Ontology-aided business rule extraction using machine learning
US9940581B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Nov 20, 2013 |
| Grant date | Apr 10, 2018 |
| Priority date | — |
| Expiry date | Oct 22, 2035 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06N5/025
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
An approach for distinguishing a business rule from a non-business rule in a computer program is provided. A rule in the program is identified based on a conditional statement within the rule. Whether the rule performs underlying operation of the program independent of a business function of the program is determined. The rule and metadata of the rule are searched for a key word which indicates part of a business transaction. Whether a sequence of program steps in the rule matches a predetermined sequence of program steps indicative of a business rule or whether the sequence of program steps in the rule matches a predetermined sequence of program steps indicative of underlying operation of the program independent of the business function of the program is determined. If the rule is a business rule is determined based on both determining steps and both searching steps.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.