|
Morphometric variation explains misclassification patterns in weight-based egg grading of Nin-Kaset Black-Boned Chickens: a machine learning approach |
|---|---|
| รหัสดีโอไอ | |
| Creator | Thanathip Suwanasopee |
| Title | Morphometric variation explains misclassification patterns in weight-based egg grading of Nin-Kaset Black-Boned Chickens: a machine learning approach |
| Contributor | Putawan Munkong, Peerayut Nilchuen, Skorn Koonawootrittriron |
| Publisher | Faculty of Agriculture |
| Publication Year | 2569 |
| Journal Title | Khon Kaen Agriculture Journal |
| Journal Vol. | 54 |
| Journal No. | 5 |
| Page no. | 1139-1155 |
| Keyword | egg grade classification, morphological traits, shape index, machine learning, Nin-Kaset Black-Boned chicken eggs |
| URL Website | https://li01.tci-thaijo.org/index.php/agkasetkaj |
| Website title | Khon Kaen Agriculture Journal |
| ISSN | 3027-6497 (Online) |
| Abstract | Egg grading in poultry production systems is traditionally based on egg weight. However, in native chicken breeds such as Nin-Kaset Black-Boned chickens, which exhibit substantial variation in egg shape, weight-based grading alone may not fully reflect the external characteristics of eggs and may lead to grading inconsistencies. Although machine learning (ML) techniques have increasingly been applied in egg grading research, studies investigating the relationship between morphometric traits and grading misclassification in native chickens remain limited. Therefore, this study aimed to evaluate the potential of morphometric traits, including egg width, egg length, and shape index (SI), to explain misclassification patterns in conventional weight-based egg grading systems without using egg weight as an input variable. A total of 8,435 eggs from Nin-Kaset Black-Boned chickens were classified into three weight-based grades: S (<42 g), M (42–52 g), and L (>52 g). Morphometric data were analyzed using six ML algorithms, namely XGBoost, Random Forest, Support Vector Machine, Decision Tree, Logistic Regression, and k-Nearest Neighbors. Model performance was evaluated using accuracy, F1-score, and confusion matrix analysis. The results showed that XGBoost achieved the highest classification performance, with an overall accuracy of 83.74%. Misclassification analysis revealed that 41.20% of S-grade eggs with relatively large morphometric characteristics were classified as M-grade eggs, whereas 20.10% of L-grade eggs with relatively round or smaller shapes were classified as M-grade eggs. These findings indicate that variation in morphometric traits is associated with misclassification patterns in conventional weight-based egg grading and may provide a useful foundation for developing egg evaluation approaches that better accommodate the phenotypic diversity of native poultry production systems in the future. |
