Automatic extraction of events from Textual Requirements specification

Events give important information about the behavior of a system in a summarized form. In the past, events have played an important role in breaking the functional requirements of the system in the ¿event partitioning approach¿. Our previous work has shown that events can be a starting point in object-oriented analysis of requirements. Every event triggers a use case in the system, hence should get a priority in identifying and analyzing requirements over use cases. In any system there is plethora of events happening, some are important to be recorded, while others are to be ignored. Moreover, there are various perspectives to define events. Thus, it becomes important to have an automated process that could help not only in extracting events but also analyze and classify them into various types. A study on various existing event extraction tools shows that they are either domain specific or take events as actions that occur at a particular time. There is no tool which extracts events that represent system behavior and at the same time gives a result that can be reused for application in multiple domains. This paper presents, a domain independent tool, developed in Java that automates the process of extraction, analysis and classification of events from textual requirements expressed in English as a natural language. This tool also assists the analysts in further refining identified events and to add some new events in the application domain. Tool has been tested on several case studies from different domains and has given very promising results.