AURO select one instance and its 2 labels to query which one is more relevant.
The query type of this method is different with the normal active learning algorithms that always query labels. To this end, the label matrix you provided can have the following additional information:
1. -1 means irrelevant. 2. A positive value means relevant, the larger, the more relevant. (However, do not use 2 which is defined as the dummy label)
References
----------
[1] Huang S J , Chen S , Zhou Z H . Multi-label active learning: query type matters[C]// Proceedings of the 24th International Joint Conference on Artificial Intelligence, pages 946-952, Buenos Aires, Argentina, July 25-31, 2015
__init__(self, X, y)
select(self, label_index, unlabel_index, **kwargs)
Select a subset from the unlabeled set, return the selected instance and label.
Parameters: |
|
---|---|
Returns: |
|
Copyright © 2018, alipy developers (BSD 3 License).