we did mention this in the class..
If k = 1, the prediction will be the set of the closest point
※ 引述《hyperstring (try and win)》之銘言：
: What is the rule to decide in multi-class K-NN problem?
: How to decide the set of classes a test instance belong to after the distances
: are computed?
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 140.112.30.33