A Big Data-Based Analysis: Color as a Predictor of Silage Quality

Ömer Mermer; Ayşe Gül Filik; Gökhan Filik

Makale detayları

A Big Data-Based Analysis: Color as a Predictor of Silage Quality

Yazarlar:

Ömer Mermer | Ayşe Gül Filik | Gökhan Filik

Yayıncı:

Karyay Karadeniz Yayımcılık

Ana Başlıklar:

Hassas Tarım Teknolojileri

Dergi:

Black Sea Journal of Agriculture Vol.8 Issue 5, pp.725-735

Belge Tipi:

Makale

Yayın Tarihi:

9/2025

Belge Dili:

İngilizce

Document ID (Org No.):

2001358067

pISSN:

2618-6578

Anahtar kelimeler:

Özet:

 In this study, the relationship between color parameters and nutrient composition was examined using silage trial data conducted on 10 different plant species or mixtures by the Filik Research Lab. The dataset, comprising a total of 6,254 observations, was analyzed using big data analytics and statistical modeling techniques. Significant associations were identified between color parameters L, a, b, C, h, ΔE, saturation, and hue and the physical, chemical, and sensory characteristics of the silage crude protein, NDF, ADF, TDN, NFC. Initially, exploratory data analysis EDA and correlation matrices were employed to observe preliminary relationships between color and nutritional components. These were followed by classical and machine learning methods, such as multiple linear regression and Random Forest, to assess the strength and direction of these associations. The analysis revealed a negative correlation r  -0.45 between L lightness and crude protein CP, , while a strong positive relationship r  0.52 was found between b yellowness and NDF. Additionally, a meaningful correlation r  0.41 was observed between color saturation and TDN. Random Forest analysis confirmed these findings and highlighted L, b, and saturation as the most influential color parameters. These results suggest that color parameters are not only indicators of visual quality but also provide indirect insights into the chemical and nutritional content of silage. This study, conducted within the scope of big data analytics, demonstrates that color data can serve as a rapid predictive tool, either as an alternative or complement to traditional chemical analyses. This presents a significant contribution toward the development of digital quality control systems in agricultural production.

 In this study, the relationship between color parameters and nutrient composition was examined using silage trial data conducted on 10 different plant species or mixtures by the Filik Research Lab. The dataset, comprising a total of 6,254 observations, was analyzed using big data analytics and statistical modeling techniques. Significant associations were identified between color parameters L, a, b, C, h, ΔE, saturation, and hue and the physical, chemical, and sensory characteristics of the silage crude protein, NDF, ADF, TDN, NFC. Initially, exploratory data analysis EDA and correlation matrices were employed to observe preliminary relationships between color and nutritional components. These were followed by classical and machine learning methods, such as multiple linear regression and Random Forest, to assess the strength and direction of these associations. The analysis revealed a negative correlation r  -0.45 between L lightness and crude protein CP, , while a strong positive relationship r  0.52 was found between b yellowness and NDF. Additionally, a meaningful correlation r  0.41 was observed between color saturation and TDN. Random Forest analysis confirmed these findings and highlighted L, b, and saturation as the most influential color parameters. These results suggest that color parameters are not only indicators of visual quality but also provide indirect insights into the chemical and nutritional content of silage. This study, conducted within the scope of big data analytics, demonstrates that color data can serve as a rapid predictive tool, either as an alternative or complement to traditional chemical analyses. This presents a significant contribution toward the development of digital quality control systems in agricultural production.

