Compositional model based on factorial evolution for realizing multi-task learning in bacterial virulent protein prediction

作者：

Highlights：

• Multitask learning framework for virulent protein prediction to handle data scarcity.

• Multifactorial evolutionary based kernel-feature pair selection.

• Least square method is used to cope with marginal distribution differences.

• MTL framework achieves the highest AUC (0.98) compared to the state-of-the-art methods.

• Proposed MTL method outperform both single task learning and multi task learning methods.

摘要

•Multitask learning framework for virulent protein prediction to handle data scarcity.•Multifactorial evolutionary based kernel-feature pair selection.•Least square method is used to cope with marginal distribution differences.•MTL framework achieves the highest AUC (0.98) compared to the state-of-the-art methods.•Proposed MTL method outperform both single task learning and multi task learning methods.

论文关键词：Multifactorial evolutionary algorithm,Multitask learning,Multi-kernel learning,Virulent protein

论文评审过程：Received 3 November 2018, Revised 1 October 2019, Accepted 5 November 2019, Available online 7 November 2019, Version of Record 18 November 2019.

论文官网地址：https://doi.org/10.1016/j.artmed.2019.101757