Title:
HUMAN BODY FEATURE PARAMETER SELECTION METHOD BASED ON IMPROVED RRELIEFF IN COMBINATION WITH MRMR
Document Type and Number:
WIPO Patent Application WO/2020/258973
Kind Code:
A1
Abstract:
The present invention relates to the technical field of biological informatization. A human body feature parameter selection method based on improved RReliefF in combination with mRMR, comprising the following steps: using RReliefF algorithm improved by means of building a sample similarity distance model for calculating each feature weight so as to remove uncorrelated features, the RReliefF algorithm being; and using mRMR algorithm for calculating the relevancy between the maximum correlation and the minimum redundancy of a feature so as to remove redundant features. The present invention has the beneficial effects of improving the RReliefF algorithm by means of building a sample similarity distance model with simultaneous consideration of the Euclidean distance and the shape-based distance of a sample and improving the screenability of the feature selection algorithm, which facilitates the obtainment of human body physiological feature parameters having greater correlation and smaller redundancy and further facilitates the building of a human body composition predication model having simplified structure and higher precision.
Inventors:
GAO XIUE (CN)
CHEN BO (CN)
CHEN SHIFENG (CN)
SANG HAITAO (CN)
XIE WENXUE (CN)
ZHANG TIANSHU (CN)
CHEN BO (CN)
CHEN SHIFENG (CN)
SANG HAITAO (CN)
XIE WENXUE (CN)
ZHANG TIANSHU (CN)
Application Number:
PCT/CN2020/083187
Publication Date:
December 30, 2020
Filing Date:
April 03, 2020
Export Citation:
Assignee:
UNIV LINGNAN NORMAL (CN)
International Classes:
G06K9/62
Foreign References:
| CN110363229A | 2019-10-22 | |||
| CN108596104A | 2018-09-28 | |||
| CN104881707A | 2015-09-02 | |||
| CN107845407A | 2018-03-27 |
Other References:
AN, YUNJIE: "The Research on Similarity Search in Time Series Based on Morphological Characteristics", CHINA MASTER’S THESES FULL-TEXT DATABASE, no. 01, 10 January 2017 (2017-01-10), DOI: 20200628204601Y
Attorney, Agent or Firm:
BEIJING HAIHONG JIACHENG INTELLECTUAL PROPERTY & PARTNERS (CN)
Download PDF: