Near-infrared spectroscopy characteristic wavelength selection method for least squares support vector machine model

Abstract

The invention discloses a near-infrared spectroscopy characteristic wavelength selection method for a least squares support vector machine model. The wavelength selection process is guided by the distribution case of a spectral information original sample reflected to a kernel function space, and the key is wavelength selection for reflecting the original sample to the kernel function space so that a large support vector boundary, a small distribution radius R and a large sample space distance are obtained and thus after the original sample is reflected by the kernel function space, a high degree of recognition is obtained. The method has a simple principle, can be realized easily and can easily acquire a model having a high distinction degree. The method carries out wavelength selection based on a structural risk minimization principle, is only based on a training sample set and guarantees model generalization ability after wavelength selection.

Description

technical field [0001] The invention relates to a method for selecting near-infrared spectrum characteristic wavelengths for a least squares support vector machine model, and belongs to the technical field of spectrum analysis. Background technique [0002] Near-infrared spectroscopy has the characteristics of fast, low cost, and non-destructive, and has become an important tool in the field of physical property analysis. Near-infrared spectrum samples often contain hundreds or even thousands of wavelength points, which contain a lot of redundant or noise information. If full-spectrum modeling is used, the complexity of modeling will be increased, and the accuracy of the model will be easily reduced due to the introduction of redundant or noise information. Therefore, optimizing the modeling wavelength points of near-infrared spectroscopy and selecting representative characteristic wavelength points is an important means to improve the robustness and accuracy of spectral mo...

AI Technical Summary

Problems solved by technology

At present, the spectral wavelength selection methods for LSSVM can be mainly divided into three categories: one is to select the combination of wavelength points with the lowest correlation according to the correlation between wavelengths for modeling, and the core is to eliminate the redundancy of wavelength points (such as continuous projection mapping), this method only considers the characteristic independence between wavelengths, and it is easy to introduce noise wavelengths; the second type is to consider the correlation between spectral wavelength and measured information, and select wavelengths with greater correlation for modeling (such as correlation coefficient method ), this method only considers the correlation between the characteristic wavelength and the measured information, and it is easy to introduce more redundant wavelengths; another type selects the wavelength through an intelligent search algorithm (such as genetic algorithm, particle swarm algorithm, etc.), and the search process of these methods It is easy to be affected by the initial parameters, and the search process is easy to fall into the local extremum, and it is difficult to obtain the globally optimal wavelength combination

Images