public
/
wwu-577
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
wwu-577/docs/hw2.md

558 B
Raw Blame History

Computer Science 477/577 - Homework 2

Matt Jensen

Very Late

Suppose that you are trying to discretize a continuous attribute by using the ChiMerge Algorithm. When you merge two rows (because the \chi^{2} statistic of the upper row falls below a threshold) you must the recompute at least one \chi^{2} value. Is the new value necessarily greater than the old one?

A formal proof is not necessary. Just an intelligent discussion. You will not lose points by giving the wrong answer, only by not being intelligent in your discussion.