baseline for multi-label classification
Posted by:
Ayush Singhal
Date: August 14, 2013 08:05AM
I have run some multi-label classifiers on 2 datasets. Now I want to compare my results with some baseline to confirm that the feature set is useful in classification.
I wanted to know if a baseline such as:
randomly generated labels for each instance and then computing similar metrics on these random baseline output,
will do this job. Also given that there is also label imbalance, should I use this information also in randomly generated labels.