Comparison between partial least square and support vector regression with a genetic algorithm wavelength selection method for the simultaneous determination of some oxygenate compounds in gasoline by FTIR spectroscopy

Catalytic Upgrading of Pyrolysis Wax Oil Obtained from Waxed Corrugated Cardboard Using Zeolite Y Catalyst

“…FTIR analysis was performed to obtain chemical group information about the thermal and catalytic pyrolysis products of solvent-extracted wax and wax oil samples (Figure ). Band assignments for all samples are based on the literature and are given in Table . − The authors had previously established that the wax oil obtained from pyrolysis of waxed cardboard was similar in structure and chemical composition to paraffin . FTIR spectra of thermal pyrolysis products (Figure f–j) remained similar to that of paraffin, and the second round of thermal pyrolysis did not change the spectra.…”

Evolutionary Neural Architecture Search for Type 2 Diabetes Mellitus Diagnosis from Salivary ATR-FTIR Spectroscopy

“…Among the usage of Genetic Algorithm (GA) with ATR-FTIR, the most frequent application is in the preprocessing step where it is used frequently to highlight the useful and most importante ATR-FTIR wavelength to facilitate the prediction or classification by a Machine Learning Algorithm. In [Asghari 2020] and in [Mohammadi 2021], this preprocess was used in a comparison between PLS and Support Vector Regression for determination of oxygenate in gasoline and for a quantitative determination of resins in oil samples, respectively. In [Zandbaaf 2022], GA was used again for wavelength selection, but with an artificial neural network to predict the breakdown voltage for transformer oils samples.…”

Rapid determination of rebaudioside A content in the macroporous resin elution process using a portable near-infrared spectrometer

“…Partial least squares regression (PLSR) is the most commonly used regression method for quantitative analysis, which aims to establish a linear connection between the spectral matrix (X) and the content (Y). 30,31 Principal component analysis (PCA) is used to extract a set of latent variables (LVs) from the spectral data, called the principal component (PC), which has a great influence on the results calculated by the model. 32 SVMR can tackle complicated nonlinear classification and regression issues.…”

“…As a learning algorithm, SVMR with generalization power and focusing on the similarity between samples can be applied to a limited number of samples and high dimensions issues. 31 SVMR can minimise prediction errors by finding an optimal parameter which defines the margin of a tube around the regression line. If it can benefit the prediction of the remaining samples, some samples may be allowed outside the tube.…”