Forecasting with ML Models
As a data science consultant, your task is to predict US hourly electricity demand. In the previous task, you cleaned and prepared the data. Now, it's time to use machine learning models to build your forecast.
We previously covered the statsforecast workflow, and now you'll apply the same principles using mlforecast.
The train and test datasets, as well as models (LGBMRegressor(), XGBRegressor(), LinearRegression()), are preloaded.
The MLForecast class has been imported from the mlforecast package, ready to use. Let's build your forecast!
Diese Übung ist Teil des Kurses
Designing Forecasting Pipelines for Production
Interaktive Übung
Vervollständige den Beispielcode, um diese Übung erfolgreich abzuschließen.
# Define the ML models to use in the forecast
ml_models = [____, XGBRegressor(), LinearRegression()]
# Set up the MLForecast object with models, frequency, lags, and date features
mlf = ____(
models= ____, # Use the ML models
freq='____', # Set the frequency of your data (hourly)
lags=list(range(1, 24)), # Define time lags as features (previous 24 hours)
date_features=['year', 'month', 'day', 'dayofweek', 'quarter', 'week', 'hour'])