Hello, Thanks for this library. I have a problem where I need to forecast monthly average monthly price change for a product over next few years. I have extra features which I need to include in the model and these are coming from different datasets. These features contain economic, demographic and other region related data. How can I approach this problem and which time series model will be best for this case. Any help is much appreciated ☺️

Hi @Avi. Almost all of the models in

neuralforecast

can incorporate exogenous variables (TFT, NHITS, NBEATSx, LSTM, TCN, etc.)! You simply need to add them in the input dataframe as columns, or in the

static_df

input. Here is a tutorial on how to add exogenous variables: https://nixtla.github.io/neuralforecast/examples/exogenous_variables.html

Hi Cristian, actually I tried experimenting with different models like TFT, NBeats and LSTMs, but not getting good forecasts. The issue is I need to do a long horizon forecasting but I don't have much data since I need to do modelling on a monthly level (around 300 data points). What do you suggest would be the best model for this case