Everything at once, two years

match ← All linear growth benchmarks

768 daily rows with automatic changepoints, two seasonal patterns, six events and four regressors.

Input

benchmark/inputs/v1/map-mixed-features-automatic-large.json: 768 history rows and 128 forecast rows.

Output

Largest difference: 0.018 (0.038% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2022-02-07	46.5893	46.5852	0.0041
2022-02-08	47.3077	47.3037	0.0040
2022-02-09	47.5411	47.5370	0.0040
2022-02-10	47.9357	47.9317	0.0040
2022-02-11	49.5180	49.5144	0.0036
2022-02-12	49.5948	49.5911	0.0037
2022-02-13	50.5748	50.5691	0.0057
2022-02-14	49.6492	49.6433	0.0059
2022-02-15	50.3784	50.3722	0.0063
2022-02-16	49.3841	49.3795	0.0046
2022-02-17	49.4535	49.4490	0.0045
2022-02-18	49.2992	49.2950	0.0042
2022-02-19	48.8937	48.8895	0.0042
2022-02-20	48.1631	48.1589	0.0042
2022-02-21	46.1708	46.1668	0.0040
2022-02-22	48.6500	48.6441	0.0060
2022-02-23	50.3412	50.3350	0.0062
2022-02-24	51.0689	51.0625	0.0063
2022-02-25	49.7884	49.7838	0.0045
2022-02-26	50.1739	50.1694	0.0045
2022-02-27	49.9278	49.9235	0.0044
2022-02-28	49.1691	49.1649	0.0042
2022-03-01	49.7516	49.7472	0.0044
2022-03-02	50.8501	50.8457	0.0044
2022-03-03	53.1047	53.0990	0.0057
2022-03-04	51.8394	51.8335	0.0059
2022-03-05	51.9578	51.9517	0.0061
2022-03-06	49.4631	49.4585	0.0046
2022-03-07	49.1809	49.1760	0.0049
2022-03-08	49.9313	49.9262	0.0050
2022-03-09	51.1905	51.1854	0.0051
2022-03-10	55.5552	55.5492	0.0061
2022-03-11	52.1450	52.1443	0.00074
2022-03-12	55.1835	55.1885	0.0051
2022-03-13	57.6355	57.6278	0.0077
2022-03-14	52.3855	52.3795	0.0060
2022-03-15	51.3122	51.3077	0.0045
2022-03-16	52.1808	52.1759	0.0049
2022-03-17	50.8839	50.8787	0.0052
2022-03-18	50.6977	50.6926	0.0052
2022-03-19	50.6932	50.6881	0.0051
2022-03-20	49.4067	49.4017	0.0050
2022-03-21	50.5980	50.5915	0.0065
2022-03-22	51.0491	51.0425	0.0066
2022-03-23	51.1985	51.1920	0.0065
2022-03-24	49.6328	49.6282	0.0047
2022-03-25	49.8497	49.8451	0.0046
2022-03-26	50.0945	50.0898	0.0047
2022-03-27	49.8799	49.8750	0.0049
2022-03-28	49.2677	49.2628	0.0049
2022-03-29	49.9483	49.9434	0.0049
2022-03-30	50.5281	50.5176	0.011
2022-03-31	48.8070	48.7889	0.018
2022-04-01	52.2733	52.2667	0.0065
2022-04-02	50.2087	50.2045	0.0042
2022-04-03	49.2698	49.2656	0.0043
2022-04-04	48.7481	48.7436	0.0045
2022-04-05	49.3442	49.3393	0.0049
2022-04-06	50.5639	50.5587	0.0052
2022-04-07	51.1417	51.1364	0.0053
2022-04-08	53.7062	53.6992	0.0070
2022-04-09	53.9648	53.9580	0.0068
2022-04-10	52.4408	52.4342	0.0065
2022-04-11	49.7563	49.7516	0.0048
2022-04-12	50.2471	50.2424	0.0048
2022-04-13	51.3120	51.3071	0.0048
2022-04-14	51.6480	51.6431	0.0049
2022-04-15	51.7430	51.7380	0.0050
2022-04-16	52.0423	52.0370	0.0053
2022-04-17	52.8585	52.8513	0.0072
2022-04-18	51.7517	51.7444	0.0073
2022-04-19	51.9584	51.9511	0.0073
2022-04-20	51.7062	51.7009	0.0053
2022-04-21	52.1354	52.1301	0.0053
2022-04-22	51.2555	51.2506	0.0050
2022-04-23	51.2791	51.2743	0.0049
2022-04-24	50.9515	50.9465	0.0050
2022-04-25	48.9940	48.9887	0.0054
2022-04-26	51.7224	51.7150	0.0074
2022-04-27	53.1850	53.1774	0.0076
2022-04-28	54.2569	54.2494	0.0075
2022-04-29	52.8437	52.8382	0.0055
2022-04-30	53.0372	53.0318	0.0053
2022-05-01	52.3164	52.3110	0.0053
2022-05-02	51.7163	51.7114	0.0049
2022-05-03	52.2148	52.2098	0.0050
2022-05-04	53.3047	53.2995	0.0052
2022-05-05	55.6085	55.6013	0.0072
2022-05-06	56.3338	56.3264	0.0074
2022-05-07	56.6657	56.6582	0.0075
2022-05-08	52.9204	52.9144	0.0060
2022-05-09	52.3415	52.3356	0.0060
2022-05-10	52.6573	52.6515	0.0058
2022-05-11	53.7158	53.7100	0.0059
2022-05-12	55.0623	55.0568	0.0055
2022-05-13	55.2339	55.2286	0.0053
2022-05-14	57.4371	57.4300	0.0072
2022-05-15	56.5342	56.5269	0.0074
2022-05-16	55.5437	55.5361	0.0076
2022-05-17	54.0251	54.0189	0.0061
2022-05-18	55.1304	55.1240	0.0064
2022-05-19	55.1005	55.0942	0.0063
2022-05-20	55.1093	55.1035	0.0058
2022-05-21	53.7155	53.7096	0.0059
2022-05-22	52.1978	52.1920	0.0059
2022-05-23	52.9182	52.9107	0.0075
2022-05-24	53.3455	53.3379	0.0076
2022-05-25	54.4649	54.4571	0.0078
2022-05-26	53.5451	53.5389	0.0062
2022-05-27	53.9458	53.9396	0.0062
2022-05-28	54.2307	54.2246	0.0061
2022-05-29	53.5561	53.5500	0.0060
2022-05-30	53.0984	53.0925	0.0059
2022-05-31	53.5633	53.5574	0.0060
2022-06-01	56.2425	56.2351	0.0073
2022-06-02	56.4150	56.4078	0.0072
2022-06-03	55.3851	55.3778	0.0073
2022-06-04	53.5581	53.5523	0.0057
2022-06-05	52.7754	52.7694	0.0060
2022-06-06	52.0527	52.0465	0.0062
2022-06-07	53.1091	53.1026	0.0065
2022-06-08	54.3485	54.3420	0.0066
2022-06-09	54.1754	54.1693	0.0061
2022-06-10	56.0820	56.0742	0.0077
2022-06-11	57.3881	57.3806	0.0076
2022-06-12	56.5552	56.5478	0.0074
2022-06-13	53.8341	53.8284	0.0056
2022-06-14	54.2716	54.2657	0.0058

