Testing and out of sample data

Learning Outcomes

• compare approaches to using data science to improve software tests
• awareness of domain-specific nature of SE machine learning needs
• risks of overfitting

Before Class

Lectures

• in class discussion, no notes

Readings

• Bhatta et al. Understanding Test Deletion in Java Applications
• Novielli, Sentiment trained on nonSE data
• Zimmermann, Cross-project Defect Prediction: A Large Scale Experiment on Data vs. Domain vs. Process

In Class

Slides

• In-class notes

• exercise

Optional Readings and Activities

• Zhang et al., exLong: Generating Exceptional Behavior Tests with Large Language Models