Building global models for fat and total protein content in raw milk based on historical spectroscopic data in the visible and short-wave near infrared range

Standard

Building global models for fat and total protein content in raw milk based on historical spectroscopic data in the visible and short-wave near infrared range. / Melenteva, Anastasiia; Galyanin, Vladislav; Savenkova, Elena; Bogomolov, Andrey.

In: Food Chemistry, Vol. 203, 15.07.2016, p. 190-198.

Research output: Contribution to journal › Article › peer-review

BibTeX

@article{0f6e940a006c45ec96ed783a2c5f77d0,

title = "Building global models for fat and total protein content in raw milk based on historical spectroscopic data in the visible and short-wave near infrared range",

abstract = "A large set of fresh cow milk samples collected from many suppliers over a large geographical area in Russia during a year has been analyzed by optical spectroscopy in the range 400-1100 nm in accordance with previously developed scatter-based technique. The global (i.e. resistant to seasonal, genetic, regional and other variations of the milk composition) models for fat and total protein content, which were built using partial least-squares (PLS) regression, exhibit satisfactory prediction performances enabling their practical application in the dairy. The root mean-square errors of prediction (RMSEP) were 0.09 and 0.10 for fat and total protein content, respectively. The issues of raw milk analysis and multivariate modelling based on the historical spectroscopic data have been considered and approaches to the creation of global models and their transfer between the instruments have been proposed. Availability of global models should significantly facilitate the dissemination of optical spectroscopic methods for the laboratory and in-line quantitative milk analysis.",

keywords = "Global modelling, Light scatter, Milk analysis, Model transfer, Short-wave near infrared spectroscopy, Variable selection, Visible spectroscopy",

author = "Anastasiia Melenteva and Vladislav Galyanin and Elena Savenkova and Andrey Bogomolov",

year = "2016",

month = jul,

day = "15",

doi = "10.1016/j.foodchem.2016.01.127",

language = "English",

volume = "203",

pages = "190--198",

journal = "Food Chemistry",

issn = "0308-8146",

publisher = "Elsevier",

}

RIS

TY - JOUR

T1 - Building global models for fat and total protein content in raw milk based on historical spectroscopic data in the visible and short-wave near infrared range

AU - Melenteva, Anastasiia

AU - Galyanin, Vladislav

AU - Savenkova, Elena

AU - Bogomolov, Andrey

PY - 2016/7/15

Y1 - 2016/7/15

N2 - A large set of fresh cow milk samples collected from many suppliers over a large geographical area in Russia during a year has been analyzed by optical spectroscopy in the range 400-1100 nm in accordance with previously developed scatter-based technique. The global (i.e. resistant to seasonal, genetic, regional and other variations of the milk composition) models for fat and total protein content, which were built using partial least-squares (PLS) regression, exhibit satisfactory prediction performances enabling their practical application in the dairy. The root mean-square errors of prediction (RMSEP) were 0.09 and 0.10 for fat and total protein content, respectively. The issues of raw milk analysis and multivariate modelling based on the historical spectroscopic data have been considered and approaches to the creation of global models and their transfer between the instruments have been proposed. Availability of global models should significantly facilitate the dissemination of optical spectroscopic methods for the laboratory and in-line quantitative milk analysis.

AB - A large set of fresh cow milk samples collected from many suppliers over a large geographical area in Russia during a year has been analyzed by optical spectroscopy in the range 400-1100 nm in accordance with previously developed scatter-based technique. The global (i.e. resistant to seasonal, genetic, regional and other variations of the milk composition) models for fat and total protein content, which were built using partial least-squares (PLS) regression, exhibit satisfactory prediction performances enabling their practical application in the dairy. The root mean-square errors of prediction (RMSEP) were 0.09 and 0.10 for fat and total protein content, respectively. The issues of raw milk analysis and multivariate modelling based on the historical spectroscopic data have been considered and approaches to the creation of global models and their transfer between the instruments have been proposed. Availability of global models should significantly facilitate the dissemination of optical spectroscopic methods for the laboratory and in-line quantitative milk analysis.

KW - Global modelling

KW - Light scatter

KW - Milk analysis

KW - Model transfer

KW - Short-wave near infrared spectroscopy

KW - Variable selection

KW - Visible spectroscopy

UR - http://www.scopus.com/inward/record.url?scp=84958166919&partnerID=8YFLogxK

U2 - 10.1016/j.foodchem.2016.01.127

DO - 10.1016/j.foodchem.2016.01.127

M3 - Article

C2 - 26948605

AN - SCOPUS:84958166919

VL - 203

SP - 190

EP - 198

JO - Food Chemistry

JF - Food Chemistry

SN - 0308-8146

ER -

ID: 41677915