Machine learning
Teach a program to find patterns, measure whether it really learned, and see why markets are a hard case.
- What is machine learning?Programs that learn rules from examples.
- Train and testThe one rule of machine learning you must never break.
- Fitting a lineYour first machine learning model, from scratch.
- Measuring errorA single number for how wrong a model is.
- Memorizing is not learningBuild a model that is perfect in training and useless after.
- ClassificationPredicting a category instead of a number.
- Nearest neighborsPredict by finding similar examples.
- Why markets are hard for machinesThe same methods that master images struggle with prices.
- Track quizFive questions. Get four right to finish the track.