Add custom seasonality in prophet model

38 views Asked by At

I have a dataset with hourly frequency for almost 2 years, what is the best parameter to include in my prophet model for better accuracy and future prediction.

date    Count
2022-01-01 00:00:00 30726592
2022-01-01 01:00:00 29027813
2022-01-01 02:00:00 26950925
2022-01-01 03:00:00 24966358
2022-01-01 04:00:00 23852430
... ...
2023-11-30 19:00:00 32653145
2023-11-30 20:00:00 32997299
2023-11-30 21:00:00 32745758
2023-11-30 22:00:00 30883692
2023-11-30 23:00:00 29626339

I tried:

model = Prophet(
  daily_seasonality = True,
  weekly_seasonality=True,
  yearly_seasonality = True,
  interval_width=0.95)

Is this the correct method?

The goal is to predict in future if Count is having major dip or spike on hourly basis and calculate the accuracy of the model.

0

There are 0 answers