The big picture
Six ideas fit on one screen.
| Idea | Remember |
|---|---|
| Guess a number | Regression learns from past numbers and guesses a new number. |
| Group with no labels | Clustering finds groups in things with no labels. What it compares decides the groups. |
| Three kinds of guess | Ask what comes out. A named group is classification. A number to guess, like a price, is regression. Groups found with no labels are clustering. |
| What a dataset holds | A dataset is a set of records. Each record is described by attributes. One way to name a dataset is by what it holds. |
| Structured or unstructured data | Structured data fits rows and columns. Unstructured data does not. A dataset with both is hybrid. |
| Train, check, test | Train on one part. Improve with a second. Test on a third it never saw. |
Ask what comes out. Then ask what the model has never seen.