Ranges over past dates

match ← All linear growth benchmarks

80% ranges from 128 simulated paths on 8 dates inside the history.

Input

benchmark/inputs/v1/linear-offset-scaling.json: 96 history rows and 24 forecast rows.

Output

Largest difference: 1.2e-11 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2020-03-29	109.2537	109.2537	1.2e-11
2020-03-30	109.2406	109.2406	1.2e-11
2020-03-31	109.4134	109.4134	1.2e-11
2020-04-01	109.7283	109.7283	1.2e-11
2020-04-02	110.0413	110.0413	1.2e-11
2020-04-03	110.2101	110.2101	1.2e-11
2020-04-04	110.1943	110.1943	1.2e-11
2020-04-05	110.0950	110.0950	1.2e-11

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Convert input rows	16 µs	472 µs	29× faster
Save model to JSON	27 µs	431 µs	16× faster
Load model from JSON	175 µs	2.50 ms	14× faster
Simulate uncertainty	331 µs	19.8 ms	Not compared: different work
First simulation in a new process	175 ms	1.48 s	Not compared: different work
Load and simulate in a new process	163 ms	1.46 s	Not compared: different work

Some steps aren't compared, because the two libraries' public calls produce different outputs. Their times show what each call costs, not which library is faster.

• Simulate uncertainty, First simulation in a new process, Load and simulate in a new process: Python's predict(vectorized=False) also builds the point forecast and every component, while Effect Prophet's predictUncertainty returns only the ranges.

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "linear-output-first-v4"
  },
  "configuration": {
    "growth": "linear",
    "scaling": "absmax",
    "seasonalityMode": "additive",
    "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
      }
    ],
    "events": [],
    "regressors": []
  },
  "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
  },
  "uncertainty": {
    "seed": 19,
    "samples": 128,
    "intervalWidth": 0.8,
    "output": "intervals",
    "algorithm": "scalar-continuous-time",
    "evidence": "docs/validation/uncertainty.md#evidence-classes",
    "samplerMemoryLimitBytes": 67108864
  }
}