wikipunk commited on
Commit
317b37f
1 Parent(s): 4117cf0

update readme

Browse files
Files changed (2) hide show
  1. README.md +178 -0
  2. requirements.txt +1 -0
README.md CHANGED
@@ -1,3 +1,181 @@
1
  ---
 
 
2
  license: mit
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ language:
3
+ - en
4
  license: mit
5
+ tags:
6
+ - knowledge-graph
7
+ - rdf
8
+ - owl
9
+ - ontology
10
+ annotations_creators:
11
+ - expert-generated
12
+ pretty_name: FIBO
13
+ size_categories:
14
+ - 100K<n<1M
15
+ task_categories:
16
+ - graph-ml
17
+ dataset_info:
18
+ features:
19
+ - name: subject
20
+ dtype: string
21
+ - name: predicate
22
+ dtype: string
23
+ - name: object
24
+ dtype: string
25
+ config_name: default
26
+ splits:
27
+ - name: train
28
+ num_bytes: 56045523
29
+ num_examples: 236579
30
+ dataset_size: 56045523
31
+ viewer: false
32
  ---
33
+
34
+ # FIBO: The Financial Industry Business Ontology
35
+
36
+ ### Overview
37
+ In the world of financial technology, the vastness of data and the
38
+ complexity of financial instruments present both challenges and
39
+ opportunities. The Financial Industry Business Ontology (FIBO) offers
40
+ a structured framework that bridges the gap between theoretical
41
+ financial concepts and real-world data. I believe machine learning
42
+ researchers interested in the financial sector could use the
43
+ relationships in FIBO to innovate in financial feature engineering to
44
+ fine-tune existing models or build new ones.
45
+
46
+ ### Use-cases
47
+ - Comprehensive Data Structure: FIBO encompasses a wide range of
48
+ financial concepts, from derivatives to securities. Its design ensures
49
+ an in-depth understanding of financial instruments from experts
50
+ in knowledge representation and the financial industry.
51
+ - Decoding Complex Relationships: The financial domain is
52
+ characterized by its intricate interdependencies. FIBO's structured
53
+ approach provides clarity on these relationships, enabling machine
54
+ learning algorithms to identify patterns and correlations within
55
+ large datasets.
56
+ - Linkage with Real-world Data: A distinguishing feature of FIBO is
57
+ its capability to associate financial concepts with real-world
58
+ financial data and controlled vocabularies. This connection is
59
+ crucial for researchers aiming to apply theoretical insights in
60
+ practical contexts in financial enterprises with their existing
61
+ data.
62
+ - Retrieval Augmented Generation: The emergence of Large Language
63
+ Models, especially when using Retrieval Augmented Generation (RAG),
64
+ has the potential to transform financial data processing and
65
+ interpretation.
66
+ - Document Classification: With the surge in financial documents,
67
+ utilizing RAG to classify financial datasets based on FIBO concepts
68
+ may help financial analysts get better accuracy and depth in data
69
+ interpretation with smart prompting.
70
+
71
+ #### Building and Verification:
72
+ 1. **Construction**: The ontology was imported using the
73
+ [AboutFIBOProd-IncludingReferenceData](https://github.com/edmcouncil/fibo/blob/master/AboutFIBOProd-IncludingReferenceData.rdf)
74
+ into Protege version 5.6.1.
75
+ 2. **Reasoning**: Due to the large size of the ontology I used the ELK
76
+ reasoner plugin to materialize (make explicit) inferences in the
77
+ ontology.
78
+ 3. **Coherence Check**: The Debug Ontology plugin in Protege was used
79
+ to ensure the ontology's coherence and consistency.
80
+ 4. **Export**: After verification, inferred axioms, along with
81
+ asserted axioms and annotations, were exported using Protege.
82
+ 5. **Encoding and Compression**: [Apache Jena's
83
+ riot](https://jena.apache.org/documentation/tools/) was used to convert the
84
+ result to ntriples, which was then compressed with gzip.
85
+
86
+ ## Features
87
+ The FIBO dataset is composed of triples representing the relationships
88
+ between different financial concepts and named individuals such as
89
+ market participants, corporations, and contractual agents.
