Method of Activity of Intelligent Agent for Semantic Analysis of Software Requirements

The paper deals with the development of the method of activity and the realization of the intelligent agent (IA) for semantic analysis of the natural language software requirements. This IA performs the analysis of the requirements, determines the number and percentage of missing indicators for determining the metrics of complexity and quality of software, displays which indicators are missing for each metric, and also forms the real ontology of the domain “Software Engineering” (part “Quality and complexity of software. Metric analysis”). During an experiment, which is conducted with the help of the realized IA, the semantic analysis of requirements for the Internet of Things (IoT)-system was performed. Such analysis provides an increase of the volume and ensures of the sufficiency of metric information in the software requirements, which, in turn, provides an improvement in the quality of the software of IoT-system.