Method and system for continuous clustering of a continuous stream of software development lifecycle (SDLC) artifacts
US12399914B1 · kind B1 · utility
Assignee
Inventors
Key dates
| Filing date | May 9, 2024 |
| Grant date | Aug 26, 2025 |
| Priority date | — |
| Expiry date | May 9, 2044 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06F16/287
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
Disclosed is a method and system for clustering a continuous stream of software development lifecycle (SDLC) artifacts. The method and system receives and preprocesses SDLC artifacts. Subsequently, a first set of topics is generated based on application of topic modelling and each SDLC artifact is clustered into a topic of the first set of topics. In the absence of topic count, a coherence process analyzes relationships and patterns within the data to determine the topic count. In case of a skewed distribution of records in the topics, skewed keywords are eliminated before rerunning the coherence process. Upon detecting an updated stream of SDLC artifacts, a second set of topics is generated. Thereafter, a similarity between the first set of topics and the second set of topics is evaluated, to perform re-clustering of the SDLC artifacts. The method and system also enables SMEs to verify and provide feedback for dynamic clustering.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.