DDI-prediction-drugbank-twoside-zhangDDI-DeepDDI

D3I-COMCF was comprehensively evaluated against state-of-the-art approaches using four standard performance metrics. Experiments were conducted on four publicly available datasets, namely DrugBank (including inductive and transductive settings) and TWOSIDES, both of which are sourced from \DSN-DDI: an accurate and generalized framework for drug\u2013drug interaction prediction by dual-view representation learning\ (dataset available at https:\/\/github.com\/microsoft\/Drug-InteractionResearch\/tree\/DSN-DDI-for-DDI-Prediction), as well as DeepDDI and ZhangDDI, which originate from \HTCL-DDI: a hierarchical triple-view contrastive learning framework for drug\u2013drug interaction prediction\ (dataset available at https:\/\/github.com\/ranzhran\/HTCL-DDI). I just collected these four data sets.  

D3I-COMCF 与当前前沿方法进行了全面对比评估，评估过程采用四项标准性能指标。实验依托四个公开数据集展开，分别为 DrugBank（涵盖归纳式（inductive）与直推式（transductive）两种设置）以及 TWOSIDES，二者的数据集均源自论文《DSN-DDI：基于双视图表征学习的药物相互作用（drug–drug interaction, DDI）预测精准通用框架》，数据集获取地址为：https://github.com/microsoft/Drug-InteractionResearch/tree/DSN-DDI-for-DDI-Prediction；另外两个数据集为 DeepDDI 与 ZhangDDI，其数据集源自论文《HTCL-DDI：面向药物相互作用（DDI）预测的分层三视图对比学习框架》，数据集获取地址为：https://github.com/ranzhran/HTCL-DDI。本次实验所用的这四个数据集均由笔者收集整理。