Machine Learning with Swift

Stemming

Stemming is the process of reducing words to their stems. The idea here is that related words can usually be reduced to a common stem.

For example: (white, whitening, whitish, whiter) → whit.

This can be used, for instance, to expand user queries. But some cases can be tricky, consider the English man and men, the Irish bhean = woman and mnà = women, or the even more extreme English am, is, are, was, were, and been. There are several popular stemmers for English.