Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
SPECTRAL SIMILARITY-BASED CORRECTION SET AND VALIDATION SET SELECTING AND MODELING METHOD
Document Type and Number:
WIPO Patent Application WO/2021/073541
Kind Code:
A1
Abstract:
A spectral similarity-based correction set and validation set selecting and modeling method. The correction set and validation set selecting method comprises: performing near infrared spectroscopy determination on original samples to obtain an original sample spectral matrix; randomly selecting m samples as an independent test set; for each sample in the independent test set, respectively calculating the spectral similarity between the sample and each sample remaining in the original samples to obtain g samples with highest similarity, and writing the g samples into the validation set; for each sample in the validation set, respectively calculating the spectral similarity between the sample and each sample remaining in the original samples to obtain n samples with highest similarity, and writing the n samples into the correction set. The model obtained by the validation set and the correction set selected based on the selection method can predict an unknown model more accurately.

Inventors:
NIE LEI (CN)
SUN YUE (CN)
ZANG HENGCHANG (CN)
ZENG YINGZI (CN)
LIU XIAOYAN (CN)
SU MEI (CN)
YUAN MENG (CN)
WANG LINLIN (CN)
JIANG HONG (CN)
CHU GUANGYI (CN)
Application Number:
PCT/CN2020/120950
Publication Date:
April 22, 2021
Filing Date:
October 14, 2020
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
UNIV SHANDONG (CN)
International Classes:
G01N21/359; G06F17/18
Foreign References:
CN110687072A2020-01-14
CN106770005A2017-05-31
CN103411893A2013-11-27
CN109324014A2019-02-12
CN105486663A2016-04-13
US6529767B12003-03-04
Attorney, Agent or Firm:
JINAN SHENGDA INTELLECTUAL PROPERTY AGENCY CO., LTD. (CN)
Download PDF: