We’re excited to announce the latest release to Tealium Predict ML, with significant improvements in the workflows to create, train, and deploy models.
These new enhancements help you train and deploy models faster and easier with clearer steps and enhanced visualizations. The new interface provides more guided experiences to help you better interpret models and act on the predictions generated by the models.
How it Works
Tealium Predict has built-in capabilities that apply machine learning insights to your data without the need for data science or machine learning expertise. Tealium Predict works with Tealium AudienceStream CDP to build and create more intelligent audiences and activate trustworthy machine learning insights.
Using Tealium Predict, you can create, optimize, and deploy data models that make meaningful predictions about visitor behavior allowing you to determine:
The driving factors that influence your key performance indicators (KPIs). For example, which types of customers are likely to make online purchases.
Which marketing channels within your website drive purchases.
Categories of online visitor segmentation to apply successful customer personalization.
New Features & Improvements
Streamlined model overviews with guided actions The new Predict includes a brand new Model Overview experience that helps you take action throughout the various phases of the model lifecycle: creation, training, deployment, and retraining. Grouping models by stage can help you focus on the next action to improve and/or take action on the model results. The new model dashboard views provide a recommended set of actions to take depending on the state of your model.
New! Live Performance view allows you to monitor the real-time performance of a model A new tabbed view separates access to insights and training metrics when viewing the live performance data of a deployed model. This Live Performance view provides the ability to see how the model is applied to real visitors in live sessions. As a result, you can see how well the live model is performing, which gives you the tools you need to make better decisions about how to use Predict models to target specific audiences.
Improved Training Details experience allows you to focus on training model metrics that will drive the most insight and value for your business Predict now provides our more advanced users with better training details on their mode, so they can more easily assess training performance and take action. For example, being able to zoom into the probability distribution graph in more detail can help drive a better understanding and validation of model quality.
Better insights and guardrails to create better models, faster We’ve improved the quality of the Predict attribute readiness ratings so that you can quickly see which customer behaviors are ready for modeling. You can also easily choose which attributes to include or exclude from your model making it easier to define the data you need to improve the accuracy of your model.