New! Smarter Prediction & Model Evaluation in ArcGIS Pro 3.6

Ready to level up your spatial machine learning workflow? ArcGIS Pro 3.6 introduces powerful new tools and enhancements to help you build, test, and trust your predictive models like never before.

1. New: Prepare Data for Prediction
   This tool helps to effortlessly split and balance your data for model training and validation. It avoids bias and data leakage with smart data balancing and spatial splitting — so your models perform reliably on real-world scenarios.
2. New: Evaluate Predictions with Cross-validation
   Go beyond the basic validation step! This tool brings spatial cross-validation to the table, letting you test your model’s accuracy across different geographic locations and with rare events. It helps you understand how your model performs in new regions and edge cases.

3. Enhanced: Predict Using Spatial Statistics Model File
   Now, with a spatial statistics model file (.ssm file), not only can you predict new locations, but you can also validate predictions against known values. Get instant feedback with metrics like accuracy, sensitivity, F1-score, and R-squared, and you’ll know exactly how your model is performing.

Why does this matter?
These amazing new features and updates make it easier to engineer training data, tune models, and evaluate results—all in one streamlined workflow. Whether you’re tackling complex machine learning problems or just want more confidence in your prediction results, the spatial statistics prediction and model evaluation tools will have you covered.

Interested in learning more? Check out the full details:

• Prepare Data for Prediction
• Evaluate Predictions with Cross-validation
• Predict Using Spatial Statistics Model File