# Hume Concepts
Hume is a Graph-Powered Insights Engine. It is composed of 3 main components that allow any enterprise to make better sense of connected data.
# Knowledge Graphs
The Schema is the central component of a Knowledge Graph. It describes the business domain together with its challenges and solutions in the form of a graph.
It consists of
relationships, when we refer to these terms we're speaking about the Schema.
The Schema is not directly bound to a database, it rather expresses the classes and their relationships in terms used by business.
Perspectives enable a view of the schema against a particular database and they are the elements in a Knowledge Graph providing a 360 degrees view of the data, Perspectives are bound to a physical Neo4j database.
Users in a Knowledge Graph are granted the usage of a particular Perspective for visualising the data.
Perspectives can have different configurations (search relevance, visibility of classes and relationships, ...) to ensure the best experience for different groups of users.
Visualisations allow end users to search, explore and act on the graph data in the context of a particular Perspective. Users can have more than one visualisation and can style it as their wish.
nodes that are of a certain class and
links that are of a certain relationship type - both defined within a schema.
The data in the perspective is persisted in the Hume application database so users can always retrieve their last visualisation state.
A Knowledge Graph is usually a part of a bigger system in an enterprise architecture and Orchestra provides the capability to integrate various sources of data and transform them into a graph representation in accordance with the Schema.
Transforming data from one silo into a new graph structure might not be sufficient to cover all classes and relationships within the Schema (as described above, the Schema represents the business domain, its challenges and solutions). Orchestra enables to use capabilities from the Hume ecosystem and third-party integrations in order to extract more information and enrich the original data. The result of this enrichement is again following the Schema and stored in the database as nodes and links. Click here to find what is Orchestra in greater detail.
# Hume Ecosystem
Ecosystem platform consists of complimentary services i.e.
Skills (for example Natural Language Processing for text analytics) and the definitions of the connections including Postgres databases, Kafka Brokers, local file system,...
The elements of the Ecosystem are granted to groups of Hume users - so called
Out of the box, Hume can provide many complimentary services of high value:
- Named Entity Recognition
- Relationship Entity Extraction
- Document Clustering
- PDF Parsing
Naturally, Hume very easily connects to existing work and capabilities available within enterprises over http protocol (API).
# Hume Labs
While Hume offers many capabilities for text understanding and information extraction, it is not feasible for Hume to cover every domain and its specific language out of the box.
Hume Labs enables users to train Hume to understand domain specific language and map it to classes and relationships in accordance with the Schema.
Let's say you would like to monitor investments within the Electric Vehicles market, users are able to feed in examples of texts and label specific words that represent entities of interest. Hume leverages this information so incoming text is automatically analysed, information extracted and mapped to nodes and relationships playing their unique role in the graph.