Additive regressor

match ← All logistic growth benchmarks

A promotion regressor that adds to the forecast.

Input

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

Output

Largest difference: 1.1e-11 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-01-01	14.1875	14.1875	9.3e-12
2020-01-02	14.7196	14.7196	9.4e-12
2020-01-03	16.4708	16.4708	1.1e-11
2020-01-04	15.8336	15.8336	9.4e-12
2020-01-05	16.4160	16.4160	9.5e-12
2020-01-06	17.0158	17.0158	9.5e-12
2020-01-07	17.6331	17.6331	9.5e-12
2020-01-08	18.2680	18.2680	9.4e-12
2020-01-09	20.1235	20.1235	1.1e-11
2020-01-10	19.5917	19.5917	9.4e-12
2020-01-11	20.2807	20.2807	9.3e-12
2020-01-12	20.9879	20.9879	9.2e-12
2020-01-13	21.7134	21.7134	9.2e-12
2020-01-14	22.4573	22.4573	9.1e-12
2020-01-15	24.4221	24.4221	1.1e-11
2020-01-16	24.0000	24.0000	8.8e-12
2020-01-17	24.7989	24.7989	8.7e-12
2020-01-18	25.6159	25.6159	8.5e-12
2020-01-19	26.4509	26.4509	8.3e-12
2020-01-20	27.3039	27.3039	8.1e-12
2020-01-21	29.3771	29.3771	9.8e-12
2020-01-22	29.0625	29.0625	7.6e-12
2020-01-23	29.9676	29.9676	7.4e-12
2020-01-24	30.8895	30.8895	7.1e-12
2020-01-25	31.8279	31.8279	6.8e-12
2020-01-26	32.7823	32.7823	6.5e-12
2020-01-27	34.9548	34.9548	8.1e-12
2020-01-28	34.7371	34.7371	5.9e-12
2020-01-29	35.7366	35.7366	5.5e-12
2020-01-30	36.7500	36.7500	5.1e-12
2020-01-31	37.7767	37.7767	4.7e-12
2020-02-01	38.8160	38.8160	4.3e-12
2020-02-02	41.0700	41.0700	5.8e-12
2020-02-03	40.9298	40.9298	3.5e-12
2020-02-04	42.0027	42.0027	3.1e-12
2020-02-05	43.0853	43.0853	2.6e-12
2020-02-06	44.1767	44.1767	2.2e-12
2020-02-07	45.2760	45.2760	1.7e-12
2020-02-08	47.5851	47.5851	3.1e-12
2020-02-09	47.4954	47.4954	7.4e-13
2020-02-10	48.6136	48.6136	2.6e-13
2020-02-11	49.7361	49.7361	2.1e-13
2020-02-12	50.8620	50.8620	7.1e-13
2020-02-13	51.9903	51.9903	1.2e-12
2020-02-14	54.3227	54.3227	2.2e-13
2020-02-15	54.2502	54.2502	2.2e-12
2020-02-16	55.3799	55.3799	2.7e-12
2020-02-17	56.5082	56.5082	3.1e-12
2020-02-18	57.6341	57.6341	3.6e-12
2020-02-19	58.7566	58.7566	4.1e-12
2020-02-20	61.0776	61.0776	2.6e-12
2020-02-21	60.9880	60.9880	5.0e-12
2020-02-22	62.0951	62.0951	5.4e-12
2020-02-23	63.1952	63.1952	5.9e-12
2020-02-24	64.2875	64.2875	6.2e-12
2020-02-25	65.3713	65.3713	6.7e-12
2020-02-26	67.6483	67.6483	5.1e-12
2020-02-27	67.5100	67.5100	7.4e-12
2020-02-28	68.5635	68.5635	7.8e-12
2020-02-29	69.6055	69.6055	8.1e-12
2020-03-01	70.6354	70.6354	8.4e-12
2020-03-02	71.6525	71.6525	8.8e-12
2020-03-03	73.8591	73.8591	7.1e-12
2020-03-04	73.6465	73.6465	9.3e-12
2020-03-05	74.6222	74.6222	9.5e-12
2020-03-06	75.5833	75.5833	9.8e-12
