一种快速高效的基因型与表型数据监督降维方法
By combining gradient boosting decision trees with related component regression, the dimensionality reduction process of genotype and phenotype data was optimized, solving the problems of low CCR computation efficiency and uncertain component number, and achieving fast and efficient dimensionality reduction and optimized prediction performance.
CN117577194BActive Publication Date: 2026-07-17HUAZHONG AGRI UNIV
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUAZHONG AGRI UNIV
- Filing Date
- 2023-10-18
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing correlation component regression methods are computationally inefficient during dimensionality reduction of genotype and phenotype data and lack the condition of determining the number of components, which affects prediction performance.
Method used
By combining gradient boosting decision tree strategy with related component regression method, the layer-by-layer accumulation of covariates is reduced by fitting the loss function, linear model is used for iteration, dimensionality reduction efficiency is optimized and compressed components are automatically determined.
Benefits of technology
It achieves a fast and efficient dimensionality reduction process, maintains the scale invariance advantage of CCR, and improves prediction performance.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure BDA0004506664970000021 
Figure BDA0004506664970000027 
Figure BDA0004506664970000031
Abstract
本发明公开了一种快速高效的基因型与表型的监督降维方法，该方法融合梯度提升决策树（Gradient Boosted Decision Tree）策略与相关分量回归（Correlated Component Regression,CCR）方法，通过对损失函数的拟合减少协变量的逐层累加，在不损失更多信息、保留CCR尺度不变性优势的同时以线性模式提升降维效率，并优化了确定压缩分量的标准，实现了最终压缩分量的可自动确定。本发明所述的方法采用线性算法，计算复杂度低，计算速度在现有基因组降维技术中最快，通过结合梯度提升策略以及采用相关组分的收敛方法，使降维后的数据具备原始数据近似或更佳的预测性能。
Need to check novelty before this filing date? Find Prior Art