Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.575750
Title: Terminology recognition in the aerospace domain
Author: Butters, Jonathan David
Awarding Body: University of Sheffield
Current Institution: University of Sheffield
Date of Award: 2012
Availability of Full Text:
Full text unavailable from EThOS. Restricted access.
Please contact the current institution’s library for further details.
Abstract:
The objects and phenomena within the world around us are frequently shared. Common understandings, intuition, and common sense allow people to communicate ideas and instructions by identifying the semantics of the words used. When we communicate these ideas through text, we rely on strings of characters to carry semantics, and through these, because of common understandings, intuition, and common sense we are often able to gain back the references to the original ideas. As people's capacity to express their ideas through text is unbounded, the task of deciphering original meanings when synonyms and term variations are involved can be challenging, trying to do this task automatically is more so. The body of work within this thesis aims to identify and then overcome the challenges in automatically recognising the terminology used for concepts within the aerospace domain. The main achievements of this thesis include new knowledge about the use of terminology in this safety critical and highly technical domain, knowledge that approaches which are often impractical in open and alternate domains may be applicable due to the aerospace domain's unique nature, and an approach to automatically recognising terminology for the purposes of knowledge management. The approach involves constructing an explicit model of the domain (afforded by its closed nature), and then leveraging it along with exploiting the combinatorial nature of aerospace domain terms in an extraction methodology. Extraction is performed using a novel, hybrid dictionary and machine learning based technique which is shown to perform better than existing dictionary, statistics, and machine learning techniques. This thesis is organised into four parts, the first part provides a survey of the aerospace domain and state of the art methods by which terminology is recognised. The second part documents several investigations into the real-world use of terminology within and across aerospace-based communities in order to provide an understanding on how terminology should be recognised and generate a set of requirements. The third part details the proposed approach and presents a novel, patent-pending methodology to fulfil the requirements. The fourth part concludes this thesis with a summary and by v answering the original research questions, before discussing possible lines of future work.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.575750  DOI: Not available
Share: