Research on Detection of Soybean Meal Quality by NIR Based on PLS-GRNN

Li-qi WANG; Jing YAO; Rui-ying WANG; Ying-shu CHEN; Shu-nian LUO; Wei-ning WANG; Yan-rong ZHANG

doi:10.3964/j.issn.1000-0593(2022)05-1433-06

Spectroscopy and Spectral Analysis, Volume. 42, Issue 5, 1433(2022)

Research on Detection of Soybean Meal Quality by NIR Based on PLS-GRNN

Li-qi WANG^1、*, Jing YAO^1、1;, Rui-ying WANG^1、1;, Ying-shu CHEN^1、1;, Shu-nian LUO^2、2;, Wei-ning WANG^2、2;, and Yan-rong ZHANG^{1、1; *;}

Author Affiliations

¹1. School of Computer and Information Engineering, Harbin University of Commerce, Heilongjiang Provincial Key Laboratory of E-commerce and Information Processing, Harbin 150028, China

²2. School of Food Engineering, Harbin University of Commerce, Harbin 150028, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(15)

Fig. 1. Near infrared spectra of soybean meal samples

Download full size

View in Article

Fig. 2. iPLS modeling results of 20 subintervals for moisture

Download full size

View in Article

Fig. 3. iPLS modeling results of 20 subintervals for protein

Download full size

View in Article

Fig. 4. iPLS modeling results of 20 subintervals for fat

Download full size

View in Article

Fig. 5. Structure of GRNN

Download full size

View in Article

Fig. 6. The changing trends of PRESS with the major factors

Download full size

View in Article

Fig. 7. Optimizing curves of spread

Download full size

View in Article

Fig. 8. Predictive effects of PLS-GRNN models for moisture, protein and fat

Download full size

View in Article

Table 1. Comparison of statistical results of various noise reduction methods
View table
View in Article
Table 1. Comparison of statistical results of various noise reduction methods
去噪方法 perc err
小波去噪(db6、 2层分解、 penalty阈值) 0.999 9 0.000 5
移动平均法 0.999 8 0.001 6
多元散射校正 0.999 7 0.090 4
标准正态变量变换 0.998 5 0.111 0

Table 2. Modeling results of two sample set partitioning methods

View table

View in Article

Table 2. Modeling results of two sample set partitioning methods

分集方法	指标	校正集		预测集
分集方法	指标	R²	RMSEC	R²	RMSEP	RSD/%
	水分	0.922 4	0.239 1	0.917 3	0.242 9	2.11
KS	蛋白质	0.914 7	0.571 7	0.910 6	0.511 2	1.10
	脂肪	0.875 8	0.154 0	0.839 9	0.178 8	16.83
	水分	0.906 4	0.243 2	0.896 4	0.251 9	2.20
SPXY	蛋白质	0.933 1	0.498 3	0.908 2	0.522 3	1.13
	脂肪	0.852 4	0.161 0	0.891 9	0.145 0	13.10

Table 3. Sample partition results

View table

View in Article

Table 3. Sample partition results

指标	样本集	数目 /个	最小值 /%	最大值 /%	均值 /%	标准差 /%
水分	校正集	278	9.79	12.82	11.437	0.646 7
水分	预测集	80	9.97	12.76	11.510	0.603 5
蛋白质	校正集	278	41.2	50.9	46.374	1.494 7
蛋白质	预测集	80	41.4	49.2	46.436	1.577 7
脂肪	校正集	278	0.44	2.49	1.044	0.352 1
脂肪	预测集	80	0.45	2.41	1.107	0.351 8

Table 4. iPLS waveband selection results of moisture for different subinterval number
View table
View in Article
Table 4. iPLS waveband selection results of moisture for different subinterval number
参数指标子区间数最佳建模波段/cm^-1 R² RMSECV
水分 20 4 904~5 200 0.926 9 0.242 4
30 5 204~5 400 0.924 1 0.246 8
40 5 216~5 364 0.913 6 0.262 6
50 4 844~4 960 0.921 5 0.250 9

Table 5. iPLS waveband selection results of protein for different subinterval number
View table
View in Article
Table 5. iPLS waveband selection results of protein for different subinterval number
参数指标子区间数最佳建模波段/cm^-1 R² RMSECV
蛋白质 20 4 304~4 600 0.939 6 0.510 8
30 4 404~4 600 0.929 9 0.549 7
40 4 456~4 604 0.928 7 0.554 0
50 4 484~4 600 0.930 6 0.546 7

Table 6. iPLS waveband selection results of fat for different subinterval number
View table
View in Article
Table 6. iPLS waveband selection results of fat for different subinterval number
参数指标子区间数最佳建模波段/cm^-1 R² RMSECV
脂肪 20 4 304~4 600 0.875 9 0.170 2
30 5 604~5 800 0.869 9 0.174 0
40 4 304~4 452 0.872 3 0.172 7
50 4 204~4 360 0.872 7 0.172 6

Table 7. Comparison of modeling effects between PLS-GRNN and PLS, BP

View table

View in Article

Table 7. Comparison of modeling effects between PLS-GRNN and PLS, BP

参数指标	建模方法	R²	RMSEP	RSD/%
	PLS	0.917 3	0.242 9	2.11
水分	BP	0.957 2	0.124 8	1.08
	PLS-GRNN	0.976 9	0.091 2	0.79
	PLS	0.910 6	0.511 2	1.10
蛋白质	BP	0.921 9	0.440 9	0.95
	PLS-GRNN	0.940 2	0.383 4	0.83
	PLS	0.891 9	0.145 0	13.10
脂肪	BP	0.909 3	0.115 3	10.90
	PLS-GRNN	0.911 1	0.113 4	8.53

Tools

Get Citation

Copy Citation Text

Li-qi WANG, Jing YAO, Rui-ying WANG, Ying-shu CHEN, Shu-nian LUO, Wei-ning WANG, Yan-rong ZHANG. Research on Detection of Soybean Meal Quality by NIR Based on PLS-GRNN[J]. Spectroscopy and Spectral Analysis, 2022, 42(5): 1433

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Research Articles

Received: Apr. 14, 2021

Accepted: --

Published Online: Nov. 10, 2022

The Author Email: Li-qi WANG (hsdwlq@163.com)

DOI:10.3964/j.issn.1000-0593(2022)05-1433-06

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Comparison of statistical results of various noise reduction methods

Table 1. Comparison of statistical results of various noise reduction methods

Table 2. Modeling results of two sample set partitioning methods

Table 2. Modeling results of two sample set partitioning methods

Table 3. Sample partition results

Table 3. Sample partition results

Table 4. iPLS waveband selection results of moisture for different subinterval number

Table 4. iPLS waveband selection results of moisture for different subinterval number

Table 5. iPLS waveband selection results of protein for different subinterval number

Table 5. iPLS waveband selection results of protein for different subinterval number

Table 6. iPLS waveband selection results of fat for different subinterval number

Table 6. iPLS waveband selection results of fat for different subinterval number

Table 7. Comparison of modeling effects between PLS-GRNN and PLS, BP

Table 7. Comparison of modeling effects between PLS-GRNN and PLS, BP