How to get the predict values from many yhats

[saksonita]

future = m.make_future_dataframe(data, periods=317)
forecast_future = m.predict(future)
fig = m.plot(forecast_future)

by writing the above code, I've got many yhats (i=365). The plot looks fine, but how can I get those values from the plot? I tried yhat1 but there were many NaN values same as the rest.

I am looking forward to your response.

[judussoari]

Hi @saksonita , thanks for opening a new discussion.
You probably used auto regression n_lags>0. In that case, it's perfectly normal to have n_lags NaN values at the beginning of the yhat columns, as no prediction can be made with less than the specified n_lags datapoints taken as an input. Your forecast_future dataframe contains the predictions you plotted.
Hope this helps. If not, feel free to post a reproducible code section, so that I can have a closer look.
Best
Julio

[saksonita]

Thank for your fast response. Is there any way to get those predicted values? Or do I have to loop to get them?
Thank again and I am looking forward to your response?

[xeonvs]

Hi. You may use something like that:
yhat = prediction[[f'yhat{i}' for i in range(1, predict_period)]].mean(axis=1)

[Alan-Wolf-VN]

If you want to get the latest forecast, you can use something like this:
latest = m.get_latest_forecast(forecast, include_history_data=True)

It will fetch the latest yhat for each future timestamp