Method for constructing liquorice near-infrared spectrum analysis model based on kernel function optimization and liquorice quality evaluation method

Abstract

The invention discloses a method for constructing a liquorice near-infrared spectrum analysis model based on kernel function optimization and a liquorice quality evaluation method. The method comprises the following steps of establishing a sample database according to collected liquorice spectrum data and a sample reference value; selecting a radial basis, a polynomial and a generalized T studentfunction as kernel functions of kernel principal component analysis to realize feature extraction of high-dimensional spectrum data; and based on this, selecting a radial basis function as a kernel function of a kernel partial least squares model. At the same time, a genetic algorithm is used to optimize relevant hyper-parameters of the kernel function, the liquorice near-infrared spectrum analysis model is established, a quality evaluation problem is transformed into a regression problem, and the quality of liquorice is analyzed. The method for constructing the liquorice near-infrared spectrum analysis model based on the kernel function optimization provided by the invention not only has the advantages of being simple in operation and lossless, but also ensures the safety and effectiveness of the quality of the liquorice.

Description

technical field [0001] The invention belongs to the technical field of near-infrared spectrum analysis and detection, and relates to a method for constructing a near-infrared spectrum detection and analysis model of licorice medicinal materials. Background technique [0002] Licorice, also known as sweet grass root and honey grass, is the root and rhizome of the leguminous plant Glycyrrhiza uralensis Fisch. At present, most of the licorice medicinal materials on the market are artificially planted. Due to the diversification of planting methods and the differences in planting environments, the quality of licorice is uneven, which affects the therapeutic effect. At present, the quality evaluation methods of licorice mainly include artificial experience method, ultramicromorphological method, DNA barcode identification method, high-performance liquid chromatography and other methods, which have disadvantages such as being greatly affected by subjective factors, complicated ope...

AI Technical Summary

Problems solved by technology

At present, most of the licorice medicinal materials on the market are artificially planted. Due to the diversification of planting methods and the differences in the planting environment, the quality of licorice is uneven, which affects the therapeutic effect.

At present, the quality evaluation methods of licorice mainly include artificial experience method, ultramicromorphological method, DNA barcode identification method, high-performance liquid chromatography and other methods, which have disadvantages such as being greatly affected by subjective factors, complicated operation, long analysis cycle, and the need to consume a large amount of reagents.

Images