Table 1. Open in new tab The performance...

Models	1:1	1:5	1:10
RF (Pedregosa et al. 2011)	0.8409	0.9016	0.9129	0.9103	0.9093	0.7836	0.9438	0.9176	0.7156
SVM (Chang and Lin 2011)	0.7993	0.8586	0.8111	0.9074	0.8917	0.6962	0.9380	0.8871	0.6078
XGBoost (Chen and Guestrin 2016)	0.8573	0.9238	0.9323	0.7982	0.8586	0.8111	0.9550	0.9311	0.7864
GCN (Kipf and Welling 2016)	0.8393	0.8938	0.8758	0.9068	0.8895	0.7100	0.9299	0.8617	0.5817
GAT (Veličković et al. 2017)	0.8219	0.8759	0.8668	0.8710	0.8558	0.6339	0.9268	0.8525	0.5340
DTI-CNN (Peng et al. 2020)	0.8523	0.9262	0.9340	0.9269	0.9281	0.8286	0.9558	0.9319	0.7957
GCNMDA (Long et al. 2020)	0.8850	0.9424	0.9347	0.9044	0.9354	0.7520	0.9302	0.9423	0.6573
MVGCN (Fu et al. 2022)	0.8489	0.9042	0.9017	0.9132	0.9209	0.7777	0.9445	0.9163	0.6959
MMGCN (Tang et al. 2021)	0.9085	0.9556	0.9122	0.9403	0.9671	0.8038	0.9582	0.9715	0.7684
GraphCDA (Dai et al. 2022)	0.8796	0.9459	0.9471	0.9221	0.9484	0.8353	0.9377	0.9133	0.6435
DTINet (Luo et al. 2017)	0.8672	0.9390	0.9432	0.8983	0.9017	0.8511	0.9029	0.9003	0.7883
MIDTI (ours)	0.9340	0.9787	0.9701	0.9413	0.9813	0.9075	0.9539	0.9794	0.8431

Models

1:1

1:5

1:10

ACC

AUC

AUPR

ACC

AUC

AUPR

ACC

AUC

AUPR

RF (Pedregosa et al. 2011)

0.8409

0.9016

0.9129

0.9103

0.9093

0.7836

0.9438

0.9176

0.7156

SVM (Chang and Lin 2011)

0.7993

0.8586

0.8111

0.9074

0.8917

0.6962

0.9380

0.8871

0.6078

XGBoost (Chen and Guestrin 2016)

0.8573

0.9238

0.9323

0.7982

0.8586

0.8111

0.9550

0.9311

0.7864

GCN (Kipf and Welling 2016)

0.8393

0.8938

0.8758

0.9068

0.8895

0.7100

0.9299

0.8617

0.5817

GAT (Veličković et al. 2017)

0.8219

0.8759

0.8668

0.8710

0.8558

0.6339

0.9268

0.8525

0.5340

DTI-CNN (Peng et al. 2020)

0.8523

0.9262

0.9340

0.9269

0.9281

0.8286

0.9558

0.9319

0.7957

GCNMDA (Long et al. 2020)

0.8850

0.9424

0.9347

0.9044

0.9354

0.7520

0.9302

0.9423

0.6573

MVGCN (Fu et al. 2022)

0.8489

0.9042

0.9017

0.9132

0.9209

0.7777

0.9445

0.9163

0.6959

MMGCN (Tang et al. 2021)

0.9085

0.9556

0.9122

0.9403

0.9671

0.8038

0.9582

0.9715

0.7684

GraphCDA (Dai et al. 2022)

0.8796

0.9459

0.9471

0.9221

0.9484

0.8353

0.9377

0.9133

0.6435

DTINet (Luo et al. 2017)

0.8672

0.9390

0.9432

0.8983

0.9017

0.8511

0.9029

0.9003

0.7883

MIDTI (ours)

0.9340

0.9787

0.9701

0.9413

0.9813

0.9075

0.9539

0.9794

0.8431

Models	1:1	1:5	1:10
RF (Pedregosa et al. 2011)	0.8409	0.9016	0.9129	0.9103	0.9093	0.7836	0.9438	0.9176	0.7156
SVM (Chang and Lin 2011)	0.7993	0.8586	0.8111	0.9074	0.8917	0.6962	0.9380	0.8871	0.6078
XGBoost (Chen and Guestrin 2016)	0.8573	0.9238	0.9323	0.7982	0.8586	0.8111	0.9550	0.9311	0.7864
GCN (Kipf and Welling 2016)	0.8393	0.8938	0.8758	0.9068	0.8895	0.7100	0.9299	0.8617	0.5817
GAT (Veličković et al. 2017)	0.8219	0.8759	0.8668	0.8710	0.8558	0.6339	0.9268	0.8525	0.5340
DTI-CNN (Peng et al. 2020)	0.8523	0.9262	0.9340	0.9269	0.9281	0.8286	0.9558	0.9319	0.7957
GCNMDA (Long et al. 2020)	0.8850	0.9424	0.9347	0.9044	0.9354	0.7520	0.9302	0.9423	0.6573
MVGCN (Fu et al. 2022)	0.8489	0.9042	0.9017	0.9132	0.9209	0.7777	0.9445	0.9163	0.6959
MMGCN (Tang et al. 2021)	0.9085	0.9556	0.9122	0.9403	0.9671	0.8038	0.9582	0.9715	0.7684
GraphCDA (Dai et al. 2022)	0.8796	0.9459	0.9471	0.9221	0.9484	0.8353	0.9377	0.9133	0.6435
DTINet (Luo et al. 2017)	0.8672	0.9390	0.9432	0.8983	0.9017	0.8511	0.9029	0.9003	0.7883
MIDTI (ours)	0.9340	0.9787	0.9701	0.9413	0.9813	0.9075	0.9539	0.9794	0.8431

