(Question) Tweak/strip-out Prophet to run linear regression (or some deterministic model; or at least remove randomization)

[tincho-mehaudy]

I noticed that Prophet does not always return the exact same result, seems some randomization is happening even if mcmc is disabled. Below's the Prophet default settings:

m = Prophet(
    growth='linear'
    , changepoints=None
    , n_changepoints=25
    , changepoint_range=0.8
    , yearly_seasonality='auto'
    , weekly_seasonality='auto'
    , daily_seasonality='auto'
    , holidays=None
    , seasonality_mode='additive'
    , seasonality_prior_scale=10.0
    , holidays_prior_scale=10.0
    , changepoint_prior_scale=0.05
    , mcmc_samples=0
    , interval_width=0.8
    , uncertainty_samples=1000
    , stan_backend=None
)

I would like to tweak/strip it in order to make it perform a simple linear regression (or some other deterministic simple model) and then start adding complexity.

Any ideas?

[AlexandroLuis]

Yes, prophet by default have randomness due to it's bayesian essence. It happens even if MCMC is disabled, so consider Disable Seasonalities, Set Holidays to None, Set Uncertainty to Zero so prophet should behave and perform a simple linear regression, you should consider use a dedicated linear regression tool instead.