Sami, Ishrat. 2024. Unsupervised Contextual Visualization of Terms for Understanding Story Planning in an NLP Corpus. Doctoral thesis, Goldsmiths, University of London [Thesis]

No full text available

Text (Unsupervised Contextual Visualization of Terms for Understanding Story Planning in an NLP Corpus) COM_thesis_SamiIR_2024.pdf - Accepted Version Permissions: Administrator Access Only until 31 July 2027. Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (23MB)

Abstract or Description

Storytelling in news, articles and books is the science of transforming knowledge into wisdom and the art of convincing readers about an author’s opinions through rhetoric. A neat and well-crafted document effectively communicates the author’s rhetoric and helps readers acquire comprehensive knowledge about the discussed topic. Understanding the variation in rhetoric about a topic of interest can be achieved by identifying story plans preserved in the chronological reporting style of existing documents. In contrast, preparing a story plan during the pre-writing phase develops rhetoric that leads to improved audience engagement through better content organization. It is challenging to design systems which can cognitively aid authors in producing creative and elegant influential content. Therefore, this PhD study aimed to identify various story planning patterns in existing documents automatically using contextual topological terms analysis. This study identified topological patterns hidden in the document, corpus and topic-based relative context. The identified single document pattern was used to extract story plans from documents. The peripheral topological nature of terms concerning time in a document collection is intuitive for identifying meaningful ideas during the prewriting phase. Motivated by these insights and the storyboard concept, this study presents a contextual Terms Board visualization for topic based document collection. Terms Board can be interpreted as a palette of ideas that helps individuals craft a storyline. The proposed approaches are evaluated via cognitive reading and writing user experiments using a news archive in this study. Finally, this research presented a demonstration of a story analysis which can guide researchers and reporters to develop their unique rhetoric through reading and planning their storyline in the prewriting phase for elegant writing.

Item Type:	Thesis (Doctoral)
Identification Number (DOI):	https://doi.org/10.25602/GOLD.00037478
Keywords:	NLP, Contextual Visualization, Story Extraction, Story Planning, Writing Patterns, Cognitive User Experiment, Data Visualization
Departments, Centres and Research Units:	Computing
Date:	31 July 2024
Item ID:	37478
Date Deposited:	16 Aug 2024 10:27
Last Modified:	16 Aug 2024 10:33
URI:	https://research.gold.ac.uk/id/eprint/37478

View statistics for this item...

Edit Record (login required)