(Part of a model building learning session series.)
This session provides the fundamentals for tackling demand forecasting use cases. A good example of a demand forecasting use case is predicting sales for a chain of stores across a hundred SKUs.
First we explain how to understand the business problem; this guides how you set up the problem and determine what data is available. Next, we consider the features you may want to include and the roles of calendar events, proper data prep, and effective partitioning.
After modeling setup is complete, we briefly explain the dominant modeling approaches for demand forecasting—from classic ARIMA to LSTMs to hierarchical modeling. We conclude the session by emphasizing the importance of explainable time series models. Join our hosts as they discuss approaches for demand forecasting use cases, and get your own questions answered.
Hosts
- Jess Lin (DataRobot, Data Scientist)
- Tony Martin (DataRobot, Data Scientist)
- Rajiv Shah (DataRobot, Data Scientist)
- Jack Jablonski (DataRobot, AI Success Manager)
More Information
Let us know what you think!
Have questions not answered during the learning session? What to continue your conversation with Jess, Tony, and Rajiv? Just Post Your Comment here or send email to learning_sessions@datarobot.com. We look forward to hearing from you!
日本語