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Convert input rows	195 µs	1.15 ms	5.9× faster
Fit	14.6 ms	38.0 ms	2.6× faster
Predict	965 µs	6.54 ms	6.8× faster
Fit and predict	11.8 ms	44.3 ms	3.8× faster
Convert, fit and predict	12.1 ms	45.5 ms	3.8× faster
First forecast in a new process	217 ms	1.48 s	6.8× faster
Save model to JSON	37 µs	1.75 ms	47× faster
Load model from JSON	347 µs	5.76 ms	17× faster
Load and predict in a new process	188 ms	1.44 s	7.6× faster

Settings

The case declaration both libraries ran.

{
  "kind": "linear-map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "linear-output-first-v4"
  },
  "configuration": {
    "effectMap": "omitted",
    "changepoints": {
      "mode": "auto",
      "count": 25,
      "range": 0.8
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [
      {
        "name": "weekly-shared",
        "periodDays": 7,
        "fourierOrder": 2,
        "priorScale": 10
      },
      {
        "name": "monthly-active",
        "periodDays": 30.5,
        "fourierOrder": 3,
        "priorScale": 8,
        "conditionName": "active"
      }
    ],
    "events": [
      {
        "name": "release",
        "date": "2020-04-10",
        "lowerWindowDays": -1,
        "upperWindowDays": 2,
        "priorScale": 10
      },
      {
        "name": "release",
        "date": "2020-10-27",
        "lowerWindowDays": -1,
        "upperWindowDays": 2,
        "priorScale": 10
      },
      {
        "name": "release",
        "date": "2022-03-11",
        "lowerWindowDays": -1,
        "upperWindowDays": 2,
        "priorScale": 10
      },
      {
        "name": "campaign",
        "date": "2020-04-11",
        "lowerWindowDays": -1,
        "upperWindowDays": 0,
        "priorScale": 6
      },
      {
        "name": "campaign",
        "date": "2021-05-15",
        "lowerWindowDays": -1,
        "upperWindowDays": 0,
        "priorScale": 6
      },
      {
        "name": "campaign",
        "date": "2022-03-31",
        "lowerWindowDays": -1,
        "upperWindowDays": 0,
        "priorScale": 6
      }
    ],
    "regressors": [
      {
        "name": "binary-auto",
        "priorScale": 10,
        "standardization": "auto"
      },
      {
        "name": "numeric-auto",
        "priorScale": 10,
        "standardization": "auto"
      },
      {
        "name": "numeric-always",
        "priorScale": 7,
        "standardization": "always"
      },
      {
        "name": "numeric-never",
        "priorScale": 5,
        "standardization": "never"
      }
    ]
  },
  "effectOptimizer": {
    "algorithm": "auto",
    "maxIterations": 10000,
    "fallback": "newton",
    "lbfgs": {
      "historySize": 5,
      "initAlpha": 0.001,
      "tolObj": 1e-12,
      "tolRelObj": 10000,
      "tolGrad": 1e-8,
      "tolRelGrad": 10000000,
      "tolParam": 1e-8
    }
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}