Ranges with mixed effects

match ← All flat growth benchmarks

80% ranges from 512 simulated paths for every past and future date.

Input

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

Output

Largest difference: 2.3e-7 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-04-06	0.2150	0.2150	2.7e-8
2020-04-07	7.1321	7.1321	1.3e-7
2020-04-08	36.2298	36.2298	2.3e-7
2020-04-09	62.8870	62.8870	2.3e-7
2020-04-10	35.0026	35.0026	1.1e-7
2020-04-11	50.4761	50.4761	3.4e-8
2020-04-12	19.5153	19.5153	1.8e-8
2020-04-13	0.2150	0.2150	2.7e-8
2020-04-14	8.3336	8.3336	1.4e-7
2020-04-15	35.0026	35.0026	1.1e-7
2020-04-16	62.8870	62.8870	2.3e-7
2020-04-17	69.7642	69.7642	1.4e-7
2020-04-18	50.4761	50.4761	3.4e-8
2020-04-19	19.5153	19.5153	1.8e-8
2020-04-20	36.2041	36.2041	1.2e-7
2020-04-21	7.1321	7.1321	1.3e-7
2020-04-22	35.0283	35.0283	2.2e-7
2020-04-23	62.8870	62.8870	2.3e-7
2020-04-24	69.7642	69.7642	1.4e-7
2020-04-25	35.0026	35.0026	1.1e-7
2020-04-26	20.7168	20.7168	1.0e-8
2020-04-27	0.2150	0.2150	2.7e-8
2020-04-28	7.1321	7.1321	1.3e-7
2020-04-29	35.0283	35.0283	2.2e-7

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Convert input rows	44 µs	498 µs	11× faster
Save model to JSON	22 µs	660 µs	30× faster
Load model from JSON	218 µs	3.56 ms	16× faster
Simulate uncertainty	1.10 ms	73.6 ms	67× faster
First simulation in a new process	178 ms	1.51 s	8.5× faster
Load and simulate in a new process	162 ms	1.51 s	9.4× faster

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "flat-map-mixed-v1"
  },
  "configuration": {
    "growth": "flat",
    "scaling": "minmax",
    "seasonalityMode": "additive",
    "holidaysMode": "additive",
    "changepoints": {
      "mode": "auto",
      "count": 25,
      "range": 0.8
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [
      {
        "name": "weekly-custom",
        "periodDays": 7,
        "fourierOrder": 2,
        "priorScale": 10,
        "conditionName": "active",
        "mode": "multiplicative"
      }
    ],
    "events": [
      {
        "name": "campaign",
        "date": "2020-01-22",
        "lowerWindowDays": 0,
        "upperWindowDays": 0,
        "priorScale": 10
      }
    ],
    "regressors": [
      {
        "name": "promotion",
        "priorScale": 10,
        "standardization": "never"
      }
    ]
  },
  "effectOptimizer": {
    "algorithm": "lbfgs",
    "maxIterations": 10000,
    "fallback": "none",
    "lbfgs": {
      "historySize": 5,
      "initAlpha": 0.001,
      "tolObj": 1e-12,
      "tolRelObj": 10000,
      "tolGrad": 1e-8,
      "tolRelGrad": 10000000,
      "tolParam": 1e-8
    }
  },
  "pythonOptimizer": {
    "algorithm": "LBFGS",
    "maxIterations": 10000,
    "newtonFallback": false
  },
  "uncertainty": {
    "seed": 19,
    "samples": 512,
    "intervalWidth": 0.8,
    "output": "intervals",
    "algorithm": "scalar-continuous-time",
    "evidence": "docs/validation/uncertainty.md#evidence-classes",
    "samplerMemoryLimitBytes": 67108864
  }
}