Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.747762
Title: Characterisation of business documents : an approach to the automation of quality assessment
Author: Thurlow, Ian
ISNI:       0000 0004 7232 4952
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Abstract:
This thesis explores a new approach to automatic characterisation of business documents of different levels of document effectiveness. Supervised text categorisation techniques are used to derive text features that characterise a specific type of business document in accordance with pre-assigned levels of document utility. The documents in question are the executive summary sections of a representative sample of sales proposal documents. The executive summaries are first rated by domain experts against a quality framework comprising pre-selected dimensions of document quality. An automatic analysis of the texts shows that certain words, word sequences, and patterns of words have the capacity to discriminate between executive summaries of varying levels of document effectiveness. Function words, which are frequently ignored in many text classification tasks, are retained and are shown to provide an important element of the word patterns. Automatic text classifiers that utilise these features are shown to categorise previously unseen executive summaries at an acceptable level of classification performance. The outcomes of the research are applied to the development of a new computer application. The application identifies, in the text of a new executive summary, word patterns that discriminate between sets of summaries previously categorised into different levels of document utility. The action of highlighting the respective categories of discriminating word patterns directs authors to areas of text that may need further attention. A trial of a prototype of the application suggests that it provides an effective way to help sales professionals improve the content and quality of the text of this type of business document. Moreover, as the approach is suitably generic, it could be applied to different types of document in different domains.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.747762  DOI: Not available
Share: