Abstract
Total polar materials (TPM) content is considered as the best indicator and the most common parameter to check the quality of deep-frying oils. The development of simpler and quicker analytical techniques than the available methods to monitor oil quality in restaurants and fried food outlets is an important topic related to the human health. This paper reports a comparison of the variable selection of near infrared (NIR) spectra by multiple linear regression (MLR-NIR) with partial least squares (PLS-NIR) models for the quantification of TPM in fried vegetable oils. The use of PLS-NIR offers an alternative in laboratory bench equipment for the determination of TPM in oils employed for frying different kinds of foods with relative prediction errors of 6.5%, a coefficient of determination for prediction of 0.99 and a residual predictive deviation (RPD) of 9.2 when selected wavenumber intervals were employed. MLR-NIR allows the selection of a reduced number of wavenumber in order to develop low cost instruments to evaluate the frying oil quality. Based on the NIR signals at four wavenumbers, the relative prediction error was 12.1%, the coefficient of determination for prediction was 0.96 and the RPD was 5.0.
© 2018 The Author(s)
PDF Article
More Like This
Cited By
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Contact your librarian or system administrator
or
Login to access Optica Member Subscription