Access this on-demand session to understand how data analysts evaluate predictive models and decide how best to deploy them in real-world settings. Experienced practitioners know that no single measure can define model accuracy, so choosing the right evaluation criteria is essential for meaningful and reliable results. This session provides a clear, practical overview of the key performance measures and deployment considerations.

You will learn how to compare different indicators of predictive accuracy, such as overall accuracy, AUC and lift values, and how to use charts like gains curves and ROC plots to assess model behaviour. The session emphasises the importance of defining practical success criteria early in a project and explores how profitability and operational impact can guide model selection. It also covers ways to test model performance and explains the different options for deploying predictive outputs across an organisation.

Designed for anyone who builds predictive models or needs to understand how they should be evaluated and implemented, this on-demand session offers a grounded, accessible guide to effective model assessment and deployment.