Machine learning on high dimensional shape data from subcortical brain surfaces: A comparison of feature selection and classification methods

作者：

Highlights：

• Classification of brain disease with high-dimensional shape data is evaluated.

• Feature selection importance increased as sample size decreased.

• LASSO improved classification efficiency but was less robust and accurate.

• Regularized random forest was highly robust but computationally expensive.

摘要

•Classification of brain disease with high-dimensional shape data is evaluated.•Feature selection importance increased as sample size decreased.•LASSO improved classification efficiency but was less robust and accurate.•Regularized random forest was highly robust but computationally expensive.

论文关键词：Feature selection,Shape analysis,Biomarker,Brain,Subcortical

论文评审过程：Received 27 January 2016, Revised 24 August 2016, Accepted 21 September 2016, Available online 22 September 2016, Version of Record 27 November 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.09.034