Additive and multiplicative together

match ← All linear growth benchmarks

A multiplicative switchable weekly pattern with an additive event and regressor, on 96 daily rows.

Input

benchmark/inputs/v1/linear-mixed-components.json: 96 history rows and 24 forecast rows.

Output

Largest difference: 1.8e-10 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-04-06	26.0176	26.0176	1.2e-10
2020-04-07	29.3874	29.3874	1.4e-10
2020-04-08	44.1660	44.1660	1.7e-10
2020-04-09	56.3191	56.3191	1.7e-10
2020-04-10	42.9853	42.9853	1.5e-10
2020-04-11	50.5414	50.5414	1.5e-10
2020-04-12	35.6534	35.6534	1.2e-10
2020-04-13	26.3580	26.3580	1.3e-10
2020-04-14	31.1089	31.1089	1.6e-10
2020-04-15	43.3841	43.3841	1.5e-10
2020-04-16	57.0518	57.0518	1.7e-10
2020-04-17	60.5134	60.5134	1.7e-10
2020-04-18	51.1965	51.1965	1.5e-10
2020-04-19	36.1147	36.1147	1.2e-10
2020-04-20	45.1353	45.1353	1.7e-10
2020-04-21	30.1549	30.1549	1.5e-10
2020-04-22	43.9599	43.9599	1.6e-10
2020-04-23	57.7845	57.7845	1.8e-10
2020-04-24	61.2891	61.2891	1.8e-10
2020-04-25	44.1815	44.1815	1.5e-10
2020-04-26	37.9433	37.9433	1.5e-10
2020-04-27	27.0388	27.0388	1.3e-10
2020-04-28	30.5386	30.5386	1.5e-10
2020-04-29	44.5183	44.5183	1.6e-10

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Convert input rows	51 µs	522 µs	10× faster
Fit	5.38 ms	23.8 ms	4.4× faster
Predict	428 µs	4.96 ms	12× faster
Fit and predict	4.91 ms	31.8 ms	6.5× faster
Convert, fit and predict	4.87 ms	31.6 ms	6.5× faster
First forecast in a new process	176 ms	1.47 s	8.4× faster
Save model to JSON	31 µs	625 µs	20× faster
Load model from JSON	141 µs	3.60 ms	25× faster
Load and predict in a new process	166 ms	1.42 s	8.5× faster

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "linear-output-first-v4"
  },
  "configuration": {
    "growth": "linear",
    "scaling": "absmax",
    "seasonalityMode": "multiplicative",
    "holidaysMode": "additive",
    "changepoints": {
      "mode": "explicit",
      "timestamps": [
        "2020-02-10T00:00:00.000Z"
      ]
    },
    "changepointPriorScale": 0.2,
    "seasonalities": [
      {
        "name": "weekly-custom",
        "periodDays": 7,
        "fourierOrder": 2,
        "priorScale": 10,
        "conditionName": "active"
      }
    ],
    "events": [
      {
        "name": "campaign",
        "date": "2020-01-22",
        "lowerWindowDays": 0,
        "upperWindowDays": 0,
        "priorScale": 10
      }
    ],
    "regressors": [
      {
        "name": "promotion",
        "priorScale": 10,
        "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
  }
}