研究目的
Investigating the performance of the closed-loop moving window-partial least squares (closed-loop MW-PLS) method in wavelength selection for near-infrared (NIR) spectroscopic analysis of soil organic matter.
研究成果
The closed-loop MW-PLS method was found to be strictly superior to the original MW-PLS method in NIR analysis of soil organic matter, offering a meaningful extension without increasing operational complexity. The method's ability to form different closed loops for different search ranges allows for more waveband combinations, potentially leading to better results.
研究不足
The study is limited to the application of the closed-loop MW-PLS method in NIR spectroscopic analysis of soil organic matter. The method's applicability to other types of samples or spectroscopic analyses was not explored.
1：Experimental Design and Method Selection:
The closed-loop MW-PLS method was proposed based on the reverse-deletion waveband idea, enabling flexible multi-band wavelength selection.
2：Sample Selection and Data Sources:
A total of 148 soil samples were collected, ground, and sifted. The organic matter content was measured using the standard potassium dichromate oxidation method.
3：List of Experimental Equipment and Materials:
An XDS Rapid Content? grating spectrometer equipped with a diffuse reflection accessory and a round sample cell was used for NIR spectroscopy.
4：Experimental Procedures and Operational Workflow:
The closed-loop MW-PLS method was applied to the NIR spectra of soil samples, and its performance was compared with the MW-PLS method.
5：Data Analysis Methods:
The root-mean-square error (SEP) and the correlation coefficients (RP) of prediction were used as evaluation indicators.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
