Genre analysis of structured e-mails for corpus profiling.
Clark, Malcolm John
MetadataShow full item record
CLARK, M., RUTHVEN, I. and HOLT, P., 2008. Genre analysis of structured e-mails for corpus profiling. In: A. DE ROECK, D. SONG and U. KRUSCHWITZ, eds. Proceedings of Workshop on Corpus Profiling for Information Retrieval and Natural Language Profiling. 18 October 2008. [online]. London: BCS-IRSG. Available from: http://www.bcs.org/upload/pdf/ewic_ir08_s1paper1.pdf [Accessed 18 March 2015]
This paper reports on our approach to the analysis of genre recognition using eyetracking. We focused on a collection of different types of email which could represent different datasets, such as, mailing lists for calls for papers, newsletters, etc. We found that genre analysis based on purpose, form and layout features is potentially effective for identifying the characteristics of these datasets and we have highlighted some of the new important features of genres. The results from a pilot study showed a clear effect, with an interaction between the email texts and the visual cues or features perceived and also the strategies employed for the processing of the texts. We found, in our small sample, that readers can determine the purpose and form of genres and that during this process some readers do skim the shape of the e-mails (form).