Assessment of integrated freshness index of different varieties of eggs using the visible and near-infrared spectroscopy

Figures & data

Figure 1. Egg visible/near infrared spectrum acquisition device.

Table 1. Total variance explained.

Component	Initial Eigenvalue			Sum of squares of extracted loads
	Total	Percentage variance	Cumulative %	Total	Percentage variance	Cumulative %	Total
1	3.220	64.394	64.394	3.220	64.394	64.394	3.153
2	1.047	20.950	85.344	1.047	20.950	85.344	1.114
3	0.400	7.994	93.338
4	0.169	3.375	96.713
5	0.164	3.287	100.000

Table 2. Component matrix after rotation ^a.

Indicators	Component
	1	2
${\mathrm{W}}_{\mathrm{i}}$	0.925	0.100
$\mathrm{H}{\mathrm{A}}_{\mathrm{i}}$	0.902	−0.121
${\mathrm{D}}_{\mathrm{i}}$	0.871	0.064
$\mathrm{H}{\mathrm{Y}}_{\mathrm{i}}$	0.849	0.325
$\mathrm{S}-\mathrm{o}\mathrm{v}{\mathrm{a}}_{\mathrm{i}}$	0.061	0.990

Note: Extraction method: Principal component analysis method.

Rotation method: Kaiser normalized maximum variance method.

a. Rotation has converged after 3 iterations.

Table 3. Egg indicator weights based on principal component analysis.

indicator	${\mathrm{W}}_{\mathrm{i}}$	$\mathrm{H}{\mathrm{A}}_{\mathrm{i}}$	${\mathrm{D}}_{\mathrm{i}}$	$\mathrm{H}{\mathrm{Y}}_{\mathrm{i}}$	$\mathrm{S}-\mathrm{o}\mathrm{v}{\mathrm{a}}_{\mathrm{i}}$
weight	0.187	0.195	0.169	0.236	0.213

Figure 2. Changing trend of IFI with storage time.

Table 4. Descriptive statistical analysis of the IFI of the sample set.

Sample set	Max	Min	Ave	Std
Total (244)	11.329	5.542	8.103	1.341
Calibration set (163)	11.191	5.780	8.001	1.252
Validation set (81)	11.329	5.542	8.325	1.492

Figure 3. The raw spectral of brown-shell eggs and pink-shell eggs.

Table 5. Full-spectrum model prediction performance of IFI with different pre-processing methods.

Models		Rc^2	RMSEC	Rp^2	RMSEP	RPD
RAW	GS-SVR	0.856	0.006	0.717	0.019	1.722
	GA-SVR	0.843	0.007	0.758	0.018	1.907
	PSO-SVR	0.897	0.005	0.617	0.026	1.153
FD	GS-SVR	0.967	0.002	0.784	0.015	1.532
	GA-SVR	0.895	0.005	0.785	0.015	1.596
	PSO-SVR	0.999	0.000	0.445	0.049	0.296
SNV	GS-SVR	0.867	0.006	0.794	0.014	2.013
	GA-SVR	0.901	0.005	0.796	0.014	2.077
	PSO-SVR	0.962	0.002	0.661	0.023	1.273

Figure 4. Screening characteristic wavelengths by competitive adaptive reweighted sampling algorithm.

Figure 5. Screening characteristic wavelengths by shuffled frog leaping algorithm.

Figure 6. Correlation between measured and predicted values of IFI in calibration set and validation set.

Table 6. Evaluation results of the prediction model based on the characteristic wavelength.

Models		Rc^2	RMSEC	Rp^2	RMSEP	RPD
CARS (19)	GS-SVR	0.849	0.007	0.796	0.014	1.852
	GA-SVR	0.848	0.007	0.808	0.013	1.969
	PSO-SVR	0.860	0.007	0.793	0.014	1.846
SFLA (63)	GS-SVR	0.856	0.007	0.803	0.014	1.852
	GA-SVR	0.900	0.005	0.816	0.012	2.009
	PSO-SVR	0.876	0.006	0.786	0.014	1.793