What is it about?
We notice that not all labels are independent in multi-label classification problems. For instance, if one instance is labeled "kungfu", it may be also labeled with "Chinese". Ho to use the correlations to help the multi-classification problems? We use a maximum spanning tree algorithm to generate a tree to represent the correlations among the labels. The tree structure graph is naturally a tree-like CRF. We learn the tree structure CRF by fitting the model on training data.
Featured Image
Photo by Diane Picchiottino on Unsplash
Why is it important?
Use the correlations among labels to help the multi-label classification problems.
Read the Original
This page is a summary of: Multi-label Classification of Short Texts with Label Correlated Recurrent Neural Networks, July 2021, ACM (Association for Computing Machinery),
DOI: 10.1145/3471158.3472246.
You can read the full text:
Contributors
The following have contributed to this page
