Table 1 Shows important findings from the literature.

From: An efficient learning based approach for automatic record deduplication with benchmark datasets

Reference

Method

Embedding Model

Alignment Model

Extra information

Prediction

CL

ML

22

AttrE

TransE

A

Attribute

CS

NO

Yes

60

BootEA

TransE variant

C

Bootstrapping

CS

Yes

Yes

61

CEA

GCN

M

Entity name

CS

Yes

Yes

62

GCN-Align

GCN

M

Attribute

MD

YES

NO

63

GM-Align

GCN

G

Entity name

MP

Yes

No

21

HGCN

GCN

M

Entity name

MD

Yes

No

64

HUMAN

GCN

M

Description, Attribute

ED

Yes

No

16

ITransE

TransE

T

Bootstrapping

ED

No

Yes

65

JAPE

TransE variant

A

Attribute

CS

Yes

No

66

KDCoE

TransE

T

Entity description

ED

Yes

No

15

KECG

GAT + TransE

M

No

ED

Yes

Yes

13

MTransE

TransE variant

T

No

ED

Yes

No

10

MuGNN

GNN

M

No

CS

Yes

Yes

67

MultiKE

TransE variant

C

Entity name

CS

No

Yes

68

NAEA

TransE variant

C

Bootstrapping

CS

Yes

Yes

69

RDGCN

DPGCNN

M

Entity name

MD

Yes

No

14

RSNs

RSNs

C

No

CS

Yes

Yes

70

TransEdge

TransEdge

C

Bootstrapping

CS

Yes

Yes

  1. (C-corpus function, M-margin based, A-attribute refined, G-graph based, T-transition, ED-Euclidian distance, CS-cosine similarity, MD-Manhattan distance, MP-matching probability, CL for cross-lingual and ML for mono-lingual).