Detection and quantification of groundnut oil adulteration with machine learning using a comparative approach with NIRS and UV-VIS

Abstract

Groundnut oil is known as a good source of essential fatty acids which are significant in the physiological development of the human body. It has a distinctive fragrant making it ideal for cooking which contribute to its demand on the market. However, some groundnut oil producers have been suspected to produce groundnut oil by blending it with cheaper oils especially palm olein at different concentrations or by adding groundnut flavor to palm olein. Over the years, there have been several methods to detect adulteration in oils which are time-consuming and expensive. Near infrared (NIR) and ultraviolet-visible (UV-Vis) spectroscopies are cheap and rapid methods for oil adulteration. This present study aimed to apply NIR and UV-Vis in combination with chemometrics to develop models for prediction and quantification of groundnut oil adulteration. Using principal component analysis (PCA) scores, pure and prepared adulterated samples showed overlapping showing similarities between them. Linear discriminant analysis (LDA) models developed from NIR and UV-Vis gave an average cross-validation accuracy of 92.61% and 62.14% respectively for pure groundnut oil and adulterated samples with palm olein at 0, 1, 3, 5, 10, 20, 30, 40 and 50% v/v. With partial least squares regression free fatty acid, color parameters, peroxide and iodine values could be predicted with R²CV's up to 0.8799 and RMSECV's lower than 3 ml/100 ml for NIR spectra and R²CV's up to 0.81 and RMSECV's lower than 4 ml/100 ml for UV-Vis spectra. NIR spectra produced better models as compared to UV-Vis spectra.

Keywords: Adulteration; Chemometrics; Free fatty acid; Iodine value; Near infrared; Peroxide value; Ultraviolet–visible.

Figure 1

Average values of physical and chemical properties of pure and laboratory adulterated samples. (A) Peroxide value. (B) Free fatty acid. (C) Iodine value. (D) Color coordinate L*. (E) Color coordinate b*. (F) Color coordinate a* (p < 0.05). Different alphabets represent samples with significant differences.

Figure 2

Raw spectra of laboratory adulteration and market samples (A) NIR raw spectra. (B) UV–VIS raw spectra.

Figure 3

Preprocessed spectra of laboratory adulteration and market samples (A) NIR preprocessed spectra. (B) UV–VIS preprocessed spectra.

Figure 4

PCA plot for laboratory adulterated laboratory and market samples (A–C) UV–Vis PCA plot for laboratory adulterated laboratory and market samples spectra (B). NIR PCA plot for laboratory adulterated laboratory and market samples.

Figure 5

LDA plot for laboratory adulterated and market samples (A–C) LDA and model parameters for UV–Vis (B–D) LDA and Model Parameters for NIRS.

Figure 6

PLSR plot for the prediction of palm olein concentration on groundnut oil using NIRS using Savitzky-Golay Pretreatment (filter 17).

Figure 7

PLSR plot for the prediction of palm olein concentration on groundnut oil using UV–Vis using Savitzky-Golay Pretreatment (filter 17).