4.2 How Machine Learning Works
Collect data, train a model, test it, deploy it, and keep improving it. It's a loop, not a one-time thing.
Chapter 4: Introduction to Machine Learning (ML)
Every ML project follows roughly the same life cycle. First you collect data and clean it up. Then you train a model - it makes guesses, checks how wrong it was, and tweaks itself to be less wrong. Repeat thousands of times.
Next you evaluate it on data it has never seen, to check it actually learned and didn't just memorise. If it passes, you deploy it so real people can use it.
Then you monitor it, because the world changes. A model trained on 2019 shopping habits would be pretty confused by 2020. New data goes back into the loop.
Key takeaways
Collect → clean → train → evaluate → deploy → monitor
Training = guess, measure error, adjust, repeat
Models need updating as the world changes