Fig. 5
From: AutoDTI++: deep unsupervised learning for DTI prediction by autoencoders
Comparison results of AutoDTI++ method with the six states-of-the-art methods (DDR, DNILMF, NRLMF, KRONRLS-MKL, BLM-NII, and COSINE) in terms of AUPR scores, using 5-repeats of tenfold CV. Results are obtained under \({S}_{p}\), \({S}_{d}\), and \({S}_{t}\) settings on NR, GPCR, IC, and E datasets. The results for DDR, DNILMF, NRLMF, KRONRLS-MKL, BLM-NII, and COSINE are captured using the best parameters published