A two-parameter egg-shape equation using only maximum length and width was tested on 120 ISA Brown eggs against fluid-displacement volume. Initial MAPE was 6.02%, reduced to 3.09% with a closed-form correction. Two-measurement grading rules reached 92.0-93.3% cross-validated accuracy, but external and cross-breed validation is absent.
Key findings
- Predicted and measured volume correlated at r=0.894 with 6.02% MAPE; correction reduced MAPE to 3.09%, below the compared empirical formula. A linear discriminant achieved 92.0% cross-validated grade accuracy, while H×W² reached 93.3%. Shape parameter and whole-egg density were nearly invariant across grades in this sample.
Why this matters globally
Industrial graders are expensive. A model based on calipers or simple imaging could support small enterprises and low-cost apps or conveyors, provided it meets weight-grade standards and operational throughput.
Thai researcher contribution
Mahasarakham University researchers and demonstration-school students collaborated with Khon Kaen University, connecting geometry to a low-cost agricultural problem across school and university levels.
Limitations to consider
Only 120 ISA Brown eggs were studied, with flock and hen-age diversity unclear. Cross-validation shared the model-development dataset and may be optimistic. Commercial grades are weight-based; volume is related but not identical. Abnormal shapes, damage and operator measurement error were not tested.
Verify the original sources
MathematicsRead the original article↗DOI: 10.3390/math14142473