<aside> 📌 PATH: 360° Customer View → Predictive Models → Select already trained ML model → Deploy

</aside>

How to deploy ML Model:

Train and Evaluate ML model

Before deploying an ML model, create a new one, train it, and evaluate whether the testing results meet your business needs. For detailed steps on how to train and evaluate an ML model, please refer to our guides:

How to Add and Train ML Model

How to Evaluate ML Model (binary and multi classification)

How to Evaluate ML Model (regression)

Prepare dataset for prediction

Following the recommendations, prepare the dataset you will use to generate predictions for actual object IDs.

• For one-time predictions, you can upload a .csv file.
• For regular predictions (daily, weekly, monthly), use a table in Google BigQuery and set up a schedule. One-time predictions are also available for Google BigQuery datasets.

How to Prepare Dataset for Prediction

Press “Deploy”

If your ML model is already trained, tested, and meets your business requirements, you are ready to deploy it.

Simply press the “Deploy” button and follow the next steps.

Select source of prediction dataset

Choose the source of the input dataset for predictions:

<aside> 📌

• CSV: For one-time predictions.
• Google BigQuery Table: For one-time or scheduled predictions (e.g., daily, weekly). </aside>