Preliminary concepts for requirements mining and classification using hidden Markov model

Tóth, László: Preliminary concepts for requirements mining and classification using hidden Markov model. In: Conference of PhD Students in Computer Science, (11). pp. 110-113. (2018)

Cikk, tanulmány, mű

Download (201kB) | Preview


Requirements specifications are crucial documents of the software development. These documents are created based on the expectations of stakeholders and the requirements of various regulations. The expectations of stakeholders might contain some ambiguities, inconsistencies and also some contradictions especially comparing with regulations. Creating specifications which do not contain issues mentioned above is one of the most important tasks of the business analysts. Reconciling the expectations and requirements can be a demanding task particularly in case of a complex system. Several attempts have been made to support the duties of business analysts using computer-aided natural language processing methods for requirements engineering. One of the most important steps during the elicitation process is the classification of the requirements collected from stakeholders considering that these requirements will be part of the formal and specific models. Investigations devoted to this task have achieved remarkable results using supervised and semi-supervised methods. However, these models need somehow prepared requirements in order to use them as their inputs. The approach presented in this article focuses on extracting and classifying requirements from unstructured documents. Hidden Markov Models are utilized in various field of natural language processing and their usability already has been proven. The idea of using HMM for processing requirements is stemmed from the success of using it for those tasks where extracting information often hidden in unstructured texts is a crucial part of the particular task. Using HMMs, which is a novel approach to processing texts containing requirements, can help also utilize various linguistic features of the sentences that could be obtained with difficulty by classical processes.

Item Type: Article
Journal or Publication Title: Conference of PhD Students in Computer Science
Date: 2018
Volume: 11
Page Range: pp. 110-113
Event Title: Conference of PhD students in computer science (11.) (2018) (Szeged)
Uncontrolled Keywords: Markov-folyamat, Markov-modell
Additional Information: Bibliogr.: p. 112-113. ; összefoglalás angol nyelven
Date Deposited: 2019. Nov. 04. 12:41
Last Modified: 2019. Nov. 04. 12:41

Actions (login required)

View Item View Item