Definition: a homonym is a word with the same pronounciation as another but with a different meaning, origin, and, usually, spelling.

Purpose: to illustrate how C4.5 and C4.5rules can be used to automate a more difficult task.

Problem: using three collected sets of instances containing homonyms, learn rules to determine, given a window of the five previous words (classified according to their respective Part Of Speech), which spelling of each homonym is preferred.

Downloadable Files

Homonym Pair filestem.names filestem.data Decision Tree Decision Rules
"Bare" versus "Bear" bare.names bare.data text or gif text or gif
"Baron" versus "Barren" baron.names baron.data text or gif text or gif
"In" versus "Inn" in.names in.data text or gif text or gif
"Wait" versus "Weight" wait.names wait.data text or gif text or gif

Summary of Results

Comment: only the immediately preceeding word/part of speech contributes to these rules.

"Bare" versus "Bear"

"Baron" versus "Barren"

"In" versus "Inn"

"Wait" versus "Weight"