 In this study, the relationship between color parameters and nutrient composition was examined using silage trial data conducted on 10 different plant species or mixtures by the Filik Research Lab. The dataset, comprising a total of 6,254 observations, was analyzed using big data analytics and statistical modeling techniques. Significant associations were identified between color parameters L, a, b, C, h, ΔE, saturation, and hue and the physical, chemical, and sensory characteristics of the silage crude protein, NDF, ADF, TDN, NFC. Initially, exploratory data analysis EDA and correlation matrices were employed to observe preliminary relationships between color and nutritional components. These were followed by classical and machine learning methods, such as multiple linear regression and Random Forest, to assess the strength and direction of these associations. The analysis revealed a negative correlation r  -0.45 between L lightness and crude protein CP, , while a strong positive relationship r  0.52 was found between b yellowness and NDF. Additionally, a meaningful correlation r  0.41 was observed between color saturation and TDN. Random Forest analysis confirmed these findings and highlighted L, b, and saturation as the most influential color parameters. These results suggest that color parameters are not only indicators of visual quality but also provide indirect insights into the chemical and nutritional content of silage. This study, conducted within the scope of big data analytics, demonstrates that color data can serve as a rapid predictive tool, either as an alternative or complement to traditional chemical analyses. This presents a significant contribution toward the development of digital quality control systems in agricultural production.

 In this study, the relationship between color parameters and nutrient composition was examined using silage trial data conducted on 10 different plant species or mixtures by the Filik Research Lab. The dataset, comprising a total of 6,254 observations, was analyzed using big data analytics and statistical modeling techniques. Significant associations were identified between color parameters L, a, b, C, h, ΔE, saturation, and hue and the physical, chemical, and sensory characteristics of the silage crude protein, NDF, ADF, TDN, NFC. Initially, exploratory data analysis EDA and correlation matrices were employed to observe preliminary relationships between color and nutritional components. These were followed by classical and machine learning methods, such as multiple linear regression and Random Forest, to assess the strength and direction of these associations. The analysis revealed a negative correlation r  -0.45 between L lightness and crude protein CP, , while a strong positive relationship r  0.52 was found between b yellowness and NDF. Additionally, a meaningful correlation r  0.41 was observed between color saturation and TDN. Random Forest analysis confirmed these findings and highlighted L, b, and saturation as the most influential color parameters. These results suggest that color parameters are not only indicators of visual quality but also provide indirect insights into the chemical and nutritional content of silage. This study, conducted within the scope of big data analytics, demonstrates that color data can serve as a rapid predictive tool, either as an alternative or complement to traditional chemical analyses. This presents a significant contribution toward the development of digital quality control systems in agricultural production.

Makale detayları

A Big Data-Based Analysis: Color as a Predictor of Silage Quality

Yazarlar:

Ömer Mermer | Ayşe Gül Filik | Gökhan Filik

Yayıncı:

Karyay Karadeniz Yayımcılık

Ana Başlıklar:

Hassas Tarım Teknolojileri

Dergi:

Black Sea Journal of Agriculture Vol.8 Issue 5, pp.725-735

Belge Tipi:

Makale

Yayın Tarihi:

9/2025

Belge Dili:

İngilizce

Document ID (Org No.):

2001358067

pISSN:

2618-6578

Silage quality

CIE Lab color space

Agricultural digitalization

Physical quality indicators

Machine learning

Agricultural decision support systemsSilage quality

CIE Lab color space

Agricultural digitalization

Physical quality indicators

Machine learning

Agricultural decision support systems

Sayı

Makale detayları

Yazarlar: Ömer Mermer | Ayşe Gül Filik | Gökhan Filik

Yayıncı: Karyay Karadeniz Yayımcılık

Ana Başlıklar: Hassas Tarım Teknolojileri

Dergi: Black Sea Journal of Agriculture Vol.8 Issue 5, pp.725-735

Belge Tipi: Makale

Yayın Tarihi: 9/2025

Belge Dili: İngilizce

Document ID (Org No.): 2001358067

pISSN: 2618-6578

Sayı

Yazarlar:

Ömer Mermer | Ayşe Gül Filik | Gökhan Filik

Yayıncı:

Karyay Karadeniz Yayımcılık

Ana Başlıklar:

Hassas Tarım Teknolojileri

Dergi:

Black Sea Journal of Agriculture Vol.8 Issue 5, pp.725-735

Belge Tipi:

Makale

Yayın Tarihi:

9/2025

Belge Dili:

İngilizce

Document ID (Org No.):

2001358067

pISSN:

2618-6578