What do we need to write in the report of homework 1?
Do we simply list the different k's we tried and the exact match ratio for
each k in the cross validation process? Then we use the best k to predict
the test data? Maybe a graph of k vs. exact match ratio for CV and real test?
Or we need to write something about our design decisions, for example: number
of folds in CV, how to choose the predicted class if there are several
multi-label appear the same number of times (random pick?)?
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 140.112.233.45