Knowledge Representation Framework for Software Requirement Specification

Abstract

The need to extract correct information has become one of the main issues when analyzing the software requirement specification (SRS) documentation. The amount of gathered knowledge depends on the size of the information. However, the complexity of software systems is continuously increasing. As software systems change to more complicated systems, the information from the SRS documents may not be easily comprehended. For example, each annotation requirements tasks target the different types of information, and these tasks require the availability of experts specialized in the field. Large scale annotation tasks require multiple experts and very costly. If the number of experts is limited, annotation tasks may overwhelm the experts. The organization would not complete their objectives if they failed to manage their data because poor knowledge management affects many operations within the organization. To extract such vast information and turn it to useful knowledge, a company needs top quality software. This technology should able to input, store, and access systematically. This paper will discuss a framework based on the knowledge-based method, an attempt to improve knowledge representation. In this approach, WordNet 2.1 would be used as the knowledge source used to identify concepts represented by each word in a text from the SRS document.