知识图谱嵌入表示再学习的药物反应预测方法
Knowledge Graph Embedding Represents a Relearning Drug Response Prediction Method
谢新平 1汪凤婷 1姜晓东 2王红强3
作者信息
- 1. 安徽建筑大学数理学院,安徽合肥 230000
- 2. 中国科学技术大学第一附属医院肿瘤科,安徽合肥 230000
- 3. 中国科学院合肥物质科学研究院智能所,安徽合肥 230000
- 折叠
摘要
为解决基于网络进行药效预测的模型中加入新样本需要重新训练模型的问题,提出基于网络的药效预测方法,并进行了对比实验.该方法可以直接使用训练好的模型编码肿瘤细胞新样本,解决了加入新样本需要重新训练编码模型的问题,提高了药效预测精度,优于其他方法.
Abstract
To solve the problem of adding new samples to the network-based model for drug efficacy prediction and retraining the model,a network-based drug efficacy prediction method was proposed and comparative experiments were conducted.This method can directly use the trained model to encode new samples of tumor cells,solving the problem of needing to retrain the encoding model to add new samples,improving the accuracy of drug efficacy prediction,and outperforming other methods.
关键词
药效预测/知识图谱嵌入/深度学习Key words
pharmacodynamic prediction/knowledge graph embedding/deep learning引用本文复制引用
出版年
2024
NETL
NSTL
万方数据