IMPLEMENTING NEW PREDICTIVE FUNCTIONAL MODEL FOR MILK FAT VALUE IN MACEDONIAN WHITE BRINED CHEESE PRODUCTION

Makarijoski, Borche and Dimitrovska, Gordana and Joshevska, Elena (2025) IMPLEMENTING NEW PREDICTIVE FUNCTIONAL MODEL FOR MILK FAT VALUE IN MACEDONIAN WHITE BRINED CHEESE PRODUCTION. Agriculture & Forestry, 71 (1). pp. 34-37. ISSN 1800-9492

Text 20250331-03 Makarijoski et al..pdf - Published Version Download (354kB)

Abstract

In the production of Macedonian white-brined cheese, milk fat content is a
crucial determinant of the final product's quality, texture, and taste. Accurate
prediction and management of milk fat levels during cheese production are
essential for maintaining consistency, optimizing yield, and ensuring consumer
satisfaction. This study presents the implementation of a new predictive
functional model specifically designed to estimate milk fat value in the
production of Macedonian white-brined cheese. The model integrates various
factors, such as the initial composition of raw milk, processing conditions, and
key technological parameters that influence fat retention and distribution
throughout the cheese making process. By using a combination of statistical
analysis and machine learning techniques, the model enables a more precise and
real-time prediction of milk fat content, addressing challenges related to seasonal
variations in milk composition and other unpredictable factors in dairy
production. Data from local dairies were used to validate the model's
performance, and results demonstrate its accuracy in predicting milk fat values
with a high degree of reliability.
The study’s most significant findings demonstrate the variation and trends
in milk fat content across four cheese variants (A, B, C, and D) during the
ripening process. The results show that Variant D consistently maintained the
highest milk fat values throughout the ripening period, with significant
differences (p<0.05) compared to the other variants. By the 60th day of ripening,
Variant D had a milk fat content of 25.41±0.02%, which was 2.3% higher than
Variant C, the variant with the lowest fat content.
The predictive functional model achieved high R² values for all variants,
ranging from 0.9673 to 0.9997, indicating its robustness in estimating milk fat dynamics. Variant A achieved an R² of 0.9997, demonstrating near-perfect
alignment between the model’s predictions and experimental data.
The implementation of this model has the potential to streamline
production processes by reducing the need for frequent laboratory analyses and
allowing producers to make real-time adjustments during manufacturing.
Furthermore, it contributes to enhanced product standardization, better resource
management, and overall improvement in cheese quality. This predictive model
offers a novel tool for dairy producers in N. Macedonia and beyond, aiming to
improve the efficiency and quality control in white-brined cheese production.
Keywords: white brined cheese, milk fat, quality control, functional model

Item Type:	Article
Subjects:	Scientific Fields (Frascati) > Agricultural Sciences > Animal and diary science
Divisions:	Faculty of Biotechnical Sciences
Depositing User:	Prof. d-r Borche Makarijoski
Date Deposited:	06 Apr 2025 17:00
Last Modified:	06 Apr 2025 17:00
URI:	https://eprints.uklo.edu.mk/id/eprint/10858

Actions (login required)

View Item