Skip to content

Latest commit

 

History

History
572 lines (481 loc) · 51.3 KB

eCommerceData.md

File metadata and controls

572 lines (481 loc) · 51.3 KB

E-Commerce Data

  • URL: https://www.kaggle.com/datasets/carrie1/ecommerce-data
  • Format: CSV
  • Description: This is a transnational data set which contains all the transactions occurring between 01/12/2010 and 09/12/2011 for a UK-based and registered non-store online retail.The company mainly sells unique all-occasion gifts. Many customers of the company are wholesalers.

Datasets

  • data.csv

    • 541909 rows x 8 columns

  • processed_data20K.csv

    • 20000 rows x 13 columns
    • From above, selection of the first 20000 rows
    • Processing of identifiers for four classes Counts data20K.csv

  • The "Description" and "CustomerID" fields are the only fields that have null values

  • No field has only unique values

Fields cardinality data20K.csv

  • (0,0) No relationship between values in the row and column fields
  • (0,1) One value in the row field is related to at most one value in the column field
  • (0,2) One value in the row field can be related to none, one or several values in the column field
  • (1,1) One value in the row filed is only related to one value in the column field
  • (1,2) One value in the row field can be related to one or several values in the column field
  • Inconsistencies:
    • The "InvoiceNo" field should have a 1 to 1 relationship with the "InvoiceDate" field.
    • The "StockCode" field should have a 1 to 1 relationship with the "Description" field and vice versa

Ontologies

  • Basic
    • One class
    • Zero object properties
    • Eight data properties
  • LLM
    • Three classes
    • Two object properties
    • Six data properties
  • Gold standard
    • Six classes
    • Six object properties
    • Eight data properties
  • Materials
    • Two classes
    • One object property
    • Two data properties

Mapping of the classes, their relationships and attributes between the different ontologies, as well as the Hits@1 metric obtained in the alignment experiments for the AttrE, BootEA, Aligne, GCN_Align and SEA methods.

Class mapping

Mapping

Mapping Dataset RDF attr_triples rel_triples
Basic processed_data20K.csv 54553 triples 53615 triples 938 triples
Gold processed_data20K.csv 95666 triples 10081 triples 85585 triples
LLM processed_data20K.csv 43164 triples 15798 triples 27366 triples
Materials processed_data20K.csv 46073 triples 4815 triples 41258 triples

Gold - Gold

High level graph generated for the Gold ontology. Nodes of the same color represent the instance, class type and attributes of the same class. The instance node indicates the number of entities in the test set. Blue arrows indicate attributes and black arrows indicate relationships. The cardinality is determined for each arrow. If it is not (1,1), the probability that it was (1,1) is indicated. The names in parentheses indicate the referenced column in the CSV source file.

Graph Gold

Input

ent_links Related Entities Split 1 Test
https://vocab.um.es/ontology/ecd/Invoice - https://vocab.um.es/ontology/ecd/Invoice 938 (3.79%) 377 (3.81%)
https://vocab.um.es/ontology/ecd/salesArticle - https://vocab.um.es/ontology/ecd/salesArticle 19425 (78.47%) 7817 (78.95%)
https://vocab.um.es/ontology/ecd/SalesSpecification - https://vocab.um.es/ontology/ecd/SalesSpecification 1407 (5.68%) 561 (5.67%)
https://vocab.um.es/ontology/ecd/Product - https://vocab.um.es/ontology/ecd/Product 2408 (9.73%) 947 (9.56%)
https://vocab.um.es/ontology/ecd/Customer - https://vocab.um.es/ontology/ecd/Customer 560 (2.26%) 192 (1.94%)
https://vocab.um.es/ontology/ecd/Country - https://vocab.um.es/ontology/ecd/Country 16 (0.07%) 7 (0.07%)
Total related entities 24754 9901
Split Training (50%) Test (40%) Validation (10%)
1 12377 9901 2476

