Skip to content

Latest commit

 

History

History
11 lines (7 loc) · 440 Bytes

File metadata and controls

11 lines (7 loc) · 440 Bytes

This exercise concerns the expressiveness of decision lists (Section learning-theory-section).

  1. Show that decision lists can represent any Boolean function, if the size of the tests is not limited.

  2. Show that if the tests can contain at most $k$ literals each, then decision lists can represent any function that can be represented by a decision tree of depth $k$.