Glossary · Phase 3 · Machine learning
Pipeline
A fixed series of steps that data goes through in order, such as fill gaps, rescale, then predict, so the steps can't get mixed up.
Try it now
The order matters: fitting any step on all the data leaks the test set into training.
Learn it properly
Preparing data · Clean, encode and scale data before a model sees it. About 30 minutes.
Other terms from the same lesson
- imputing: Filling in a missing value with a sensible guess, such as the column's median.
- median: The middle value once the numbers are sorted. Unlike the average, one huge value barely moves it.
- one-hot encoding: Turning a category into columns of 0s and 1s, with a 1 in the column of the matching category.
- outlier: A value far from the rest, such as one order of $5,000 among orders of about $50. It can be a real event or a mistake.
- standardization: Rescaling a column so its average is 0 and its standard deviation is 1, so columns measured in different units become comparable.
Never trained an AI?
Teach a small one to write, right in your browser. It takes about 20 seconds and needs no account.