Model construction method, substance identification method and device, storage medium and program product

The model is constructed through machine learning methods, and the mass spectrometry data of gas samples is used to identify elements and functional groups, which solves the accuracy of I/SVOC component identification and classification in the prior art, and realizes efficient analysis of component functional groups.

CN119719763APending Publication Date: 2025-03-28TSINGHUA UNIVERSITY
6 Cites 0 Cited by

Patent Information

Application Number
CN202411651561.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-18
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

It is difficult for the prior art to accurately identify and classify medium volatile and semi-volatile organic components (I/SVOCs) in gas samples. Due to the limitation of separation effect, the phenomenon of components co-flow is serious. Traditional methods only achieve a more accurate classification of about 50% of the components.

Method used

The model is constructed using machine learning methods, and feature vectors and labels are constructed by obtaining the mass spectral data of standard substances, and machine learning models are trained and tested to identify the elements and functional groups of the substance to be tested and determine their substance categories.

Benefits of technology

The proportion of qualitative classification components has been significantly improved, the resolution of the functional group composition of the full spectrum components has been improved, and the rapid and accurate identification of large samples composed of all outflow components is achieved by machine learning.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

A model construction method, a substance identification method and device, a storage medium and a program product, the model construction method comprising: obtaining a training sample, the training sample comprising a plurality of standard substance mass spectra; constructing a feature vector of the training sample, wherein the feature vector of the training sample comprises relative peak intensities of a plurality of mass-to-charge ratio positions of the standard substance mass spectrum; a training sample label is constructed, the training sample label comprises an element label and a structure label, the element label comprises a plurality of element categories, and the structure label comprises a plurality of functional group structure categories; and constructing a machine learning model, and training and checking the machine learning model based on the feature vector of the training sample and the training sample label.
Need to check novelty before this filing date? Find Prior Art