Output

Approach Split Aligned Ent H@1 (%) H@5 (%) H@10 (%) MR MRR [0,1] Run time Invoice (#/%) salesArticle (#/%) SalesSpecification (#/%) Product (#/%) Customer (#/%) Country (#/%)
AlignE 1 9901 98.47 99.66 99.88 1 0.99 21406 365 (96.82%) 7750 (99.14%) 547 (97.50%) 890 (93.98%) 190 (98.96%) 7 (100%)
AliNet 1 9901 94.46 97.57 97.87 23 0.96 8656 - - - - - -
AttrE 1 9901 99.52 99.97 100 1 1.00 9264 372 (98.67%) 7783 (99.57%) 555 (98.93%) 945 (99.79%) 191 (99.48%) 7 (100%)
BootEA 1 9901 99.63 99.90 99.95 1 1.00 21594 375 (99.47%) 7804 (98.83%) 558 (99.47%) 928 (97.99%) 192 (100%) 7 (100%)
BootEA_RotatE 1 9901 99.61 100 100 1 1.00 35040 - - - - - -
BootEA_TransH 1 9901 E r r o r
Conve 1 9901 E r r o r
GCN_Align 1 9901 94.71 99.68 99.74 1 0.97 1700 375 (99.47%) 7336 (93.85%) 555 (98.93%) 912 (96.30%) 192 (100%) 7 (100%)
GMNN 1 9901 E r r o r
HolE 1 9901 89.33 93.50 94.38 37 0.91 36516 - - - - - -
IMUSE 1 9901 94.62 98.76 99.16 2 0.97 2603 - - - - - -
IPTransE 1 9901 96.39 98.83 99.27 1 0.97 9568 - - - - - -
JAPE 1 9901 67.11 82.33 85.10 255 0.74 4321 - - - - - -
KDCoE 1 9901 E r r o r
MTransE 1 9901 40.85 55.10 58.17 898 0.47 2974 - - - - - -
MultiKE 1 9901 E r r o r
ProjE 1 9901 55.53 78.99 80.86 76 0.66 15925 - - - - - -
RDGCN 1 9901 76.25 90.33 92.55 10 0.83 77837 - - - - - -
RotatE 1 9901 99.40 100 100 1 1.00 16049 - - - - - -
RSN4EA 1 9901 95.79 98.73 99.07 5 0.97 54845 - - - - - -
SEA 1 9901 97.58 99.85 99.98 1 0.99 3143 367 (97.35%) 7623 (97.52%) 554 (98.75%) 930 (98.20%) 180 (93.75%) 7 (100%)
SimplE 1 9901 91.89 95.16 96.19 3 0.94 2760 - - - - - -
TransD 1 9901 94.45 96.79 97.44 5 0.96 10276 - - - - - -
TransH 1 9901 95.90 98.26 98.88 2 0.97 8940 - - - - - -
TransR 1 9901 0.32 1.33 2.35 2082 0.01 8775 - - - - - -

Basic - Basic

High level graph generated for the Gold ontology. Nodes of the same color represent the instance, class type and attributes of the same class. The instance node indicates the number of entities in the test set. Blue arrows indicate attributes and black arrows indicate relationships. The cardinality is determined for each arrow. If it is not (1,1), the probability that it was (1,1) is indicated. The names in parentheses indicate the referenced column in the CSV source file.

Graph Basic

Input

ent_links Related Entities Split 1 Test
https://vocab.um.es/ontology/ecd/Invoice - https://vocab.um.es/ontology/ecd/Invoice 938 (100%) 375 (100%)
Total related entities 938 375
Split Training (50%) Test (40%) Validation (10%)
1 469 375 94

Output

Approach Split Aligned Ent H@1 (%) H@5 (%) H@10 (%) MR MRR [0,1] Run time Invoice (#/%)
AlignE 1 375 0.27 0.80 1.87 191 0.02 110 1 (0.27%)
AliNet 1 375 E r r o r
AttrE 1 375 91.20 97.6 100 1 0.94 495 342 (91.20%)
BootEA 1 375 0.80 2.67 4.53 191 0.03 121 3 (0.80%)
BootEA_RotatE 1 375 0.53 2.13 3.73 185 0.02 454 -
BootEA_TransH 1 375 E r r o r
Conve 1 375 E r r o r
GCN_Align 1 375 0.00 0.27 1.60 189 0.01 18 0 (0.00%)
GMNN 1 375 100 100 100 1 1.00 2997 -
HolE 1 375 0.53 1.07 2.40 191 0.02 158 -
IMUSE 1 375 27.47 28.27 28.53 140 0.29 25 -
IPTransE 1 375 E r r o r
JAPE 1 375 0.80 1.07 3.20 191 0.02 91 -
KDCoE 1 375 E r r o r
MTransE 1 375 0.53 1.33 2.13 194 0.02 29 -
MultiKE 1 375 E r r o r
ProjE 1 375 E r r o r
RDGCN 1 375 0.27 1.33 2.67 188 0.02 814 -
RotatE 1 375 0.27 1.87 3.73 193 0.02 175 -
RSN4EA 1 375 E r r o r
SEA 1 375 0.00 1.33 1.87 190 0.01 24 0 (0.0%)
SimplE 1 375 0.00 0.53 1.33 189 0.01 43 -
TransD 1 375 0.00 1.07 1.07 195 0.01 47 -
TransH 1 375 0.27 1.60 3.47 189 0.02 48 -
TransR 1 375 0.27 1.33 2.67 188 0.02 80 -

