Unlock instant, AI-driven research and patent intelligence for your innovation.

An Improved Team Progress Algorithm for Near Infrared Spectroscopy Wavelength Screening

A technology of near-infrared spectroscopy and screening methods, which is applied in the fields of genetic law, material analysis by optical means, instruments, etc., can solve the problem of insufficient prediction accuracy, and achieve the goal of ensuring prediction accuracy, improving detection accuracy, and reducing the number of wavelength variables. Effect

Active Publication Date: 2021-06-25
JIANGNAN UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The TPA algorithm is much simpler than the GA algorithm, and the number of wavelengths screened out is very small, but the prediction accuracy is slightly lower than that of the GA algorithm.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • An Improved Team Progress Algorithm for Near Infrared Spectroscopy Wavelength Screening
  • An Improved Team Progress Algorithm for Near Infrared Spectroscopy Wavelength Screening
  • An Improved Team Progress Algorithm for Near Infrared Spectroscopy Wavelength Screening

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0045] like figure 1 As shown, this embodiment proposes a near-infrared spectral wavelength screening method that improves the team's progressive algorithm, and applies it to a set of standard corn near-infrared spectral data sets. The spectral data set is referenced from the open source corn sample spectral data set on the eigenvector website, address https: / / eigenvector.com / resources / data-sets / . The data set includes 80 corn samples, which were measured by three spectroscopic instruments (m5, mp5, mp6). The wavelength range is 1100-2498nm in 2nm intervals (700 variables) and includes moisture, oil, protein and starch values ​​for each sample. These data were originally collected at Cargill. The experimental data used the sample data collected by the device mp5 in this data set and the corresponding protein content value.

[0046] The methods include:

[0047] Step 1: Outlier elimination and sample set division. Considering that the abnormal spectrum obtained due to the ...

Embodiment 2

[0074] In order to investigate the effect of the variable screening algorithm proposed in Example 1 on modeling prediction, the total number of members is set to 35, of which the number of elite groups is 10, the number of ordinary groups is 10, and the number of garbage collection groups is 15 indivual. Set the probability l of freshman members choosing learning as 0.35, that is, the probability of choosing exploration behavior as 0.65. The shrinkage index of the elite group is 20, and the shrinkage index of the ordinary group is generally half of that of the elite group, which is 10. Set the number of iterations to 1000. Figure 4 What is shown is the evaluation value of the optimal band of the elite group during the iterative process of the algorithm.

[0075] Using the classic near-infrared spectral wavelength screening algorithm Genetic Algorithm (GA), Principal Component Analysis (PCA), Team Progressive Algorithm (TPA) and the improved Team Progressive Algorithm (iTPA)...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a near-infrared spectrum wavelength screening method based on an improved team progress algorithm, belonging to the field of near-infrared spectrum detection. The method includes: according to the evaluation value from high to low, divide the near-infrared spectrum bands to be screened into elite group, common group and garbage collection group; The wavelength point of the band is selected from a random band in the selection group, and the wavelength point of the new wave band inherits the selected wavelength point; the inherited new wave band selects a learning behavior or exploration behavior by setting a probability to update its own Wavelength points, generate candidate bands, and select the band with the highest evaluation value in the elite group as the band to be screened. Under the condition that the prediction accuracy of the model is ensured, the invention greatly reduces the number of wavelength variables, reduces the complexity of the algorithm, and improves the precision of non-destructive detection in crops.

Description

technical field [0001] The invention relates to the field of near-infrared spectrum detection, in particular to a near-infrared spectrum wavelength screening method that improves the team progress algorithm. Background technique [0002] With the substantial improvement of my country's comprehensive strength, people have higher and higher requirements for product quality, including fruits, crops, food, etc. In the field of industrial process detection, it is necessary to detect the quality of components such as oils and chemicals, which not only requires fast and non-destructive detection technology, but also pursues detection accuracy. With the continuous improvement of urbanization, the contradiction between the demand for more accurate and convenient detection technology and the current relatively backward detection technology level is escalating. In the context of many applications, near-infrared spectroscopy has emerged. [0003] As an online analysis technology for t...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G01N21/359G01N21/3554G06N3/12
CPCG01N21/359G01N21/3554G06N3/126G01N2201/129G01N2021/8466G01N2201/12
Inventor 高美凤陶焕明于力革
Owner JIANGNAN UNIV