AhmedGaver commited on
Commit
a668c65
·
verified ·
1 Parent(s): 83c6c22

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +101 -0
README.md ADDED
@@ -0,0 +1,101 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # LexisNexis KYC Screening Dataset
2
+
3
+ ## Dataset Description
4
+
5
+ This dataset contains parsed LexisNexis Bridger Insight screening records for Know Your Customer (KYC) name screening and entity resolution. It includes PEP (Politically Exposed Persons) and sanctions list matching data.
6
+
7
+ ### Dataset Structure
8
+
9
+ ```json
10
+ {
11
+ "screening_metadata": {
12
+ "record_id": "string",
13
+ "alert_state": "string",
14
+ "status": "string",
15
+ "is_false_positive": "boolean",
16
+ "date_created": "string",
17
+ "predefined_search_name": "string"
18
+ },
19
+ "input_profile": {
20
+ "name": {
21
+ "full": "string",
22
+ "first": "string",
23
+ "middle": "string",
24
+ "last": "string"
25
+ },
26
+ "gender": "string",
27
+ "addresses": [
28
+ {
29
+ "city": "string",
30
+ "country": "string",
31
+ "state": "string"
32
+ }
33
+ ]
34
+ },
35
+ "hit_profiles": [
36
+ {
37
+ "names": [
38
+ {
39
+ "full": "string",
40
+ "first": "string",
41
+ "last": "string",
42
+ "score": "string",
43
+ "is_best_match": "boolean"
44
+ }
45
+ ],
46
+ "genders": ["string"],
47
+ "nationalities": [
48
+ {
49
+ "country": "string",
50
+ "category": "string"
51
+ }
52
+ ],
53
+ "addresses": [
54
+ {
55
+ "city": "string",
56
+ "country": "string"
57
+ }
58
+ ]
59
+ }
60
+ ]
61
+ }
62
+ ```
63
+
64
+ ### Use Cases
65
+
66
+ - **Entity Resolution**: Match customer profiles against PEP/sanctions lists
67
+ - **Name Screening**: Cross-reference names with watchlists
68
+ - **RAG Pipelines**: Build retrieval systems for compliance screening
69
+ - **SPARQL Integration**: Generate queries for Wikidata name disambiguation
70
+
71
+ ### Fields
72
+
73
+ #### Input Profile (Customer/Subject)
74
+ - Full name, parsed name components
75
+ - Gender
76
+ - Current address and country
77
+
78
+ #### Hit Profiles (Watchlist Matches)
79
+ - Multiple name aliases and variations
80
+ - Match scores (0-100)
81
+ - Gender information
82
+ - Nationalities and citizenships
83
+ - Associated addresses
84
+
85
+ ### Statistics
86
+
87
+ - Total Records: 150
88
+ - Records with Hits: 150
89
+ - False Positives: 42
90
+
91
+ ## Intended Use
92
+
93
+ This dataset is for:
94
+ - Compliance and KYC research
95
+ - Name entity resolution algorithm development
96
+ - Training models for entity matching
97
+ - Testing SPARQL query generation for name disambiguation
98
+
99
+ ## License
100
+
101
+ Proprietary - LexisNexis Bridger Insight data