Skip to main content

Table 1 Edge- and User-level Performances in Entry_DS

From: Unveiling the silent majority: stance detection and characterization of passive users on social media using collaborative filtering and graph convolutional networks

Models

Edge

User

nDCG

MAP

Acc.

RMSE

Least

Middle

Most

Null Model

0.0176

0.0066

0.4628

0.7858

0.3672

0.4559

0.5474

CSMF

0.0457

0.0180

0.3929

0.7792

0.4061

0.4015

0.3717

NGCF

0.0539

0.0192

0.7192

0.5300

0.5862

0.7530

0.8184

LightGCN

0.0879

0.0377

0.7387

0.5079

0.5706

0.7389

0.9066

WLGCN (norm by user)

0.0728

0.0307

0.7445

0.4946

0.5955

0.7373

0.9006

WLGCN (norm by TF-IDF)

0.0668

0.0271

0.7399

0.4951

0.6004

0.7329

0.8864

WLGCN

0.0889

0.0378

0.7589

0.4891

0.6073

0.7593

0.9101

WLGCN + (\(E_{\mathrm{HT}}\))

0.0896

0.0381

0.7678

0.4803

0.6367

0.7730

0.8938

WLGCN + (\(\mathcal{G}_{\mathrm{PathSim}}\))

0.0860

0.0364

0.7673

0.4809

0.6391

0.7590

0.9038

WLGCN + (\(E_{\mathrm{HT}}\), \(\mathcal{G}_{\mathrm{PathSim}}\))

\(\textbf{0.0921}^{*}\)

\(\textbf{0.0393}^{*}\)

0.7915

0.4540

0.6772

\(\textbf{0.8010}^{*}\)

0.8963

WLGCN + (\(\mathcal{G}_{\mathrm{Soc}}\))

0.0857+

0.0357+

0.7704

0.4790

0.6479+

0.7569

0.9064

WLGCN + (\(E_{\mathrm{HT}}\), \(\mathcal{G}_{\mathrm{Soc}}\))

0.0865

0.0356

0.7741

0.4738+

0.6314

0.7761

\(\textbf{0.9147}^{*}\)

WLGCN + (\(\mathcal{G}_{\mathrm{PathSim}}\), \(\mathcal{G}_{\mathrm{Soc}}\))

0.0838

0.0357

0.7757

0.4720

0.6539

0.7710

0.9022

WLGCN + (\(E_{\mathrm{HT}}\), \(\mathcal{G}_{\mathrm{PathSim}}\), \(\mathcal{G}_{\mathrm{Soc}}\))

0.0917

0.0392+

\(\textbf{0.7954}^{*}\)

\(\textbf{0.4507}^{*}\)

\(\textbf{0.6895}^{*}\)

0.7988

0.8977

  1. We marked WLGCN variants’ results that are statistically significant (t-test) compared to the baseline LightGCN: (p<0.05), +(p<0.1).