90
+
91
+ ### Usage
92
+ First make sure you have the requirements installed:
93
+
94
+ ```python
95
+ pip install datasets
96
+ pip install rdflib
97
+ ```
98
+
99
+ You can load the dataset using the Hugging Face Datasets library with the following Python code:
100
+
101
+ ```python
102
+ from datasets import load_dataset
103
+ dataset = load_dataset('wikipunk/fibo2023Q3', split='train')
104
+ ```
105
+
106
+ #### Note on Format:
107
+ The subject, predicate, and object features are stored in N3 notation
108
+ with no prefix mappings. This allows users to parse each component
109
+ using `rdflib.util.from_n3` from the RDFLib Python library.
110
+
111
+ ### Example
112
+ Here is an example of a triple in the dataset:
113
+ - Subject: `"<https://spec.edmcouncil.org/fibo/ontology/FBC/FunctionalEntities/MarketsIndividuals/ServiceProvider-L-JEUVK5RWVJEN8W0C9M24>"`
114
+ - Predicate: `"<http://www.w3.org/1999/02/22-rdf-syntax-ns#type>`
115
+ - Object: `"<https://spec.edmcouncil.org/fibo/ontology/BE/FunctionalEntities/FunctionalEntities/FunctionalEntity>"`
116
+
117
+ This triple represents the statement that the market individual
118
+ "ServiceProvider-L-JEUVK5RWVJEN8W0C9M24" has a type of "FunctionalEntity".
119
+
120
+ ---
121
+
122
+ ## Ideas for Deriving Graph Neural Network Features from FIBO:
123
+ Graph Neural Networks (GNNs) have emerged as a powerful tool for
124
+ machine learning on structured data. FIBO, with its structured
125
+ ontology, can be leveraged to derive features for GNNs.
126
+
127
+ ### Node Features:
128
+
129
+ - **rdf:type**: Each entity in FIBO has one or more associated `rdf:type`,
130
+ `<http://www.w3.org/1999/02/22-rdf-syntax-ns#type>`, that
131
+ indicates its class or category. This can serve as a primary node
132
+ feature to encode.
133
+
134
+ - **Entity Attributes**: Attributes of each entity, such as names or
135
+ descriptions, can be used as additional node features. Consider
136
+ embedding descriptions using a semantic text embedding model.
137
+
138
+ ### Edge Features:
139
+
140
+ - **RDF Predicates**: The relationships between entities in FIBO are
141
+ represented using RDF predicates. These predicates can serve as edge
142
+ features in a GNN, capturing the nature of the relationship between
143
+ nodes.
144
+
145
+ ### Potential Applications:
146
+
147
+ 1. **Entity Classification**: Using the derived node and edge
148
+ features, GNNs can classify entities into various financial
149
+ categories, enhancing the granularity of financial data analysis.
150
+
151
+ 2. **Relationship Prediction**: GNNs can predict potential
152
+ relationships between entities, aiding in the discovery of hidden
153
+ patterns or correlations within the financial data.
154
+
155
+ 3. **Anomaly Detection**: By training GNNs on the structured data from
156
+ FIBO and interlinked financial datasets, anomalies or
157
+ irregularities in them may be detected, ensuring data integrity and
158
+ accuracy.
159
+
160
+ ### Acknowledgements
161
+ We extend our sincere gratitude to the FIBO contributors for their
162
+ meticulous efforts in knowledge representation. Their expertise and
163
+ dedication have been instrumental in shaping a comprehensive and
164
+ insightful framework that serves as a cornerstone for innovation in
165
+ the financial industry.
166
+
167
+ If you are interested in modeling the financial industry you should
168
+ consider [contributing to
169
+ FIBO](https://github.com/edmcouncil/fibo/blob/master/CONTRIBUTING.md).
170
+
171
+ ### Citation
172
+ ```bibtex
173
+ @misc{fiboQ32023,
174
+ title={Financial Industry Business Ontology (FIBO) Q32023 Release},
175
+ author={EDM Council and Various Contributors},
176
+ year={2023},
177
+ note={Derived from the AboutFIBOProd-IncludingReferenceData.rdf},
178
+ howpublished={\url{https://spec.edmcouncil.org/fibo/}},
179
+ license={MIT License, \url{https://opensource.org/licenses/MIT}}
180
+ }
181
+ ```
requirements.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ rdflib>=6.0.0