作者: Lei Zhang , Fengchun Tian , Xiongwei Peng , Lijun Dang , Guorui Li
DOI: 10.1016/J.SNB.2012.11.113
关键词: Principal component analysis 、 Electronic nose 、 Projection (set theory) 、 Calibration 、 Artificial neural network 、 Sensor array 、 Pattern recognition 、 Affine transformation 、 Computer science 、 Signal 、 Artificial intelligence
摘要: Abstract The shift in sensor signal measured by identical gas array system (commonly called an electronic nose) makes the analysis of merged measurement data difficult. This would grossly affect quantification accuracy such nose (E-nose) instruments. Thus, a real-time calibration transfer based on reference alcohol projection model (RAPT) was designed this paper which aims to project onto hazardous and set up “bridge” from instrument through three artificial neural networks (ANN), attempt solve problem between E-nose instruments array. Besides, principal component (PCA) is also used for validation different models space. For comparison, previous four including univariate direct standardization (UDS), partial least square (PLS), neural, global affine transformation robust weighted (GAT-RWLS) are presented. Qualitative quantitative results demonstrate that proposed RAPT competitive standardization.