Multicollinearity Problem in Prediction of Egg Albumen Index and Some Regression Methods

Research on Japanese quail (Coturnix coturnix japonica) eggs examined factors affecting albumen index prediction and evaluated the multicollinearity problem among egg quality traits. The study included egg weight, egg width, egg length, shape index, and Haugh unit as explanatory variables for predicting albumen index. Correlation analysis results showed a high positive relationship (r=0.85) between egg weight and egg width and a high negative relationship (r=−0.91) between egg length and shape index. Variance inflation factor (VIF) values ranging from 1.41 to 30.30 indicated the presence of strong multicollinearity within the dataset. To overcome the effects of multicollinearity, principal component regression (PCR) and penalty-based regression methods including Ridge, LASSO, Liu, and Elastic Net were applied. PCR analysis showed that the first two principal components explained 75% of the total variance, demonstrating the effectiveness of dimension reduction. Ridge regression stabilized coefficient estimates by retaining all variables, whereas LASSO controlled model complexity through coefficient shrinkage. Liu regression provided stable coefficient estimates under correlated variable structures, while Elastic Net combined L1 and L2 penalties to regulate coefficient selection. The comparison of model performance criteria demonstrated that PCR, Ridge, and LASSO regression models provided the highest adjusted R² values (0.889). In addition, LASSO regression showed the lowest AIC, cAIC, and BIC values, indicating the best balance between model fit and complexity among the evaluated approaches. The findings demonstrated that classical regression approaches may be insufficient when multicollinearity exists among biological traits. Therefore, PCR and regularized regression methods provide reliable alternatives for predicting albumen index and improving model stability in datasets containing highly correlated variables.

Authors

Institutions

Publication Details

Journal
Black Sea Journal of Agriculture
Published
2026-11-15
DOI
https://doi.org/10.47115/bsagriculture.1964459
Primary Topic
Advanced Statistical Methods and Models
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Multicollinearity Problem in Prediction of Egg Albumen Index and Some Regression Methods

Ömer Faruk Karaokur, Mustafa Şahin
Black Sea Journal of Agriculture
Advanced Statistical Methods and Models
article

Multicollinearity Problem in Prediction of Egg Albumen Index and Some Regression Methods

Ömer Faruk Karaokur, Mustafa Şahin
article en

Abstract

Research on Japanese quail (Coturnix coturnix japonica) eggs examined factors affecting albumen index prediction and evaluated the multicollinearity problem among egg quality traits. The study included egg weight, egg width, egg length, shape index, and Haugh unit as explanatory variables for predicting albumen index. Correlation analysis results showed a high positive relationship (r=0.85) between egg weight and egg width and a high negative relationship (r=−0.91) between egg length and shape index. Variance inflation factor (VIF) values ranging from 1.41 to 30.30 indicated the presence of strong multicollinearity within the dataset. To overcome the effects of multicollinearity, principal component regression (PCR) and penalty-based regression methods including Ridge, LASSO, Liu, and Elastic Net were applied. PCR analysis showed that the first two principal components explained 75% of the total variance, demonstrating the effectiveness of dimension reduction. Ridge regression stabilized coefficient estimates by retaining all variables, whereas LASSO controlled model complexity through coefficient shrinkage. Liu regression provided stable coefficient estimates under correlated variable structures, while Elastic Net combined L1 and L2 penalties to regulate coefficient selection. The comparison of model performance criteria demonstrated that PCR, Ridge, and LASSO regression models provided the highest adjusted R² values (0.889). In addition, LASSO regression showed the lowest AIC, cAIC, and BIC values, indicating the best balance between model fit and complexity among the evaluated approaches. The findings demonstrated that classical regression approaches may be insufficient when multicollinearity exists among biological traits. Therefore, PCR and regularized regression methods provide reliable alternatives for predicting albumen index and improving model stability in datasets containing highly correlated variables.

Black Sea Journal of AgricultureVol. 9(6)
Kahramanmaraş Sütçü İmam University (TR)
Advanced Statistical Methods and Models
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon