Function point counting automation based on Swagger specifications
Function Points, Machine Learning, Swagger, Automation
This work presents an approach for automated measurement of function points using Swagger specifications as input for identification of base functional components and machine learning algorithms as a resource for classifying features identified based on function point metrics.The proposal explores the technology independence and standardization of Swagger specifications with the resources provided by machine learning classification algorithms to enable an identification and classification of the analyzed functionalities. The applicability of machine learning classification algorithms was evaluated, where initial results were obtained with good accuracy, despite the initial experiments using a certain software architecture. The next steps of the work will use Swagger specifications, to identify the information necessary for the classification of features independently of technology and application of the algorithms already used to classify the features.