Weighted classifier ensemble based on quadratic form

作者：

Highlights：

• A new weighted classifier ensemble method is proposed.

• An approximation form of the ensemble error is introduced.

• An optimal weight vector is sought based on minimizing the approximation form.

• It is converted into maximizing quadratic forms by invoking a known weight vector.

• The larger the value of the quadratic form is, the lower the ensemble error is.

摘要

•A new weighted classifier ensemble method is proposed.•An approximation form of the ensemble error is introduced.•An optimal weight vector is sought based on minimizing the approximation form.•It is converted into maximizing quadratic forms by invoking a known weight vector.•The larger the value of the quadratic form is, the lower the ensemble error is.

论文关键词：Ensemble learning,Weighted classifier ensemble,Quadratic form

论文评审过程：Received 23 July 2013, Revised 20 August 2014, Accepted 12 October 2014, Available online 20 November 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.10.017