No changepoints, loose prior, minmax scaling

match ← All logistic growth benchmarks

The same loose-prior fit, scaled between the lowest floor and the highest ceiling.

Input

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

Output

Largest difference: 1.4e-11 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-01-01	15.1103	15.1103	2.0e-12
2020-01-02	15.1577	15.1577	2.1e-12
2020-01-03	15.2076	15.2076	2.2e-12
2020-01-04	15.2602	15.2602	2.3e-12
2020-01-05	15.3154	15.3154	2.4e-12
2020-01-06	15.3736	15.3736	2.5e-12
2020-01-07	15.4348	15.4348	2.6e-12
2020-01-08	15.4993	15.4993	2.7e-12
2020-01-09	15.5671	15.5671	2.8e-12
2020-01-10	15.6384	15.6384	2.9e-12
2020-01-11	15.7134	15.7134	3.0e-12
2020-01-12	15.7924	15.7924	3.2e-12
2020-01-13	15.8754	15.8754	3.3e-12
2020-01-14	15.9627	15.9627	3.4e-12
2020-01-15	16.0546	16.0546	3.6e-12
2020-01-16	16.1511	16.1511	3.7e-12
2020-01-17	16.2527	16.2527	3.9e-12
2020-01-18	16.3594	16.3594	4.0e-12
2020-01-19	16.4716	16.4716	4.2e-12
2020-01-20	16.5896	16.5896	4.3e-12
2020-01-21	16.7135	16.7135	4.5e-12
2020-01-22	16.8437	16.8437	4.7e-12
2020-01-23	16.9805	16.9805	4.8e-12
2020-01-24	17.1242	17.1242	5.0e-12
2020-01-25	17.2751	17.2751	5.2e-12
2020-01-26	17.4336	17.4336	5.4e-12
2020-01-27	17.6000	17.6000	5.6e-12
2020-01-28	17.7746	17.7746	5.8e-12
2020-01-29	17.9578	17.9578	6.0e-12
2020-01-30	18.1501	18.1501	6.2e-12
2020-01-31	18.3518	18.3518	6.4e-12
2020-02-01	18.5634	18.5634	6.7e-12
2020-02-02	18.7852	18.7852	6.9e-12
2020-02-03	19.0177	19.0177	7.1e-12
2020-02-04	19.2614	19.2614	7.4e-12
2020-02-05	19.5166	19.5166	7.6e-12
2020-02-06	19.7840	19.7840	7.8e-12
2020-02-07	20.0639	20.0639	8.1e-12
2020-02-08	20.3569	20.3569	8.3e-12
2020-02-09	20.6636	20.6636	8.6e-12
2020-02-10	20.9843	20.9843	8.8e-12
2020-02-11	21.3196	21.3196	9.1e-12
2020-02-12	21.6701	21.6701	9.4e-12
2020-02-13	22.0364	22.0364	9.6e-12
2020-02-14	22.4189	22.4189	9.9e-12
2020-02-15	22.8182	22.8182	1.0e-11
2020-02-16	23.2348	23.2348	1.0e-11
2020-02-17	23.6694	23.6694	1.1e-11
2020-02-18	24.1224	24.1224	1.1e-11
2020-02-19	24.5945	24.5945	1.1e-11
2020-02-20	25.0861	25.0861	1.1e-11
2020-02-21	25.5977	25.5977	1.2e-11
2020-02-22	26.1300	26.1300	1.2e-11
2020-02-23	26.6833	26.6833	1.2e-11
2020-02-24	27.2582	27.2582	1.2e-11
2020-02-25	27.8552	27.8552	1.3e-11
2020-02-26	28.4746	28.4746	1.3e-11
2020-02-27	29.1169	29.1169	1.3e-11
2020-02-28	29.7824	29.7824	1.3e-11
2020-02-29	30.4715	30.4715	1.3e-11
2020-03-01	31.1845	31.1845	1.3e-11
2020-03-02	31.9216	31.9216	1.4e-11
2020-03-03	32.6830	32.6830	1.4e-11
2020-03-04	33.4689	33.4689	1.4e-11
2020-03-05	34.2793	34.2793	1.4e-11
2020-03-06	35.1143	35.1143	1.4e-11