LLM - LLM

High level graph generated for the Gold ontology. Nodes of the same color represent the instance, class type and attributes of the same class. The instance node indicates the number of entities in the test set. Blue arrows indicate attributes and black arrows indicate relationships. The cardinality is determined for each arrow. If it is not (1,1), the probability that it was (1,1) is indicated. The names in parentheses indicate the referenced column in the CSV source file.

Graph LLM

Input

ent_links Related Entities Split 1 Test
https://vocab.um.es#invoice - https://vocab.um.es#invoice 938 (26.40%) 383 (26.95%)
https://vocab.um.es#salesArticle - https://vocab.um.es#salesArticle 2408 (67.77%) 952 (67.00%)
http://schema.org/UnitPriceSpecification - http://schema.org/UnitPriceSpecification 207 (5.83%) 86 (6.05%)
Total related entities 3553 1421
Split Training (50%) Test (40%) Validation (10%)
1 1776 1421 356

Output

Approach Split Aligned Ent H@1 (%) H@5 (%) H@10 (%) MR MRR [0,1] Run time Invoice (#/%) salesArticle (#/%) UnitPriceSpecification (#/%)
AlignE 1 1421 94.93 98.45 98.80 2 0.97 2597 351 (91.64%) 921 (96.74%) 77 (89.53%)
AliNet 1 1421 86.98 94.23 95.71 4 0.90 1278 - - -
AttrE 1 1421 96.34 99.51 99.65 1 0.98 1370 360 (93.99%) 937 (98.42%) 72 (83.72%)
BootEA 1 1421 95.22 98.94 99.51 1 0.97 2167 358 (93.47%) 921 (96.74%) 74 (86.05%)
BootEA_RotatE 1 1421 96.13 99.23 99.79 1 0.98 9076 - - -
BootEA_TransH 1 1421 E r r o r
Conve 1 1421 E r r o r
GCN_Align 1 1421 88.46 94.23 95.43 4 0.91 102 340 (88.77%) 841 (88.34%) 76 (88.37%)
GMNN 1 1421 100 100 100 1 1.00 13676 - - -
HolE 1 1421 88.74 96.06 97.61 4 0.92 13609 - - -
IMUSE 1 1421 87.26 96.76 98.38 2 0.92 738 - - -
IPTransE 1 1421 E r r o r
JAPE 1 1421 21.46 37.65 46.38 59 0.30 541 - - -
KDCoE 1 1421 E r r o r
MTransE 1 1421 13.02 19.07 22.24 294 0.16 449 - - -
MultiKE 1 1421 E r r o r
ProjE 1 1421 37.23 56.02 66.08 22.57 0.47 4402 - - -
RDGCN 1 1421 71.22 87.61 93.60 3 0.79 2869 - - -
RotatE 1 1421 96.48 99.16 99.93 1 0.98 3208 - - -
RSN4EA 1 1421 91.13 97.47 98.03 3 0.94 13963 - - -
SEA 1 1421 95.43 99.16 99.93 1 0.97 459 355 (92.69%) 924 (97.06%) 77 (89.53%)
SimplE 1 1421 40.46 68.33 80.08 8 0.53 413 - - -
TransD 1 1421 82.83 91.84 94.16 5 0.87 1319 - - -
TransH 1 1421 88.04 95.36 97.04 2 0.91 1596 - - -
TransR 1 1421 1.9 5.98 7.74 308 0.05 1942 - - -

Materials - Materials

High level graph generated for the Gold ontology. Nodes of the same color represent the instance, class type and attributes of the same class. The instance node indicates the number of entities in the test set. Blue arrows indicate attributes and black arrows indicate relationships. The cardinality is determined for each arrow. If it is not (1,1), the probability that it was (1,1) is indicated. The names in parentheses indicate the referenced column in the CSV source file.

