This is a Bayesian model so it of course provides a probability distribution for both the parameters and the forecasted values. P(model | historic_data) = P(hi

This is a Bayesian model so it of course provides a probability distribution for both the parameters and the forecasted values. P(model | historic_data) = P(historic_data | model) P(model) / (sum(model) P(historic_data | model) P(model)) P(future_data | historic_data) = sum(model) P(future_data | model) P(model | historic_data) = (sum(model) P(future_data | model) P(historic_data | model) P(model)) / (sum(model) P(historic_data | model) P(model))