1. Learn
    2. /
  2. Courses
    3. /
  3. Designing Forecasting Pipelines for Production
Connected

Exercise

Defining the forecasting models

As a data science consultant, you've been tasked with predicting US hourly electricity demand. Before diving into training and testing, you first need to define your machine learning models: ElasticNet, KNeighborsRegressor, and MLPRegressor. Then, you'll initialize the MLForecast object with key parameters.

To capture temporal dependencies, you'll regress the time series against the last 24 lags and include seasonal features like the day of the week and hour of the day. This setup will form the foundation for building robust forecasts.

Instructions 1/2

undefined XP
    1
    2
  • Define ml_models as a dictionary with model names as keys and their instances as values.