Week 3: Feature engineering, modeling, and evaluation

DATA 441/641 · Week 3

Notes

• 3.1 Feature Engineering(slides)
• 3.2 Modeling(slides)
• 3.3 Evaluation(slides)
• 3.4 Deployment and Monitoring(slides)

Labs

• Lab 3: A Short Fake-News Pipeline(PDF)
• Lab 3 solutions (password)

Additional readings

• Practical NLP, Ch. 2 continued
• sklearn: Precision, recall, F-score

Discussion

• Give one intrinsic and one extrinsic way to score a tokenizer.
• What would you monitor after a spam filter goes live?