FBProphet forecast results are wildly different using the same codes and data

Question

I am using python 3.5 FBProphet forecast results are wildly different using the same codes and data, but the forecasting y^ can be very different between each run using the same code and data sets. The followings are my FBprophet forecasting codes

    m = Prophet(interval_width=1)

    m.fit(df)

    future = m.make_future_dataframe(periods=365)

    future.tail()

    forecast = m.predict(future)

This post mentioned that we can choose different forecasting algos from Newton to L-BFGS, and Newton should provide more repeatable results. However, in the. latest FBProphet that I used, there is no algos options anymore.

https://github.com/facebook/prophet/issues/253

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=0.80,
uncertainty_samples=1000,

Is there any other parameters that can cause this randomness? Thank you!

Answer 1

Try with:

m.fit(algorithm='Newton')