GraphDB Reasoning Engine
GraphDB Reasoning Engine
Powerful knowledge expression and reasoning and graph data storing in GraphDB Suite enables the discovery and storing of new facts from converted data through reasoning. In particular, schema axiom and instance axiom are provided for high-level and low-level reasoning and relationship reasoning, and concepts in the data model.
The GraphDB Suite supports RDFS, OWL1 and OWL2 that are W3C’s graph models, and the (Labeled) Property Graph model can be saved and queried. The user can use two data models selectively in accordance with the purpose.
The GraphDB reasoning engine provides two reasoning strategies (forward-chaining and backward-chaining) for rule-based reasoning. The GraphDB reasoning engine provides the forward-chaining-based reasoning strategy, with the backward-chaining reasoning engine used if and as necessary. In regard to advantages and disadvantages of forward-chaining-based reasoning, inferred fact is expanded after data transactions in the repository, so it may be slow when uploading, storing, adding or deleting new facts. In the GraphDB reasoning engine, reasoning begins when data is inputted. However, the forward-chaining-based reasoning method creates and stores reasoning results for all data in advance, so it provides a reasonably fast query and search performance. Generally, the backward-chaining reasoning method creates reasoning when querying or searching, and complicated deductive reasoning, conformance testing or another reasoning is likely to occur, which may affect performance.
Graph data storing and analysis provide various analysis functions for data integration and graph data through the support of powerful knowledge expression and multiple data models. The main features of graph data storing and analysis are as follows.
Main Functions and Specifications
- Reasoning function based on various rules
The graph data reasoning engine provides optimized rule sets according to the data model. This engine provides RDF, OWL-Horst, OWL2-QL and OWL2-RL and the user can use these functions by optimizing the rule.
- Schema update transaction function
GraphDB provides a function to modify schema if it is necessary to modify incorrectly entered schema. This function performs and reflects reasoning for existing data according to schema changes promptly, so it is carried out in the engine internally.
- Consistency verification function
The graph data reasoning engine is the function that checks whether or not graph data is configured according to schema axiom.
- Re-reasoning function
This function deletes previously reasoned data and re-creates reasoned data in order to secure consistency of graph data or delete incorrectly reasoned data.