Observed spillovers and unseen future names

match ← All flat growth benchmarks

96 daily rows from 2010; sparse prediction years 2021 and 2060. December 31's observed New Year's Day is retained while unseen Juneteenth names are ignored.

Input

benchmark/inputs/v5/country-boundary-96.json: 96 history rows and 4 forecast rows.

Output

Largest difference: 5.2e-12 (< 0.0001% of the input's range).

Every forecast row from both libraries.

Date	Effect Prophet	Python Prophet	Difference
2021-06-18	20.0158	20.0158	4.1e-12
2021-06-19	20.0158	20.0158	4.1e-12
2021-12-31	20.0158	20.0158	2.5e-12
2060-01-01	23.0350	23.0350	5.2e-12

Timing

Median of all measured repetitions.

Step	Effect Prophet	Python Prophet	Effect Prophet is
Convert input rows	13 µs	469 µs	35× faster
Fit	23.1 ms	67.1 ms	2.9× faster
Predict	223 µs	5.26 ms	24× faster
Fit and predict	21.9 ms	72.7 ms	3.3× faster
Convert, fit and predict	21.6 ms	73.5 ms	3.4× faster
First forecast in a new process	194 ms	1.74 s	9.0× faster
Save model to JSON	77 µs	561 µs	7.3× faster
Load model from JSON	779 µs	3.36 ms	4.3× faster
Load and predict in a new process	160 ms	1.64 s	10× faster

Settings

The case declaration both libraries ran.

{
  "kind": "map",
  "comparison": {
    "kind": "equivalent-objective",
    "evidenceId": "country-calendar-public-v1"
  },
  "configuration": {
    "growth": "flat",
    "countryHolidays": {
      "country": "US",
      "coverage": "reject-known-gaps",
      "priorScale": 10
    },
    "scaling": "absmax",
    "seasonalityMode": "additive",
    "holidaysMode": "additive",
    "changepoints": {
      "mode": "auto",
      "count": 25,
      "range": 0.8
    },
    "changepointPriorScale": 0.05,
    "seasonalities": [],
    "events": [],
    "regressors": []
  },
  "pythonOptimizer": {
    "algorithm": "Auto",
    "maxIterations": 10000,
    "newtonFallback": true,
    "sigFigs": 12
  }
}