2020-03-07	76.5292	76.5292	1.0e-11
2020-03-08	77.4597	77.4597	1.0e-11
2020-03-09	79.5770	79.5770	8.5e-12
2020-03-10	79.2729	79.2729	1.1e-11
2020-03-11	80.1552	80.1552	1.1e-11
2020-03-12	81.0209	81.0209	1.1e-11
2020-03-13	81.8700	81.8700	1.1e-11
2020-03-14	82.7023	82.7023	1.1e-11
2020-03-15	84.7204	84.7204	9.2e-12
2020-03-16	84.3162	84.3162	1.1e-11
2020-03-17	85.0977	85.0977	1.1e-11
2020-03-18	85.8622	85.8622	1.1e-11
2020-03-19	86.6097	86.6097	1.1e-11
2020-03-20	87.3404	87.3404	1.1e-11
2020-03-21	89.2570	89.2570	9.4e-12
2020-03-22	88.7516	88.7516	1.1e-11
2020-03-23	89.4323	89.4323	1.1e-11
2020-03-24	90.0966	90.0966	1.1e-11
2020-03-25	90.7447	90.7447	1.1e-11
2020-03-26	91.3768	91.3768	1.1e-11
2020-03-27	93.1957	93.1957	9.1e-12
2020-03-28	92.5938	92.5938	1.1e-11
2020-03-29	93.1791	93.1791	1.1e-11
2020-03-30	93.7493	93.7493	1.1e-11
2020-03-31	94.3048	94.3048	1.1e-11
2020-04-01	94.8456	94.8456	1.0e-11
2020-04-02	96.5748	96.5748	8.4e-12
2020-04-03	95.8847	95.8847	1.0e-11
2020-04-04	96.3836	96.3836	1.0e-11
2020-04-05	96.8691	96.8691	9.9e-12
2020-04-06	97.3414	97.3414	9.8e-12
2020-04-07	97.8009	97.8009	9.6e-12
2020-04-08	99.4506	99.4506	7.5e-12
2020-04-09	98.6828	98.6828	9.3e-12
2020-04-10	99.1058	99.1058	9.2e-12
2020-04-11	99.5171	99.5171	9.0e-12
2020-04-12	99.9172	99.9172	8.8e-12
2020-04-13	100.3064	100.3064	8.7e-12
2020-04-14	101.8875	101.8875	6.6e-12
2020-04-15	101.0529	101.0529	8.3e-12
2020-04-16	101.4109	101.4109	8.2e-12
2020-04-17	101.7592	101.7592	8.0e-12
2020-04-18	102.0979	102.0979	7.8e-12
2020-04-19	102.4274	102.4274	7.7e-12
2020-04-20	103.9507	103.9507	5.6e-12
2020-04-21	103.0600	103.0600	7.3e-12
2020-04-22	103.3635	103.3635	7.1e-12
2020-04-23	103.6590	103.6590	7.0e-12
2020-04-24	103.9466	103.9466	6.8e-12
2020-04-25	104.2265	104.2265	6.6e-12
2020-04-26	105.7018	105.7018	4.6e-12
2020-04-27	104.7647	104.7647	6.3e-12
2020-04-28	105.0233	105.0233	6.2e-12
2020-04-29	105.2753	105.2753	6.0e-12

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Fit	7.94 ms	15.8 ms	2.0× faster
Predict	1.08 ms	3.29 ms	3.1× faster
Save model to JSON	28 µs	480 µs	17× faster
Load model from JSON	131 µs	2.44 ms	19× faster
Load and predict in a new process	158 ms	1.43 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": [
        "2020-02-08T00:00:00.000Z"
      ]
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [],
    "events": [],
    "regressors": [
      {
        "name": "promotion",
        "priorScale": 10,
        "standardization": "never"
      }
    ]
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}