Robust feature selection to predict tumor treatment outcome

作者：

Highlights：

• A novel wrapper method searches forward in a hierarchical feature subset space.

• Prior domain knowledge is incorporated into the selection procedure.

• Promising results are obtained on two cancer-patient datasets.

摘要

Highlights•A novel wrapper method searches forward in a hierarchical feature subset space.•Prior domain knowledge is incorporated into the selection procedure.•Promising results are obtained on two cancer-patient datasets.

论文关键词：Hierarchical forward feature selection,Prior knowledge,Prediction,Small sample,Positron emission tomography,Support vector machine

论文评审过程：Received 6 January 2015, Revised 1 July 2015, Accepted 1 July 2015, Available online 14 August 2015, Version of Record 2 September 2015.

论文官网地址：https://doi.org/10.1016/j.artmed.2015.07.002