Eigenvector University Online - LIVE classes start September 15. Get details here!

EVRI-thing You Need to Know About Cross-Validation

Aug 19, 2020

August 19, 2020


Our webinar series continues on Wednesday, August 19 with “EVRI-thing You Need to Know About Cross-Validation.” Join Eigenvector President Barry M. Wise for an in-depth discussion of this ubiquitous but often misused machine learning procedure.

Cross-validation, where calibration data sets are repeatedly split into training and test sets, is used extensively in the development of both predictive (e.g. Partial Least Squares regression, PLS) and descriptive models (e.g. Principal Components Analysis, PCA), for assessing performance and optimizing meta-parameters (e.g. number of components). But in order for cross-validation to give valid results it must be set up and performed correctly. In this webinar we’ll review the basics of cross-validation and show how to make it work properly (or not).

In this webinar we’ll discuss:

  • Fundamentals of cross-validation
  • Alternative ways of segmenting data sets
  • Cross-validation in the context of regression, classification and PCA
  • Customizing cross-validation sets
  • How to make cross-validation go bad

We will include some live demos with PLS_Toolbox/Solo to show you how to use the cross-validation settings for best results.

Register for “EVRI-thing You Need to Know About Cross-Validation” through your existing Eigenvector account, or create one. The webinar is free, but you must register to attend. Find the webinar near the bottom of the page under the “Purchase” tab.

A question and answer session will follow the webinar. Don’t miss your chance to quiz Barry and the Eigen-Guys about how to get the most out of cross-validation.

Reserve your seat today. The webinar will be live on Wednesday, August 19, at 7:00am PDT, (that’s 16:00 CEST). We will send you a WebEx invitation the day before the webinar. We hope you can attend live, but if you sign up and can’t make it, we’ll send you a link to view the recording the day after the webinar.