Tutorial 2A: Time Series Recipe Tutorial - Retail Sales Forecasting

In this tutorial, you will learn what a time series experiment is, concept behind time series, how to set a time series experiment on Driverless AI, time series and more. 

Key:

Complete
Failed
Available
Locked
Objective
Select the "Read" button to begin.
Select the "Read" button to begin. Explore the goal of this tutorial
Prerequisites
Select the "Read" button to begin.
Select the "Read" button to begin. Find general information about what you need in order to complete this tutorial
Task 1: Launch Time Series Experiment
Select the "Read" button to begin.
Select the "Read" button to begin. Take a quick look at the dataset, and do the initial set up of your experiment.
Task 2: Time Series Experiment Settings
Select the "Read" button to begin.
Select the "Read" button to begin. Explore and update the experiment settings and launch your experiment.
Task 3: Time Series Experiment Concepts
Select the "Read" button to begin.
Select the "Read" button to begin. Review general concepts about Time Series, and the concepts behind Time Series Forecasting with Driverless AI
Task 4: Experiment Results Summary
Select the "Read" button to begin.
Select the "Read" button to begin. Check the initial results from your experiment
Task 5: Model Interpretability
Select the "Read" button to begin.
Select the "Read" button to begin. Check out the model interpretability for Time Series experiments, and explore some of the results from your experiment
Task 6: Analysis
Select the "Read" button to begin.
Select the "Read" button to begin. Analyze the results from your experiment and compare them to the results obtained from a pre-ran experiment.
Next Steps
Select the "Read" button to begin.
Select the "Read" button to begin. Check out the next tutorial about Machine Learning Interpretability.
Appendix A: Add the Datasets
Select the "Read" button to begin.
Select the "Read" button to begin. Explore how to import datasets into Driverless AI
Time Series Tutorial Quiz
25 Questions  |  2 attempts  |  20/25 points to pass
25 Questions  |  2 attempts  |  20/25 points to pass
Time Series ML Intro Badge
  |  Badge available
  |  Badge available
You must be logged in to post to the discussion
  • ERICA ESCUDERO

    Hi, 

    I am experimenting with different time series in your tool and I have a question, how do you prepare the train and test files? must contain the same amount of time, for example 26 weeks each file or is it indistinct?
    Another query is how do I load both files in driverless AI to be able to perform the experiment?
    where you can get more detailed information about time series in driverless ai, thanks

    Reply
  • SR

    Hi, 

        Is there any threshold for Gap and Horizon for the test datasets or can we provide data with time periods (2009-2010 daily data) and ask prediction for (2020 daily data)?

    Reply
  • FV

    Hi Senthilkumar, you can predict on data for 2020, but we expect the pattern to match that of 2011. So if sales have increased from 2011 to 2020, then the model won’t know of this change. That’s where you would need to retrain or refit the model

    Reply