Graph Materials

Input

ent_links Related Entities Split 1 Test
https://purl.org/ontologies/MT/SalesArticle - https://purl.org/ontologies/MT/SalesArticle 19425 (88.97%) 7780 (89.09%)
https://purl.org/ontologies/MT/sales_product - https://purl.org/ontologies/MT/sales_product 2408 (11.03%) 953 (10.91%)
Total related entities 21833 8733
Split Training (50%) Test (40%) Validation (10%)
1 10916 8733 2184

Output

Approach Split Aligned Ent H@1 (%) H@5 (%) H@10 (%) MR MRR [0,1] Run time salesArticle (#/%) Product (#/%)
AlignE 1 8733 29.03 68.75 86.00 103 0.46 7044 1722 (22.13%) 813 (85.31%)
AliNet 1 8733 29.60 68.95 85.48 176 0.46 5316 - -
AttrE 1 8733 27.91 66.19 83.39 79 0.44 3134 1599 (20.55%) 838 (87.93%)
BootEA 1 8733 29.11 68.72 86.15 109 0.46 9055 1728 (22.21%) 814 (85.41%)
BootEA_RotatE 1 8733 29.21 69.44 86.43 39 0.46 14761 - -
BootEA_TransH 1 8733 E r r o r
Conve 1 8733 E r r o r
GCN_Align 1 8733 29.73 69.19 86.79 8 0.47 1474 1779 (22.87%) 817 (85.73%)
GMNN 1 8733 100 100 100 1 1.00 11541
HolE 1 8733 25.20 58.75 75.42 139 0.40 12896 - -
IMUSE 1 8733 28.43 67.37 84.46 58 0.45 969 - -
IPTransE 1 8733 E r r o r
JAPE 1 8733 E r r o r
KDCoE 1 8733 E r r o r
MTransE 1 8733 24.92 58.47 73.56 361 0.40 1494 - -
MultiKE 1 8733 E r r o r
ProjE 1 8733 20.26 59.90 77.01 120 0.37 7861 - -
RDGCN 1 8733 14.62 36.46 49.14 712 0.25 13482 - -
RotatE 1 8733 29.65 68.67 86.15 96 0.46 6673 - -
RSN4EA 1 8733 29.70 68.69 85.70 143 0.46 19243 - -
SEA 1 8733 29.88 68.87 86.13 98 0.47 1180 1796 (23.08%) 813 (85.31%)
SimplE 1 8733 28.24 67.23 85.23 57 0.45 970 - -
TransD 1 8733 22.67 49.61 64.29 245 0.35 2317 - -
TransH 1 8733 26.44 61.61 78.08 140 0.42 2834 - -
TransR 1 8733 0.08 0.40 0.76 2939 0.01 4534 - -

Gold - Basic

Graph Gold-Basic

Input

ent_links Related Entities Split 1 Test
https://vocab.um.es/ontology/ecd/Invoice - https://vocab.um.es/ontology/ecd/Invoice 938 (100%) 375 (100%)
Total related entities 938 375
Split Training (50%) Test (40%) Validation (10%)
1 469 375 94

Output

Approach Split Aligned Ent H@1 (%) H@5 (%) H@10 (%) MR MRR [0,1] Run time Invoice (#/%)
AlignE 1 375 0.8 1.33 2.40 180 0.02 2123 3 (0.80%)
AliNet 1 375 E r r o r
AttrE 1 375 1.33 5.33 11.2 132 0.05 1254 5 (1.33)
BootEA 1 375 0.0 1.07 1.60 193 0.01 2117 0 (0.00%)
BootEA_RotatE 1 375 0.8 2.13 3.20 188 0.02 6864 -
BootEA_TransH 1 375 E r r o r
Conve 1 375 E r r o r
GCN_Align 1 375 0.0 1.60 3.20 188 0.02 215 0 (0.00%)
GMNN 1 375 98.67 100 100 1 0.99 20177 -
HolE 1 375 0.27 1.60 3.20 187 0.02 4278 -
IMUSE 1 375 52.27 60.53 64.53 60 0.56 449 -
IPTransE 1 375 0.0 1.01 1.87 187 0.01 731 -
JAPE 1 375 0.53 1.33 2.93 187 0.02 747 -
KDCoE 1 375 E r r o r
MTransE 1 375 0.53 1.33 2.40 190 0.02 659 -
MultiKE 1 375 E r r o r
ProjE 1 375 0.27 1.60 1.87 181 0.02 3128 -
RDGCN 1 375 0.27 1.33 3.47 166 0.02 1847 -
RotatE 1 375 0.27 1.33 2.40 184 0.02 3800 -
RSN4EA 1 375 E r r o r
SEA 1 375 0.27 0.53 1.87 184 0.02 266 1 (0.27%)
SimplE 1 375 0.0 2.13 4.0 187 0.02 237 -
TransD 1 375 0.53 1.60 2.40 187 0.02 796 -
TransH 1 375 0.0 1.07 2.13 186 0.01 696 -
TransR 1 375 0.53 2.40 3.47 194 0.02 2197 -

