Repeated dates

match ← All logistic growth benchmarks

96 rows over 48 dates, each appearing twice with different values. Every row counts as its own measurement.

Input

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

Output

Largest difference: 2.6e-11 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-01-01	14.1880	14.1880	2.4e-11
2020-01-01	14.1880	14.1880	2.4e-11
2020-01-02	15.2680	15.2680	2.5e-11
2020-01-02	15.2680	15.2680	2.5e-11
2020-01-03	16.4149	16.4149	2.5e-11
2020-01-03	16.4149	16.4149	2.5e-11
2020-01-04	17.6304	17.6304	2.5e-11
2020-01-04	17.6304	17.6304	2.5e-11
2020-01-05	18.9161	18.9161	2.6e-11
2020-01-05	18.9161	18.9161	2.6e-11
2020-01-06	20.2730	20.2730	2.6e-11
2020-01-06	20.2730	20.2730	2.6e-11
2020-01-07	21.7020	21.7020	2.6e-11
2020-01-07	21.7020	21.7020	2.6e-11
2020-01-08	23.2031	23.2031	2.5e-11
2020-01-08	23.2031	23.2031	2.5e-11
2020-01-09	24.7762	24.7762	2.5e-11
2020-01-09	24.7762	24.7762	2.5e-11
2020-01-10	26.4205	26.4205	2.4e-11
2020-01-10	26.4205	26.4205	2.4e-11
2020-01-11	28.1344	28.1344	2.4e-11
2020-01-11	28.1344	28.1344	2.4e-11
2020-01-12	29.9159	29.9159	2.3e-11
2020-01-12	29.9159	29.9159	2.3e-11
2020-01-13	31.7621	31.7621	2.2e-11
2020-01-13	31.7621	31.7621	2.2e-11
2020-01-14	33.6696	33.6696	2.0e-11
2020-01-14	33.6696	33.6696	2.0e-11
2020-01-15	35.6344	35.6344	1.9e-11
2020-01-15	35.6344	35.6344	1.9e-11
2020-01-16	37.6515	37.6515	1.7e-11
2020-01-16	37.6515	37.6515	1.7e-11
2020-01-17	39.7156	39.7156	1.6e-11
2020-01-17	39.7156	39.7156	1.6e-11
2020-01-18	41.8205	41.8205	1.4e-11
2020-01-18	41.8205	41.8205	1.4e-11
2020-01-19	43.9598	43.9598	1.2e-11
2020-01-19	43.9598	43.9598	1.2e-11
2020-01-20	46.1265	46.1265	1.0e-11
2020-01-20	46.1265	46.1265	1.0e-11
2020-01-21	48.3139	48.3139	7.9e-12
2020-01-21	48.3139	48.3139	7.9e-12
2020-01-22	50.5137	50.5137	5.8e-12
2020-01-22	50.5137	50.5137	5.8e-12
2020-01-23	52.7181	52.7181	3.7e-12
2020-01-23	52.7181	52.7181	3.7e-12
2020-01-24	54.9193	54.9193	1.5e-12
2020-01-24	54.9193	54.9193	1.5e-12
2020-01-25	57.1094	57.1094	6.5e-13
2020-01-25	57.1094	57.1094	6.5e-13
2020-01-26	59.2810	59.2810	2.7e-12
2020-01-26	59.2810	59.2810	2.7e-12
2020-01-27	61.4267	61.4267	4.8e-12
2020-01-27	61.4267	61.4267	4.8e-12
2020-01-28	63.5396	63.5396	6.7e-12
2020-01-28	63.5396	63.5396	6.7e-12
2020-01-29	65.6133	65.6133	8.6e-12
2020-01-29	65.6133	65.6133	8.6e-12
2020-01-30	67.6420	67.6420	1.0e-11
2020-01-30	67.6420	67.6420	1.0e-11
2020-01-31	69.6202	69.6202	1.2e-11
2020-01-31	69.6202	69.6202	1.2e-11
2020-02-01	71.5433	71.5433	1.3e-11
