Missing values in test set

The front end.
Post Reply
liid
Posts: 32
Joined: Tue Jun 19, 2018 5:19 am

Missing values in test set

Post by liid »

Hi,

I know that the EM algorithm takes care of missing values in the training set. How do you deal with missing values in the test set? do you simply ignore that specific feature value for that specific sample? or do you do something more sophisticated (assuming I am not actively imputing them or something similar). Thanks!
shooltz[BayesFusion]
Site Admin
Posts: 1417
Joined: Mon Nov 26, 2007 5:51 pm

Re: Missing values in test set

Post by shooltz[BayesFusion] »

For each record in the test set we set evidence if the value is present in the record and the node is not a class node. The inference is then performed and most likely outcome is selected as prediction. Prediction is compared with actual value in the test set, and the confusion matrix is updated accordingly.
Post Reply