Gold - LLM

Graph Gold-LLM

Input

ent_links Related Entities Split 1 Test
https://vocab.um.es/ontology/ecd/Invoice - https://vocab.um.es#invoice 938 (4.31%) 364 (4.18%)
https://vocab.um.es/ontology/ecd/salesArticle - https://vocab.um.es#salesArticle 19425 (89.23%) 7802 (89.60%)
https://vocab.um.es/ontology/ecd/SalesSpecification - http://schema.org/UnitPriceSpecification 1407 (6.46%) 542 (6.22%)
Total related entities 21770 8708
Split Training (50%) Test (40%) Validation (10%)
1 10885 8708 2177

Output

Approach Split Aligned Ent H@1 (%) H@5 (%) H@10 (%) MR MRR [0,1] Run time Invoice (#/%) salesArticle (#/%) SalesSpecification (#/%)
AlignE 1 8708 15.81 42.43 59.75 27 0.29 19146 296 (81.32%) 4449 (57.02%) 161 (29.70%)
AliNet 1 8708 E r r o r
AttrE 1 8708 17.33 45.90 60.85 31 0.31 3640 308 (84.62%) 3673 (47.08%) 172 (31.73%)
BootEA 1 8708 14.23 39.62 55.71 30 0.27 18663 291 (79.95%) 4214 (54.01%) 160 (29.52%)
BootEA_RotatE 1 8708 16.74 47.20 65.41 25 0.31 33143 - - -
BootEA_TransH 1 8708 E r r o r
Conve 1 8708 E r r o r
GCN_Align 1 8708 10.57 33.34 48.51 186 0.22 888 106 (29.12%) 3776 (48.40%) 127 (23.43%)
GMNN 1 8708 E r r o r
HolE 1 8708 4.42 13.56 20.42 432 0.10 14042 - - -
IMUSE 1 8708 52.27 60.53 64.53 60 0.56 449 - - -
IPTransE 1 8708 11.77 32.40 42.80 193 0.22 2372 - - -
JAPE 1 8708 11.02 23.79 27.40 1835 0.17 2055 - - -
KDCoE 1 8708 E r r o r
MTransE 1 8708 19.36 55.11 70.36 263 0.35 973 - - -
MultiKE 1 8708 E r r o r
ProjE 1 8708 19.07 56.75 72.38 133 0.35 9444 - - -
RDGCN 1 8708 5.87 25.32 40.65 789 0.16 14299 - - -
RotatE 1 8708 12.02 31.73 44.41 72 0.22 7652 - - -
RSN4EA 1 8708 E r r o r
SEA 1 8708 19.40 52.10 69.87 18 0.34 1046 327 (89.84%) 4920 (63.06%) 225 (41.51%)
SimplE 1 8708 8.00 29.18 42.43 347 0.18 1246 - - -
TransD 1 8708 E r r o r
TransH 1 8708 12.72 35.45 47.67 118 0.24 3732 - - -
TransR 1 8708 0.21 0.77 1.32 2451 0.01 4278 - - -

Gold - Materials

Graph Gold-Materials

Input

ent_links Related Entities Split 1 Test
https://vocab.um.es/ontology/ecd/salesArticle - https://purl.org/ontologies/MT/SalesArticle 19425 (88.97%) 7785 (89.14%)
https://vocab.um.es/ontology/ecd/Product - https://purl.org/ontologies/MT/sales_product 2408 (11.03%) 948 (10.86%)
Total related entities 21833 8733
Split Training (50%) Test (40%) Validation (10%)
1 10916 8733 2184

