When analyzing data using binary data, other values ​​are additionally created in the middle.

[boozesounds]

m = NeuralProphet(
    n_forecasts=3,
    n_lags=3,
    learning_rate=0.1,
    normalize=False
)
m = m.highlight_nth_step_ahead_of_each_forecast(3)
metrics = m.fit(df, freq="D")

I only put 2 and 3 data. Likewise When you put binary data like 0 and 1, 1 and -1 another value is added to the middle line of the data to start the analysis. How can I fix this situation that occurs when I put in binary time series data?

and this problem occurs when using lag .

[leoniewgnr]

Hi @boozesounds, thanks for your question! Do you have a dataset that you can share which leads to that behaviour or is it only data consisting of 2 and 3?

[judussoari]

@boozesounds I assume the values are imputed. Try setting impute_missing=False when defining m = NeuralProphet().