A hybrid approach to classifying Wikipedia article quality flaws with feature fusion framework

作者：

Highlights：

• Combine pretrained models with deep learning for quality flaw classification.

• Manually designed features and automatically extracted features are combined.

• Proposed method achieves notably improved precision, recall, and accuracy.

• Provide a best practice for feature selection of quality flaw classification.

摘要

•Combine pretrained models with deep learning for quality flaw classification.•Manually designed features and automatically extracted features are combined.•Proposed method achieves notably improved precision, recall, and accuracy.•Provide a best practice for feature selection of quality flaw classification.

论文关键词：Quality flaw,Deep learning,Fusion framework,Text classification

论文评审过程：Received 10 January 2021, Revised 16 April 2021, Accepted 17 April 2021, Available online 22 April 2021, Version of Record 19 May 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115089