Application of LS-SVM to non-linear phenomena in NIR spectroscopy: development of a robust and portable sensor for acidity prediction in grapes

Abstract

Nowadays, near infrared (NIR)technology is being transferred from the laboratory to the industrial world for on-line and portable applications. As a result, new issues are arising, such as the need for increased robustness, or the ability to compensate for non-linearities in the calibration or instrument. Semi-parametric modeling has been suggested as a means for adapting to these complications. In this article, Least-Squared Support Vector Machine (LS-SVM) regression, a semi-parametric modeling technique, is used to predict the acidity of three different grape varieties using NIR spectra. The performance and robustness of LS-SVM regression are compared to Partial Least Square Regression (PLSR) and Multivariate Linear Regression (MLR). LS-SVM regression produces more accurate prediction. However SNV pretreatment is required to improve the model robustness.NIR Spectroscopy Robust calibration LS-SVM PLSR MLR Grapes tartaric and malic acidity.

Determination Of Gas Mixture Components Using Fluctuation Enhanced Sensing And The LS-SVM Regression Algorithm

“…A disadvantage of both methods is that the training time increases with the square of the number of training samples (N) and linearly with the number of variables (dimension of the investigated spectra), which is opposite to the case of classic least-squares methods using principal component analysis algorithm [16].…”

“…The fundamentals of this method, and various applications for predicting chemical compound concentration, have been presented elsewhere [6,8,[11][12][13][14][15]. The algorithm attracted huge interest among scientists because it is based on a very simple idea and leads to high performance in numerous practical applications [14,[16][17][18]. The algorithm was developed originally for pattern recognition by learning from exemplary data belonging to two opposite sets.…”

“…Here we used the latter because the main error of the recorded data is the random error of the estimated power spectral densities due to limited averaging time. The RBF kernel function should work well with such a source of data inaccuracy, as considered also in other regression problems [8,16].…”

Nondestructive Determination of Fruit Based on Sensory and Near-Infrared Technology

“…The performance and robustness of LS-SVM regression are compared to Partial Least Square Regression (PLSR) and Multivariate Linear Regression (MLR). LS-SVM regression produces more accurate prediction [7].…”

Applications of integrated sensing and processing in spectroscopic imaging and sensing

“…Higher order algorithms such as support vector machines (SVM) and kernel PCA (KPCA) can be employed in such situations, such as multidimensional image and multivariate data analysis. SVMs have been used for analysis of mid-infrared spectra [10], nonlinear parametric models called multilayer perceptrons [11], NIR spectra affected by temperature-induced spectral variation [12], NIR spectra for acidity prediction in grapes [13], and neural networks [14]. KPCA has been used for optical character recognition and analysis of DNA [15], images [16], and NIR spectra [17].…”