Title | ||
---|---|---|
Well-calibrated confidence measures for multi-label text classification with a large number of labels |
Abstract | ||
---|---|---|
•We propose a novel approach to address the computationally demanding nature of the Label Powerset (LP) Inductive Conformal Prediction (ICP) multi-label classification with a high number of labels. We mathematically establish the validity of the proposed approach and provide experimental results that highlight its computational efficiency.•We present prediction set results for data-sets in for multi-label text classification problems where it was previously computationally challenging and show that can be practically useful.•Results show that Bert classifier surpasses the non-contextualised based by a large margin. In addition, to the best of our knowledge, our bert implementation achieved state-of-the-art results in the data-sets used. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1016/j.patcog.2021.108271 | Pattern Recognition |
Keywords | DocType | Volume |
Text classification,Multi-label,Word2vec,Bert,Conformal prediction,Label powerset,Computational efficiency,Nonconformity measure,Confidence measure | Journal | 122 |
Issue | ISSN | Citations |
1 | 0031-3203 | 1 |
PageRank | References | Authors |
0.40 | 0 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Lysimachos Maltoudoglou | 1 | 1 | 0.40 |
Andreas Paisios | 2 | 1 | 0.40 |
Ladislav Lenc | 3 | 57 | 16.54 |
Jirí Martínek | 4 | 2 | 3.82 |
Pavel Král | 5 | 85 | 21.99 |
Harris Papadopoulos | 6 | 219 | 26.33 |