Visualizing Paragraph Closeness for Academic Writing Support

Author

O´Rourke, Stephen T. ; Calvo, Rafael A.

Author_Institution

Sch. of Electr. & Inf. Eng., Univ. of Sydney, Sydney, NSW, Australia

fYear

2009

fDate

15-17 July 2009

Firstpage

688

Lastpage

692

Abstract

In this paper, we describe a novel visualization to support formative assessment in academic writing. The visualization makes use of text mining techniques to provide insight on the flow of topics in an essay. We propose that visualization can be used to mitigate many of problems associated with the subjectivity of essay assessment by bringing greater insight to an essaypsilas latent features. The proposed visualization method involves a process of non-negative matrix factorization (NMF), to uncover topics in an essay, follow by multidimensional scaling (MDS), to map the topic closeness of the essaypsilas paragraphs. We evaluate the visualization method with a corpus of 44 short essays written by university students.

Keywords

computer aided instruction; data mining; data visualisation; text analysis; academic writing; multidimensional scaling; nonnegative matrix factorization; paragraph closeness; text mining technique; Algorithm design and analysis; Collaboration; Feedback; Humans; Multidimensional systems; Reflection; Software tools; Text mining; Visualization; Writing;

fLanguage

English

Publisher

ieee

Conference_Titel

Advanced Learning Technologies, 2009. ICALT 2009. Ninth IEEE International Conference on

Conference_Location

Riga

Print_ISBN

978-0-7695-3711-5

Electronic_ISBN

978-0-7695-3711-5

Type

conf

DOI

10.1109/ICALT.2009.145

Filename

5194339