A Software System Evolution in Human-Centric Environment Driven by New user Intention Detection using CRF
The Software Service Evolution can easily determine through requests for changes, improvement, and
enablement of knowledge development continuously from users’, as compared to the other factors. It is unavoidable for
almost all software and can be seen as the development of system-user interactions. The ability to precisely and effectively
monitor users’ volatile requirements is perilous that requires to make a timely improved system for adaptation of fast varying
environments. In this research, a methodology applies Conditional Random Fields (CRF) as a mathematical foundation to
discover the users’ potential desires and requirements in order to deliver a quantitative exploration of system-user
interactions. By examining users’ run-time behavioral patterns, domain knowledge experts can predict how users’ intentions
shift. The results also show the effects of different regularization algorithms of CRF on the training model. Our supreme
objective is to accelerate software service evolution by using machine learning techniques. To detect users’ intentions using
the CRF method, an experiment on an open-source software is performed.
Keywords - Conditional Random Fields, intention, requirement, software service evolution, target.