Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Pine nut protein content prediction method based on ensemble learning calibration model

A technique for protein content, calibration models

Active Publication Date: 2020-11-27
NORTHEAST FORESTRY UNIVERSITY
View PDF1 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to solve the problem that the existing method does not make full use of the complex spatial characteristics in the near-infrared spectrum when establishing the near-infrared calibration model, resulting in the low accuracy rate of protein content prediction in pine nuts by using the existing near-infrared calibration model, A method for predicting the protein content of pine nuts based on an ensemble learning calibration model was proposed

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
  • Pine nut protein content prediction method based on ensemble learning calibration model
  • Pine nut protein content prediction method based on ensemble learning calibration model
  • Pine nut protein content prediction method based on ensemble learning calibration model

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0024] Specific implementation mode one: combine Figure 5 This embodiment will be described. A method for predicting protein content of pine nuts based on an integrated learning calibration model described in this embodiment, the method is specifically implemented through the following steps:

[0025] Step 1, collect the original spectral data samples of the pine nut sample, the number of samples is m, and then preprocess each original spectral data sample respectively, and obtain the preprocessed near-infrared spectral data;

[0026] Step 2. Use LTSA (Local Tangent Space Alignment), isomap (Isometric Feature Mapping), LLE (Local Linear Embedding) and PCA (Principal Component Analysis) to extract the features of the preprocessed near-infrared spectral data, and obtain four The four sets of feature vectors extracted by the method;

[0027] Step 3, select the boosting integrated learning algorithm to establish a calibration model for pine nut protein content prediction, and t...

specific Embodiment approach 2

[0031] Embodiment 2: This embodiment is a further detailed description of Embodiment 1. In the first step, each original spectral data sample is preprocessed respectively, and the method used in the preprocessing is standard normal variate , SNV) and SG (Savitzky-Golay) smoothing filtering.

[0032] In this embodiment, the near-infrared spectral data of the pine nut sample is preprocessed, the purpose of which is to eliminate the interference of sample surface scattering, baseline drift and noise on the spectral data, and enhance the data difference. Obtain near-infrared spectral data after removing interference.

specific Embodiment approach 3

[0033] Specific implementation mode three: this implementation mode is a further detailed description of specific implementation mode one, and the specific process of the step one is:

[0034] Step one, figure 2 Be the original spectrum of the pine nut sample collected by the spectrometer, for the i-th original spectrum data sample, set the optical program position number as j, j=1,2,...,l, l represents the total number of optical paths, Near-infrared spectral data obtained by fitting the original spectral data at the jth optical path position by using a p-order polynomial;

[0035]

[0036] Among them, a j′ is the weight coefficient, j'=0,1,...,p, the window width of p-order polynomial fitting is 2q+1, and λ is the absorbance in the wavelength range of the window width;

[0037] When the value of q is 4, the window width is 9, and when fitting the original spectral data at the 5th optical path position, then λ is the absorbance within the wavelength range from the 1st ...

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 pine nut protein content prediction method based on an ensemble learning calibration model, and belongs to the technical field of food component detection. According to the method, the problem of low accuracy of predicting the protein content in the pine nuts by utilizing an existing near-infrared calibration model is solved. The method comprises the following steps: preprocessing near infrared spectrum data of pine nuts, and selecting local tangent space alignment, equidistant feature mapping, local linear embedding and principal component analysis to perform featureextraction on the preprocessed spectrum data after preprocessing is finished; establishing a partial least square model of pine nut protein content and spectral data by using the extracted characteristic data set; and finally, outputting a final pine nut protein content result by taking a stacking method as an integration strategy and a BP neural network as a secondary learner. According to the method, the spectral data utilization degree is higher, complex spatial characteristics in the near-infrared spectrum are fully utilized, and the prediction accuracy of the calibration model is improved. The method can be applied to prediction of protein content in pine nuts.

Description

technical field [0001] The invention relates to the technical field of food component detection, in particular to a method for predicting protein content of pine nuts based on an integrated learning calibration model. Background technique [0002] The near-infrared spectrum modeling technology is to measure a series of reflectance, transmittance and physical and chemical properties of the experimental sample in the near-infrared spectral band of the experimental sample, and use statistical methods to optimize the spectral band and establish the physical and chemical properties of the experimental sample and the near-infrared spectrum. Calibration model, a technique for predicting the physicochemical properties of other samples using the established calibration model. Most of the calibration models use traditional statistical models such as PLS, PCR, and MCR. In recent years, with the development of machine learning and data mining, more and more machine learning methods have...

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
IPC IPC(8): G16B40/10G16B40/30G06N20/20
CPCG06N20/20G16B40/10G16B40/30
Inventor 张冬妍蒋大鹏李鸿博李丹丹曹军
Owner NORTHEAST FORESTRY UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products