10× faster

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "logistic-python-reconciliation-v1"
  },
  "fitRequest": "growth-only",
  "configuration": {
    "growth": "logistic",
    "scaling": "absmax",
    "seasonalityMode": "additive",
    "holidaysMode": "additive",
    "changepoints": {
      "mode": "auto",
      "count": 25,
      "range": 0.8
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [
      {
        "name": "weekly",
        "periodDays": 7,
        "fourierOrder": 3,
        "priorScale": 10
      }
    ],
    "events": [],
    "regressors": []
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}