Weekly pattern, absmax scaling

match ← All flat growth benchmarks

96 daily rows around one level with a custom weekly seasonality.

Input

benchmark/inputs/v1/flat-seasonal.json: 96 history rows and 24 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
2020-04-06	33.6351	33.6351	1.1e-11
2020-04-07	33.9097	33.9097	1.1e-11
2020-04-08	35.0039	35.0039	1.1e-11
2020-04-09	36.0959	36.0959	1.1e-11
2020-04-10	36.3665	36.3665	1.1e-11
2020-04-11	35.6084	35.6084	1.1e-11
2020-04-12	34.3909	34.3909	1.1e-11
2020-04-13	33.6351	33.6351	1.1e-11
2020-04-14	33.9097	33.9097	1.1e-11
2020-04-15	35.0039	35.0039	1.1e-11
2020-04-16	36.0959	36.0959	1.1e-11
2020-04-17	36.3665	36.3665	1.1e-11
2020-04-18	35.6084	35.6084	1.1e-11
2020-04-19	34.3909	34.3909	1.1e-11
2020-04-20	33.6351	33.6351	1.1e-11
2020-04-21	33.9097	33.9097	1.1e-11
2020-04-22	35.0039	35.0039	1.1e-11
2020-04-23	36.0959	36.0959	1.1e-11
2020-04-24	36.3665	36.3665	1.1e-11
2020-04-25	35.6084	35.6084	1.1e-11
2020-04-26	34.3909	34.3909	1.1e-11
2020-04-27	33.6351	33.6351	1.1e-11
2020-04-28	33.9097	33.9097	1.1e-11
2020-04-29	35.0039	35.0039	1.1e-11

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Fit	155 ms	603 ms	3.9× faster
Predict	222 µs	3.06 ms	14× faster
Save model to JSON	29 µs	436 µs	15× faster
Load model from JSON	245 µs	2.10 ms	8.6× faster
Load and predict in a new process	157 ms	1.43 s	9.1× faster

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "flat-growth-reduced-v1"
  },
  "configuration": {
    "growth": "flat",
    "scaling": "absmax",
    "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
      }
    ],
    "events": [],
    "regressors": []
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}