Problem with make_future_dataframe()

[Mo-Abdelhameed]

I am running the following code:

import pandas as pd
df = pd.read_csv("https://github.com/ourownstory/neuralprophet-data/raw/main/kaggle-energy/datasets/tutorial04.csv")
df.head()

from neuralprophet import NeuralProphet, set_log_level
set_log_level("ERROR")
m = NeuralProphet(
# Disable trend changepoints
n_changepoints=10,
# Disable seasonality components
yearly_seasonality=True,
weekly_seasonality=True,
daily_seasonality=True,
# Add the autogression
n_lags=10,
epochs=1,
n_forecasts=5
)
m.add_lagged_regressor("temperature")
metrics = m.fit(df)

df_future = m.make_future_dataframe(df=df, periods=10, n_historic_predictions=2)

The future dataframe looks like this

I am specifying periods = 10 but the future data points created is 5 which is equal to n_forecasts parameter. What is the use of periods then? Also the n_historic_predictions does not seem to be working?

When i run this line afterwards

forecast = m.predict(df_future)

this is what i get:

I am having a problem understanding what the output means and why does it have NaN values. Can someone help me please?

What I want to do on general is to forecast a series and add lagged and future regressors. I have a problem with lagged regressors because i get results like the one i am showing and i do not understand what it means?

Thanks!

[rafiuddinkhan]

Yes it seems to be confusing in latest version but you can try these:

You get the latest forecast from the predicted values.

df_forecast = m.get_latest_forecast(forecast)

or you can try:

def exctract_yhat(forecasts, size):
  columns = forecasts.columns[3:]
  newframe = forecasts[['ds', 'yhat1']].iloc[-size:].copy()
  for col in columns:
      if 'yhat' in col:
         newframe['yhat1'] = newframe['yhat1'].fillna(forecasts[col])
  return newframe

p=exctract_yhat(forecast,5)

For details look:
#1541

[ourownstory]

Maybe this tutorial may be of help: https://neuralprophet.com/how-to-guides/feature-guides/collect_predictions.html
have a closer look at the 'raw' format.

[suntochukwu]

@Mo-Abdelhameed could you explain or share the code on how you solved this? I'm having the same isssue.