Output

Approach Split Aligned Ent H@1 (%) H@5 (%) H@10 (%) MR MRR [0,1] Run time salesArticle (#/%) Product (#/%)
AlignE 1 8733 27.43 66.07 82.87 149 0.44 16219 1598 (20.53%) 797 (84.07%)
AliNet 1 8733 29.01 68.63 84.15 140 0.46 11314 - -
AttrE 1 8733 23.62 55.90 72.24 153 0.38 4462 1269 (16.30) 794 (83.76%)
BootEA 1 8733 27.78 67.84 83.88 153 0.45 13993 1628 (20.91%) 798 (84.18%)
BootEA_RotatE 1 8733 26.80 65.03 82.60 112 0.43 31140 - -
BootEA_TransH 1 8733 E r r o r
Conve 1 8733 E r r o r
GCN_Align 1 8733 27.04 66.13 82.88 146 0.44 1059 1588 (20.40%) 773 (81.54%)
GMNN 1 8733 100 100 100 1 1.00 20826 - -
HolE 1 8733 11.44 28.83 39.90 291 0.20 19268 - -
IMUSE 1 8733 E r r o r
IPTransE 1 8733 23.98 55.30 70.25 276 0.38 4063 - -
JAPE 1 8733 12.54 23.33 28.96 1367 0.18 2048 - -
KDCoE 1 8733 E r r o r
MTransE 1 8733 6.79 9.88 11.92 1775 0.09 1489 - -
MultiKE 1 8733 E r r o r
ProjE 1 8733 19.32 58.95 74.79 167 0.36 21328 - -
RDGCN 1 8733 7.97 25.76 36.28 1271 0.16 36333 - -
RotatE 1 8733 26.13 63.86 81.12 139 0.43 16412 - -
RSN4EA 1 8733 28.35 68.30 84.32 169 0.45 30261 - -
SEA 1 8733 27.23 64.59 81.26 140 0.43 3639 1584 (20.35%) 794 (83.76%)
SimplE 1 8733 24.26 63.30 80.27 52 0.41 2217 - -
TransD 1 8733 20.88 46.33 58.38 357 0.32 4032 - -
TransH 1 8733 23.52 53.99 68.79 222 0.37 5718 - -
TransR 1 8733 0.10 0.32 0.54 3223 0.01 6244 - -

Basic - LLM

Graph Basic-LLM

Input

ent_links Related Entities Split 1 Test
https://vocab.um.es/ontology/ecd/Invoice - https://vocab.um.es#invoice 938 (100%) 375 (100%)
Total related entities 938 375
Split Training (50%) Test (40%) Validation (10%)
1 469 375 94

Output

Approach Split Aligned Ent H@1 (%) H@5 (%) H@10 (%) MR MRR [0,1] Run time Invoice (#/%)
AlignE 1 375 0.00 0.53 2.13 177 0.01 617 0 (0.00%)
AliNet 1 375 E r r o r
AttrE 1 375 8.00 22.93 35.73 44 0.17 761 30 (8.00%)
BootEA 1 375 0.27 2.13 3.73 182 0.02 641 1 (0.27%)
BootEA_RotatE 1 375 0.27 1.33 3.20 188 0.02 2926 -
BootEA_TransH 1 375 E r r o r
Conve 1 375 E r r o r
GCN_Align 1 375 0.53 1.33 3.47 182 0.02 55 2 (0.53%)
GMNN 1 375 100 100 100 1 1.00 5211 -
HolE 1 375 0.80 2.13 3.20 189 0.02 1534 -
IMUSE 1 375 E r r o r
IPTransE 1 375 E r r o r
JAPE 1 375 0.00 1.60 2.13 191 0.02 212 -
KDCoE 1 375 E r r o r
MTransE 1 375 0.27 1.87 3.47 188 0.02 133 -
MultiKE 1 375 E r r o r
ProjE 1 375 E r r o r
RDGCN 1 375 0.53 2.13 2.93 190 0.02 1115 -
RotatE 1 375 0.00 0.53 2.13 189 0.01 1289 -
RSN4EA 1 375 E r r o r
SEA 1 375 0.00 1.60 3.47 194 0.02 75 0 (0.0%)
SimplE 1 375 0.27 0.80 2.40 187 0.02 110 -
TransD 1 375 0.27 0.80 1.60 189 0.01 223 -
TransH 1 375 0.00 2.13 3.20 195 0.02 212 -
TransR 1 375 0.00 0.80 1.87 194 0.01 756 -

Materials - LLM

Graph Materials-LLM

Input

ent_links Related Entities Split 1 Test
https://purl.org/ontologies/MT/SalesArticle - https://vocab.um.es#salesArticle 19425 (100%) 7770 (100%)
Total related entities 19425 7770
Split Training (50%) Test (40%) Validation (10%)
1 9712 7770 1943

