A boosting extreme learning machine for nearinfrared spectral quantitative analysis of diesel fuel and edible blend oil samples
writer：Xihui Bian*, Caixia Zhang, Xiaoyao Tan, Michal Dymek, Yugao Guo, Ligang Lin, Bowen Cheng, Xiaoyu
keywords：Extreme learning machine, Ensemble modeling, Boosting, Complex samples, Near-infrared spectroscopy
source：期刊
specific source：Analytical Methods, 2017, 9, 2983-2989
Issue time：2017年
Extreme learning machine (ELM) has drawn
increasing attention due to its characteristics of simple structure, high
learning speed and excellent performance. However, a single ELM tends to low
predictive accuracy and instability in dealing with quantitative analysis of
complex samples. To further improve the predictive accuracy and stability of
ELM, a new quantitative model, called boosting ELM is proposed. In the
approach, a large number of ELM sub-models are sequentially built by selecting
a certain number of samples from the original training set according to the
distribution of the sampling weights, and then their predictions aggregate by
weighted median. Activation function and the hidden nodes number of ELM sub-model
are determined simultaneously by the ratio of mean value and standard deviation
of correlation coefficients (MSR). The performance of the proposed method is
tested with diesel fuel and blended edible oil samples. Compared with partial
least squares (PLS) and ELM, the results demonstrate that boosting ELM is an
efficient ensemble model and has obvious superiorities in predictive accuracy
and stability. Therefore, the proposed method may be an alternative for
near-infrared (NIR) spectral quantitative analysis of complex samples