Node Classification
ノード分類は、グラフデータ分析における機械学習タスクの一つで、グラフ内のノードの属性とそれらの関係性を基に未知のノード属性を予測することを目指しています。このタスクは、グラフニューラルネットワークなどのモデルを用いて実装され、その性能は通常、Cora、Citeseer、Pubmedなどのベンチマークデータセットを用いて精度(Accuracy)やF1スコアなどで評価されます。ノード分類は、ソーシャルネットワーク分析、推薦システム、バイオインフォマティクスなど、多くの分野で重要な応用価値を持っています。
20NEWS
GraRep
Actor
AIFB
R-GCN
AM
BoP
Amazon Computers
GAT
Amazon-Fraud
GTAN
Amazon Photo
GraphSAGE
amazon-ratings
Amazon2M
Cluster-GCN
AMPLUS
R-GCN
AMZ Comp
HH-GCN
AMZ Computers
CPF-ind-GAT
AMZ Computers: fixed 20 node per class
AMZ Photo
Exphormer
arXiv-year
FaberNet
AVA
ASDNet [ASDNet_ICCV2021]
BGP
PathNet
BGS
SCENE
Bitcoin-Alpha
GraphMix (GCN)
Bitcoin-OTC
GraphMix (GCN)
BlogCatalog
ISNE
Brazil Air-Traffic
CellTypeGraph Benchmark
plantcelltype-EdgeDeeperGCN
Chameleon
DJ-GNN
Chameleon (48%/32%/20% fixed splits)
GREAD-BS
Chameleon (60%/20%/20% random splits)
Chameleon(60%/20%/20% random splits)
Citeseer
TransGNN
CiteSeer (0.5%)
MT-GCN
CiteSeer (1%)
VCHN
Citeseer (48%/32%/20% fixed splits)
Geom-GCN
CiteSeer (60%/20%/20% random splits)
Citeseer: fixed 20 node per class
Citeseer Full-supervised
IncepGCN+DropEdge
Citeseer random partition
GraphMix (GCN)
CiteSeer with Public Split: fixed 20 nodes per class
CiteSeer with Public Split: fixed 5 nodes per class
GraphMix (GCN)
CLUSTER
GRIT
Coauthor CS
CoLinkDist
Coauthor Phy
GCN-LPA
Coauthor Physics
CoLinkDist
COCO-SP
NeuralWalker
Cora
SSP
Cora (0.5%)
CPF-ind_APPNP
Cora (1%)
VHCN
Cora (3%)
Truncated Krylov
Cora (48%/32%/20% fixed splits)
NLGAT
Cora (60%/20%/20% random splits)
Cora: fixed 10 node per class
CPF-tra-GCNII
Cora: fixed 20 node per class
Self-supervised GraphMAE
Cora: fixed 20 nodes per class
Cora: fixed 5 node per class
CPF-tra-APPNP
Cora Full
CoLinkDist
Cora Full-supervised
GCNII
Cora Full with Public Split
CoLinkDist
Cora random partition
GraphMix (GCN)
Cora with Public Split: fixed 20 nodes per class
OGC
Cornell
Cornell (48%/32%/20% fixed splits)
Cornell (60%/20%/20% random splits)
HH-GraphSAGE
Crocodile
LW-GCN
DBLP
GRACE
Deezer Croatia
Deezer Hungary
Deezer Romania
DMG777K
RR-GCN-PPV
DMGFULL
RR-GCN-PPV
Electronics
Europe Air-Traffic
GCN_cheby (Kipf and Welling, 2017)
Facebook
GNNMoE(GCN-like P)
Film(48%/32%/20% fixed splits)
Film (60%/20%/20% random splits)
Flickr
GCN+GAugM (Zhao et al., 2021)
genius
GESN
MAG-scholar-C
FastGCN
MAG240M-LSC
MDGENRE
minesweeper
MS ACADEMIC
APPNP
MuMiN-large
MuMiN-medium
MuMiN-small
LaBSE
MUTAG
RR-GCN-PPV
NBA
NELL
DFNet-ATT
PascalVOC-SP
SAN+LapPE
PATTERN
CKGCN
PATTERN 100k
EGT
Penn94
GloGNN++
Placenta
GraphSAGE
pokec
NeuralWalker
PPI
g2-MLP
Pubmed
ClusterGCN
PubMed (0.03%)
Truncated Krylov
PubMed (0.05%)
PubMed (0.1%)
Truncated Krylov
PubMed (48%/32%/20% fixed splits)
GCNII
PubMed (60%/20%/20% random splits)
Pubmed: fixed 20 node per class
Pubmed Full-supervised
GraphSAGE+DropEdge
Pubmed random partition
GraphMix (GCN)
PubMed with Public Split: fixed 20 nodes per class
OGC
questions
GCN
Reddit
BNS-GCN
roman-empire
Squirrel
Squirrel (48%/32%/20% fixed splits)
Squirrel (60%/20%/20% random splits)
Squirrel(60%/20%/20% random splits)
NCSAGE
Telegram (Directed Graph label rate 60%)
ScaleNet
Texas
2-HiGCN
Texas (48%/32%/20% fixed splits)
Texas (60%/20%/20% random splits)
ACM-GCN++
tolokers
Polynormer
twitch-gamers
Dual-Net GNN
USA Air-Traffic
DEMO-Net(weight)
Wiki
DAOR
wiki
A2DUG
Wiki-CS
CGT
Wiki-Vote
DEMO-Net(weight)
Wikipedia
GraphGAN
Wisconsin
5-HiGCN
Wisconsin (48%/32%/20% fixed splits)
Wisconsin (60%/20%/20% random splits)
ACM-GCN++
Yelp-Fraud
YouTube
LINE
DAOR