Data Cloud Video Hub

17. Building MultiClass Models

This episode of the Einstein Studio demo focuses on building multi-class machine learning models in Data 360, explaining key differences between binary and multi-class classification. Utilizing a channel optimization dataset, it demonstrates how to configure a model to predict from multiple outcomes, such as email, phone, social, and messaging channels. Viewers learn about the unique metrics reported for each class, the automatic selection of the best algorithm, and the significance of performance metrics like AUC for individual classes. The session concludes with insights on effectively training and assessing a multi-class model’s predictions.