A hybrid approach combining extreme learning machine and sparse representation

A simple but very effective combination of ELM and SRC.

By incorporating extreme learning machine (ELM) and sparse representation (SRC) into a unified framework, the proposed hybrid classifier not only has advantage of fast testing (the merit of ELM) but also shows notable classification accuracy (the merit of SRC).
We test it for AR face recognition, and it achieves a high accuracy of 95%, better than ELM (91%) and SRC (93.5%).
The bridge between ELM and SRC is the ELM misclassification metric and adaptive protein classes selection.
For more details, please refer to the paper "Luo M, Zhang K. A hybrid approach combining extreme learning machine and sparse representation for image classification[J]. Engineering Applications of Artificial Intelligence, 2014, 27: 228-235.".
Note that it is an improved version of the above paper.