2020-02-01	71.5433	71.5433	1.3e-11
2020-02-02	73.4074	73.4074	1.5e-11
2020-02-02	73.4074	73.4074	1.5e-11
2020-02-03	75.2091	75.2091	1.6e-11
2020-02-03	75.2091	75.2091	1.6e-11
2020-02-04	76.9457	76.9457	1.7e-11
2020-02-04	76.9457	76.9457	1.7e-11
2020-02-05	78.6152	78.6152	1.8e-11
2020-02-05	78.6152	78.6152	1.8e-11
2020-02-06	80.2163	80.2163	1.8e-11
2020-02-06	80.2163	80.2163	1.8e-11
2020-02-07	81.7481	81.7481	1.9e-11
2020-02-07	81.7481	81.7481	1.9e-11
2020-02-08	83.2104	83.2104	1.9e-11
2020-02-08	83.2104	83.2104	1.9e-11
2020-02-09	84.6036	84.6036	2.0e-11
2020-02-09	84.6036	84.6036	2.0e-11
2020-02-10	85.9282	85.9282	2.0e-11
2020-02-10	85.9282	85.9282	2.0e-11
2020-02-11	87.1855	87.1855	2.0e-11
2020-02-11	87.1855	87.1855	2.0e-11
2020-02-12	88.3768	88.3768	2.0e-11
2020-02-12	88.3768	88.3768	2.0e-11
2020-02-13	89.5039	89.5039	2.0e-11
2020-02-13	89.5039	89.5039	2.0e-11
2020-02-14	90.5687	90.5687	2.0e-11
2020-02-14	90.5687	90.5687	2.0e-11
2020-02-15	91.5736	91.5736	1.9e-11
2020-02-15	91.5736	91.5736	1.9e-11
2020-02-16	92.5207	92.5207	1.9e-11
2020-02-16	92.5207	92.5207	1.9e-11
2020-02-17	93.4125	93.4125	1.8e-11
2020-02-17	93.4125	93.4125	1.8e-11
2020-02-18	94.2514	94.2514	1.8e-11
2020-02-19	94.9670	94.9670	1.7e-11
2020-02-20	95.7075	95.7075	1.7e-11
2020-02-21	96.3286	96.3286	1.6e-11
2020-02-22	96.9809	96.9809	1.6e-11
2020-02-23	97.5179	97.5179	1.5e-11
2020-02-24	98.0918	98.0918	1.4e-11
2020-02-25	98.5545	98.5545	1.4e-11
2020-02-26	99.0594	99.0594	1.3e-11
2020-02-27	99.4568	99.4568	1.3e-11
2020-02-28	99.9012	99.9012	1.2e-11
2020-02-29	100.2417	100.2417	1.2e-11
2020-03-01	100.6336	100.6336	1.1e-11
2020-03-02	100.9246	100.9246	1.0e-11
2020-03-03	101.2709	101.2709	9.8e-12
2020-03-04	101.5192	101.5192	9.4e-12
2020-03-05	101.8263	101.8263	8.8e-12
2020-03-06	102.0378	102.0378	8.4e-12
2020-03-07	102.3110	102.3110	7.9e-12
2020-03-08	102.4909	102.4909	7.5e-12
2020-03-09	102.7352	102.7352	7.0e-12
2020-03-10	102.8880	102.8880	6.6e-12
2020-03-11	103.1076	103.1076	6.3e-12
2020-03-12	103.2373	103.2373	5.8e-12

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Fit	6.01 ms	16.7 ms	2.8× faster
Predict	688 µs	4.14 ms	6.0× faster
Save model to JSON	26 µs	475 µs	18× faster
Load model from JSON	149 µs	2.41 ms	16× faster
Load and predict in a new process	155 ms	1.59 s	10× 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": [
        "2020-01-20T00:00:00.000Z"
      ]
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [],
    "events": [],
    "regressors": []
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}