CTF-DDI: Constrained tensor factorization for drug-drug interactions prediction

Guosheng Han, Lingzhi Peng, Aocheng Ding, Yan Zhang, Xuan Lin

Abstract: Highlights•A novel tensor factorization method with effective constraints is proposed.•CTF-DDI leverages Hessian and L2,1 regularization as constraints.•CTF-DDI combines constraints-based tensor factorization and deep neural networks.•CTF-DDI can reduce data sparsity and provide more effective representation.•Experiments show CTF-DDI outperforms tensor factorization-based and classic methods.