eonu/sequentia

Improve tests for KNNClassifier

eonu opened this issue · 1 comments

eonu commented

A lot of the tests for KNNClassifier are useless tests for verboseness. These should be removed and replaced with more meaningful tests.

In general, tests for KNNClassifier and GMMHMM/HMMClassifier tests should eventually move towards using some real sample datasets rather than the random numbers we are using currently. (see here).

eonu commented

Implemented in #213 - dataset improvement will be done in #194.