Output

Approach Split Aligned Ent H@1 (%) H@5 (%) H@10 (%) MR MRR [0,1] Run time salesArticle (#/%)
AlignE 1 7770 19.99 63.36 82.23 288 0.39 13961 7295 (93.89%)
AliNet 1 7770 E r r o r
AttrE 1 7770 18.65 60.95 77.21 125 0.37 11907 6344 (81.65%)
BootEA 1 7770 19.50 61.78 80.56 300 0.38 12074 7226 (93.00%)
BootEA_RotatE 1 7770 19.46 63.10 82.06 391 0.38 2982 -
BootEA_TransH 1 7770 E r r o r
Conve 1 7770 E r r o r
GCN_Align 1 7770 16.72 52.97 71.88 218 0.33 1388 1326 (70.16%)
GMNN 1 7770 E r r o r
HolE 1 7770 12.69 37.22 48.26 263 0.24 17829 -
IMUSE 1 7770 E r r o r
IPTransE 1 7770 E r r o r
JAPE 1 7770 14.72 40.14 46.78 1225 0.25 2247 -
KDCoE 1 7770 E r r o r
MTransE 1 7770 19.37 63.23 82.07 205 0.38 916 -
MultiKE 1 7770 E r r o r
ProjE 1 7770 20.12 63.02 81.60 244 0.39 7503 -
RDGCN 1 7770 0.01 0.06 0.14 3653 0.01 13335 -
RotatE 1 7770 18.70 58.92 76.78 219 0.36 8735 -
RSN4EA 1 7770 E r r o r
SEA 1 7770 20.06 63.00 82.15 283 0.39 2778 7299 (93.94%)
SimplE 1 7770 17.79 61.15 80.04 354 0.36 2128 -
TransD 1 7770 15.86 47.58 59.83 432 0.30 5643 -
TransH 1 7770 16.62 45.48 56.81 455 0.29 4411 -
TransR 1 7770 0.10 0.44 0.80 2359 0.01 4226 -

Results

Figure Methods Boxplot. Boxplot of the OpenEA methods according to the Hits@1 metric of the 9 pairwise comparisons of the KGs generated from the E-Commerce Data dataset. The orange solid line represents the median and the red diamond the mean of the experiments that did not lead to an error. The number of experiments without error is shown in brackets next to the name of the method. BootEA-TransH, ConvE, KDCoE and MultiKE are not shown because their error rate is equal to 1 for all datasets.

Modules Boxplot E-Commerce Data

Figure Methods Clustering. Hierarchical clustering (Ward algorithm) of 25 OpenEA methods according to the 5 metrics for the 9 pairwise comparisons of the KGs generated from the E-Commerce Data dataset.

Modules Clusters E-Commerce Data

Figure Experiments Boxplot. Boxplot of the 9 types of pairwise comparison experiments of KGs based on the Hits@1 score obtained by the 25 methods carried out in the E-Commerce Data dataset. The orange solid line represents the median and the red diamond the mean of the experiments that did not lead to an error. The number of modules that did not generate an error is shown in brackets under the name of the experiment represented.

Datasets Boxplot E-Commerce Data

Figure Experiments Clustering. Hierarchical clustering, with method Ward, of the 9 types of pairwise comparison experiments of KG, according to the 5 metrics, with all the values scaled to the range [0,1], of the 25 (at most) OpenEA modules carried out in the E-Commerce Data dataset.

