A deep fusion framework for unlabeled data-driven tumor recognition
作者：
Highlights：
• A fusion model is constructed by integrating deep representation learning and classification into one model, guiding and reinforcing each other.
• The model achieves good performance for binary classification even the simplest linear regression classifier is used.
• The model has good generalization ability and stability for small sample and classification imbalance.
• The system is open and can be improved as needed.
• The performance is verified on genetic-based tumor recognition.
摘要
•A fusion model is constructed by integrating deep representation learning and classification into one model, guiding and reinforcing each other.•The model achieves good performance for binary classification even the simplest linear regression classifier is used.•The model has good generalization ability and stability for small sample and classification imbalance.•The system is open and can be improved as needed.•The performance is verified on genetic-based tumor recognition.
论文关键词：Unlabeled data,Deep representation learning,Non-negative matrix factorization,Tumor recognition
论文评审过程：Received 5 August 2020, Revised 11 May 2021, Accepted 16 May 2021, Available online 4 June 2021, Version of Record 22 June 2021.
论文官网地址：https://doi.org/10.1016/j.patcog.2021.108066
