Outlier-robust extreme learning machine for regression problems

An extension of extreme learning machine (ELM).

Handling data containing outliers for regression and binary classification by extreme learning machine.
The objective is to minimize a combination of L1-norm loss function and L2-norm output weights.
For more details, please refer to the paper "Zhang K, Luo M. Outlier-robust extreme learning machine for regression problems[J]. Neurocomputing, 2015, 151: 1519-1527.".