Table 1.

Open in new tab

The performance of MIDTI as well as other baseline approaches for predicting DTI under different ratios on Luo’s dataset.^a

Models	1:1			1:5			1:10
	ACC	AUC	AUPR	ACC	AUC	AUPR	ACC	AUC	AUPR
RF (Pedregosa et al. 2011)	0.8409	0.9016	0.9129	0.9103	0.9093	0.7836	0.9438	0.9176	0.7156
SVM (Chang and Lin 2011)	0.7993	0.8586	0.8111	0.9074	0.8917	0.6962	0.9380	0.8871	0.6078
XGBoost (Chen and Guestrin 2016)	0.8573	0.9238	0.9323	0.7982	0.8586	0.8111	0.9550	0.9311	0.7864
GCN (Kipf and Welling 2016)	0.8393	0.8938	0.8758	0.9068	0.8895	0.7100	0.9299	0.8617	0.5817
GAT (Veličković et al. 2017)	0.8219	0.8759	0.8668	0.8710	0.8558	0.6339	0.9268	0.8525	0.5340
DTI-CNN (Peng et al. 2020)	0.8523	0.9262	0.9340	0.9269	0.9281	0.8286	0.9558	0.9319	0.7957
GCNMDA (Long et al. 2020)	0.8850	0.9424	0.9347	0.9044	0.9354	0.7520	0.9302	0.9423	0.6573
MVGCN (Fu et al. 2022)	0.8489	0.9042	0.9017	0.9132	0.9209	0.7777	0.9445	0.9163	0.6959
MMGCN (Tang et al. 2021)	0.9085	0.9556	0.9122	0.9403	0.9671	0.8038	0.9582	0.9715	0.7684
GraphCDA (Dai et al. 2022)	0.8796	0.9459	0.9471	0.9221	0.9484	0.8353	0.9377	0.9133	0.6435
DTINet (Luo et al. 2017)	0.8672	0.9390	0.9432	0.8983	0.9017	0.8511	0.9029	0.9003	0.7883
MIDTI (ours)	0.9340	0.9787	0.9701	0.9413	0.9813	0.9075	0.9539	0.9794	0.8431

Models	1:1			1:5			1:10
	ACC	AUC	AUPR	ACC	AUC	AUPR	ACC	AUC	AUPR
RF (Pedregosa et al. 2011)	0.8409	0.9016	0.9129	0.9103	0.9093	0.7836	0.9438	0.9176	0.7156
SVM (Chang and Lin 2011)	0.7993	0.8586	0.8111	0.9074	0.8917	0.6962	0.9380	0.8871	0.6078
XGBoost (Chen and Guestrin 2016)	0.8573	0.9238	0.9323	0.7982	0.8586	0.8111	0.9550	0.9311	0.7864
GCN (Kipf and Welling 2016)	0.8393	0.8938	0.8758	0.9068	0.8895	0.7100	0.9299	0.8617	0.5817
GAT (Veličković et al. 2017)	0.8219	0.8759	0.8668	0.8710	0.8558	0.6339	0.9268	0.8525	0.5340
DTI-CNN (Peng et al. 2020)	0.8523	0.9262	0.9340	0.9269	0.9281	0.8286	0.9558	0.9319	0.7957
GCNMDA (Long et al. 2020)	0.8850	0.9424	0.9347	0.9044	0.9354	0.7520	0.9302	0.9423	0.6573
MVGCN (Fu et al. 2022)	0.8489	0.9042	0.9017	0.9132	0.9209	0.7777	0.9445	0.9163	0.6959
MMGCN (Tang et al. 2021)	0.9085	0.9556	0.9122	0.9403	0.9671	0.8038	0.9582	0.9715	0.7684
GraphCDA (Dai et al. 2022)	0.8796	0.9459	0.9471	0.9221	0.9484	0.8353	0.9377	0.9133	0.6435
DTINet (Luo et al. 2017)	0.8672	0.9390	0.9432	0.8983	0.9017	0.8511	0.9029	0.9003	0.7883
MIDTI (ours)	0.9340	0.9787	0.9701	0.9413	0.9813	0.9075	0.9539	0.9794	0.8431

The best results are marked in bold and the second best is underlined.

This Feature Is Available To Subscribers Only