DP-3007: Train and Deploy a Machine Learning Model with Azure Machine Learning — Lesson 2: Model Training, Evaluation, and Experiment Tracking

Learning Objectives

• Train a machine learning model with appropriate features, algorithm, and parameters
• Track runs, metrics, artifacts, logs, and comparison evidence
• Evaluate model performance, errors, bias, overfitting, and business acceptance criteria

Detailed Topics

• Train a machine learning model with appropriate features: Train a machine learning model with appropriate features, algorithm, and parameters
• Track runs: Track runs, metrics, artifacts, logs, and comparison evidence
• Evaluate model performance: Evaluate model performance, errors, bias, overfitting, and business acceptance criteria

Guided Demonstration

1. Review the requirement, environment, dependencies, roles, risks, and expected outcome.
2. Demonstrate the relevant configuration, analysis, code, or administration workflow step by step.
3. Capture settings, outputs, logs, screenshots, calculations, or test evidence.
4. Compare the observed result with the stated acceptance criteria.

Hands-On Activity

Train and compare model runs, then select a candidate using defensible metrics.

• Document assumptions and the planned sequence before making changes.
• Complete the activity and preserve evidence of normal and exception behavior.
• Record issues, corrective actions, and reusable operational or development notes.

Troubleshooting Focus

• Separate symptoms from causes and collect evidence before applying a fix.
• Check configuration, permissions, data, dependencies, capacity, connectivity, and recent changes.
• Validate rollback, recovery, support ownership, and user or business impact.

Validation Checklist

• The requirement and expected outcome are clear.
• The selected configuration, design, query, code, or process is documented.
• Normal, boundary, failure, security, and operational conditions are reviewed.
• The result is supported by evidence and the next improvement is prioritized.

Lesson Outcome

Learners can explain and apply Model Training, Evaluation, and Experiment Tracking in a realistic scenario and validate the result with documented evidence.