Research Article
Dictionary-Based Transfer Learning for One-Class Text Classification
@INPROCEEDINGS{10.4108/eai.26-5-2023.2334296, author={Haoxin Xie and Bo Liu and Yanshan Xiao}, title={Dictionary-Based Transfer Learning for One-Class Text Classification}, proceedings={Proceedings of the 2nd International Conference on Mathematical Statistics and Economic Analysis, MSEA 2023, May 26--28, 2023, Nanjing, China}, publisher={EAI}, proceedings_a={MSEA}, year={2023}, month={7}, keywords={one-class learning transfer learning dictionary learning}, doi={10.4108/eai.26-5-2023.2334296} }- Haoxin Xie
Bo Liu
Yanshan Xiao
Year: 2023
Dictionary-Based Transfer Learning for One-Class Text Classification
MSEA
EAI
DOI: 10.4108/eai.26-5-2023.2334296
Abstract
One-class classification requires only positive samples to build the model. In this paper, we propose a novel method for one-class text learning based on transfer learning and dictionary learning (DTOC). We integrate transfer learning and dictionary learning into one-class text model. In order to solve DTOC model, the interactive framework updates each variable by fixing other variables, and updates one-class classifier and dictionary alternatively to obtain the predictive classifier. Finally, extensive experiments show that our method has higher competitive classification accuracy and less sensitive to label noise than existing state-of-the-art methods.
Copyright © 2023–2024 EAI
