Determining Software Inter-Dependency Patterns for Integration Testing by applying Machine learning on Logs and Telemetry data
R Rajaraman, P. K. Kapur, Deepak Kumar · 2020
Businesses running software applications are moving more towards vendor-agnostic approaches. Integration testing becoming more demanding and complex than ever. Determining integration testing requires a solid dependency pattern among different software applications. We would like to use here an NLP (Natural language processing) & Machine learning Classification model-based approach to identify the dependency graph between different component across applications. There are existing approaches eliciting about package dependencies and source code-based dependencies. In this paper, we would like to introduce an approach to understand the communication dependencies between products rather than just high-level package or install dependencies. The communication dependencies will be prioritized based on the frequency and criticality of usage. The prioritization helps us in determining the software inter-dependency patterns for Integration testing.