Quantitative Analysis of Gas Phase IR Spectra Based on Extreme Learning Machine Regression Model

Abstract

Advanced chemometric analysis is required for rapid and reliable determination of physical and/or chemical components in complex gas mixtures. Based on infrared (IR) spectroscopic/sensing techniques, we propose an advanced regression model based on the extreme learning machine (ELM) algorithm for quantitative chemometric analysis. The proposed model makes two contributions to the field of advanced chemometrics. First, an ELM-based autoencoder (AE) was developed for reducing the dimensionality of spectral signals and learning important features for regression. Second, the fast regression ability of ELM architecture was directly used for constructing the regression model. In this contribution, nitrogen oxide mixtures (i.e., N2O/NO2/NO) found in vehicle exhaust were selected as a relevant example of a real-world gas mixture. Both simulated data and experimental data acquired using Fourier transform infrared spectroscopy (FTIR) were analyzed by the proposed chemometrics model. By comparing the numerical results with those obtained using conventional principle components regression (PCR) and partial least square regression (PLSR) models, the proposed model was verified to offer superior robustness and performance in quantitative IR spectral analysis.

Towards the Direct Detection of Viral Materials at the Surface of Protective Face Masks via Infrared Spectroscopy

“…After spectral data acquisition, pre-processing of the spectra was performed for extracting the relevant information. 26,29 The IR spectra were analyzed using MATLAB (Release 2020a, MathWorks, Natick, USA) with the PLS toolbox 8.7 (Eigenvector Research Inc., Manson, USA). Typically, a variety of pre-processing steps are used to remove unwanted variance (i.e., variance not characteristic for the target analytes) from individual samples and numerically prepare data for modelling.…”

“…Principal components analysis (PCA) followed by principal components regression (PCR) is among the most commonly applied multivariate algorithms projecting variances contained within the data set onto a small number of Eigenvectors (i.e., principal components; PCs) with the aim of reducing dimensionality. 26 Conversely, partial least squares (PLS) regression is a method related to PCA, yet besides ful lling the criterion that a PC should describe the maximum residual variance simultaneously relates the latent variables to the dependent variables in an optimized way. PLS in combination with linear classi cation methods yields so-called PLS-based linear discriminant analysis routines (PLS-DA).…”

Rapid Quantitative Analysis of IR Absorption Spectra for Trace Gas Detection by Artificial Neural Networks Trained with Synthetic Data

“…With the use of machine learning-based methods, it is possible to build regression models, e.g., partial least-square regression (PLSR), for the prediction of gas concentrations [ 16 ]. In particular, models based on artificial neural networks (ANNs) have shown promising results in building regression models and processing acquired data in real time [ 17 , 18 ]. ANNs consist of nodes that are typically arranged in layers.…”

Application Of Mathematical Models And Digital Filters And Their Processors Of Spectral Analysis For Aromatic Compounds Gas In A Fluorescent Chemical

“…Aromatic compounds are broadly utilized in cutting edge industry. It is necessary to know their concentration in the air, particularly in industrial and populated ranges, due to their harmfulness indeed at parts per billion concentrations [1], these compounds such as benzene, toluene and xylene, cause extraordinary hurt to human health,(for example, anaemia, sterility and a few sorts of cancers), in addition to environmental dangers, despite the health issues caused by these aromatic compounds, we note that the production of these compounds increments universally [2].…”