2020-03-07	35.9738	35.9738	1.4e-11
2020-03-08	36.8578	36.8578	1.4e-11
2020-03-09	37.7659	37.7659	1.4e-11
2020-03-10	38.6981	38.6981	1.4e-11
2020-03-11	39.6539	39.6539	1.4e-11
2020-03-12	40.6330	40.6330	1.4e-11
2020-03-13	41.6348	41.6348	1.4e-11
2020-03-14	42.6587	42.6587	1.4e-11
2020-03-15	43.7041	43.7041	1.3e-11
2020-03-16	44.7702	44.7702	1.3e-11
2020-03-17	45.8562	45.8562	1.3e-11
2020-03-18	46.9612	46.9612	1.3e-11
2020-03-19	48.0842	48.0842	1.2e-11
2020-03-20	49.2241	49.2241	1.2e-11
2020-03-21	50.3797	50.3797	1.2e-11
2020-03-22	51.5498	51.5498	1.1e-11
2020-03-23	52.7331	52.7331	1.1e-11
2020-03-24	53.9283	53.9283	1.1e-11
2020-03-25	55.1340	55.1340	1.0e-11
2020-03-26	56.3487	56.3487	9.7e-12
2020-03-27	57.5709	57.5709	9.3e-12
2020-03-28	58.7991	58.7991	8.8e-12
2020-03-29	60.0317	60.0317	8.3e-12
2020-03-30	61.2672	61.2672	7.8e-12
2020-03-31	62.5039	62.5039	7.3e-12
2020-04-01	63.7403	63.7403	6.8e-12
2020-04-02	64.9749	64.9749	6.3e-12
2020-04-03	66.2059	66.2059	5.7e-12
2020-04-04	67.4320	67.4320	5.2e-12
2020-04-05	68.6515	68.6515	4.6e-12
2020-04-06	69.8631	69.8631	4.1e-12
2020-04-07	71.0653	71.0653	3.6e-12
2020-04-08	72.2567	72.2567	3.1e-12
2020-04-09	73.4360	73.4360	2.5e-12
2020-04-10	74.6020	74.6020	2.0e-12
2020-04-11	75.7535	75.7535	1.5e-12
2020-04-12	76.8894	76.8894	1.0e-12
2020-04-13	78.0086	78.0086	6.0e-13
2020-04-14	79.1101	79.1101	1.4e-13
2020-04-15	80.1932	80.1932	3.1e-13
2020-04-16	81.2571	81.2571	7.4e-13
2020-04-17	82.3009	82.3009	1.1e-12
2020-04-18	83.3241	83.3241	1.5e-12
2020-04-19	84.3262	84.3262	1.9e-12
2020-04-20	85.3066	85.3066	2.2e-12
2020-04-21	86.2650	86.2650	2.6e-12
2020-04-22	87.2011	87.2011	2.8e-12
2020-04-23	88.1146	88.1146	3.1e-12
2020-04-24	89.0053	89.0053	3.4e-12
2020-04-25	89.8733	89.8733	3.6e-12
2020-04-26	90.7183	90.7183	3.9e-12
2020-04-27	91.5405	91.5405	4.1e-12
2020-04-28	92.3400	92.3400	4.3e-12
2020-04-29	93.1168	93.1168	4.4e-12

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Fit	15.2 ms	20.5 ms	1.3× faster
Predict	727 µs	2.93 ms	4.0× faster
Save model to JSON	24 µs	449 µs	19× faster
Load model from JSON	169 µs	2.15 ms	13× faster
Load and predict in a new process	154 ms	1.69 s	11× faster

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "logistic-python-reconciliation-v1"
  },
  "configuration": {
    "growth": "logistic",
    "scaling": "minmax",
    "seasonalityMode": "additive",
    "holidaysMode": "additive",
    "changepoints": {
      "mode": "explicit",
      "timestamps": []
    },
    "changepointPriorScale": 10,
    "seasonalities": [],
    "events": [],
    "regressors": []
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}