Datasets Clusters E-Commerce Data

Merged methods

Pair Approach Invoice (#/%) salesArticle (#/%) SalesSpecification (#/%) Product (#/%) Customer (#/%) Country (#/%)
Basic-Basic AttrE 342 (91.20%) - - - - -
Basic-Basic BootEA 3 (0.80%) - - - - -
Basic-Basic AlignE 1 (0.27%) - - - - -
Basic-Basic SEA 0 (0.0%) - - - - -
Basic-Basic AttrE-BootEA 342 (91.20%) - - - - -
Basic-Basic AttrE-AlignE 342 (91.20%) - - - - -
Basic-Basic AttrE-SEA 342 (91.20%) - - - - -
Basic-Gold AttrE 5 (1.33%) - - - - -
Basic-Gold BootEA 0 (0.00%) - - - - -
Basic-Gold AlignE 3 (0.80%) - - - - -
Basic-Gold SEA 1 (0.27%) - - - - -
Basic-Gold AttrE-BootEA 5 (1.33%) - - - - -
Basic-Gold AttrE-AlignE 8 (2.13%) - - - - -
Basic-Gold AttrE-SEA 6 (1.60%) - - - - -
Basic-LLM AttrE 30 (8.00%) - - - - -
Basic-LLM BootEA 1 (0.27%) - - - - -
Basic-LLM AlignE 0 (0.00%) - - - - -
Basic-LLM SEA 0 (0.0%) - - - - -
Basic-LLM AttrE-BootEA 31 (8.27%) - - - - -
Basic-LLM AttrE-AlignE 30 (8.00%) - - - - -
Basic-LLM AttrE-SEA 30 (8.00%) - - - - -
Gold-Gold AttrE 372 (98.67%) 7783 (99.57%) 555 (98.93%) 945 (99.79%) 191 (99.48%) 7 (100%)
Gold-Gold BootEA 375 (99.47%) 7804 (98.83%) 558 (99.47%) 928 (97.99%) 192 (100%) 7 (100%)
Gold-Gold AlignE 365 (96.82%) 7750 (99.14%) 547 (97.50%) 890 (93.98%) 190 (98.96%) 7 (100%)
Gold-Gold SEA 367 (97.35%) 7623 (97.52%) 554 (98.75%) 930 (98.20%) 180 (93.75%) 7 (100%)
Gold-Gold AttrE-BootEA 375 (99.47%) 7810 (99.91%) 558 (99.47%) 945 (99.79%) 192 (100%) 7 (100%)
Gold-Gold AttrE-AlignE 374 (99.20%) 7796 (99.73%) 557 (99.29%) 945 (99.79%) 191 (99.48%) 7 (100%)
Gold-Gold AttrE-SEA 373 (98.94%) 7801 (99.80%) 557 (99.29%) 945 (99.79%) 192 (100%) 7 (100%)
Gold-LLM AttrE 308 (84.62%) 3673 (47.08%) 172 (31.73%) - - -
Gold-LLM BootEA 291 (79.95%) 4214 (54.01%) 160 (29.52%) - - -
Gold-LLM AlignE 296 (81.32%) 4449 (57.02%) 161 (29.70%) - - -
Gold-LLM SEA 327 (89.84%) 4920 (63.06%) 225 (41.51%) - - -
Gold-LLM AttrE-BootEA 308 (84.62%) 3673 (47.08%) 172 (31.73%) - - -
Gold-LLM AttrE-AlignE 308 (84.62%) 3673 (47.08%) 172 (31.73%) - - -
Gold-LLM AttrE-SEA 343 (94.23%) 5721 (73.33%) 253 (46.68%) - - -
Gold-Materials AttrE - 1269 (16.30%) - 794 (83.76%) - -
Gold-Materials BootEA - 1628 (20.91%) - 798 (84.18%) - -
Gold-Materials AlignE - 1598 (20.53%) - 797 (84.07%) - -
Gold-Materials SEA - 1584 (20.35%) - 794 (83.76%) - -
Gold-Materials AttrE-BootEA - 2276 (29.24%) - 802 (84.60%) - -
Gold-Materials AttrE-AlignE - 2276 (29.24%) - 802 (84.60%) - -
Gold-Materials AttrE-SEA - 2293 (29.45%) - 799 (84.28%) - -
LLM-LLM AttrE 360 (93.99%) 937 (98.42%) 72 (83.72%) - - -
LLM-LLM BootEA 358 (93.47%) 921 (96.74%) 74 (86.05%) - - -
LLM-LLM AlignE 351 (91.64%) 921 (96.74%) 77 (89.53%) - - -
LLM-LLM SEA 355 (92.69%) 924 (97.06%) 77 (89.53%) - - -
LLM-LLM AttrE-BootEA 365 (95.30%) 943 (99.05%) 78 (90.70%) - - -
LLM-LLM AttrE-AlignE 365 (95.30%) 943 (99.05%) 78 (90.70%) - - -
LLM-LLM AttrE-SEA 366 (95.56%) 945 (99.26%) 77 (89.53%) - - -
Materials-LLM AttrE - 6344 (81.65%) - - - -
Materials-LLM BootEA - 7226 (93.00%) - - - -
Materials-LLM AlignE - 7295 (93.89%) - - - -
Materials-LLM SEA - 7299 (93.94%) - - - -
Materials-LLM AttrE-BootEA - 7310 (94.08%) - - - -
Materials-LLM AttrE-AlignE - 7310 (94.08%) - - - -
Materials-LLM AttrE-SEA - 7311 (94.09%) - - - -
Materials-Materials AttrE - 1599 (20.55%) - 838 (87.93%) - -
Materials-Materials BootEA - 1728 (22.21%) - 814 (85.41%) - -
Materials-Materials AlignE - 1722 (22.13%) - 813 (85.31%) - -
Materials-Materials SEA - 1796 (23.08%) - 813 (85.31%) - -
Materials-Materials AttrE-BootEA - 2572 (33.06%) - 839 (88.04%) - -
Materials-Materials AttrE-AlignE - 2572 (33.06%) - 839 (88.04%) - -
Materials-Materials AttrE-SEA - 2633 (33.84%) - 839 (88.04%) - -