The writing style of a number of authors writing in
English was empirically investigated to detect
stylistic patterns in relation to advancing age. The
aim was to
identify the type of stylistic markers among lexical,
syntactical, phonemic, entropic,
character-based, and content ones most able to
discriminate between early,
middle, and late works of the selected authors, and
the best classification or prediction
algorithm most suited for this task. A pilot
examination of personal letters and poetry by
Christina Georgina Rossetti and Edgar Allan Poe
suggested that authors and genres vary
inconsistently. Examination of selected variables on
Shakespeare s plays which emphasised their predictive
power, led to four experiments on personal
correspondence and poetry from Edna St Vincent Millay and
William Butler Yeats. The methods used were stepwise
multiple linear regression, regression trees, ordinal
logistic regression and artificial
neural networks. Prediction was found influenced by
method, authorship, and genre whereas classification
was found sensitive to all factors. Given the current
experiments, generalizable conclusions for the wider
author population have been avoided.
English was empirically investigated to detect
stylistic patterns in relation to advancing age. The
aim was to
identify the type of stylistic markers among lexical,
syntactical, phonemic, entropic,
character-based, and content ones most able to
discriminate between early,
middle, and late works of the selected authors, and
the best classification or prediction
algorithm most suited for this task. A pilot
examination of personal letters and poetry by
Christina Georgina Rossetti and Edgar Allan Poe
suggested that authors and genres vary
inconsistently. Examination of selected variables on
Shakespeare s plays which emphasised their predictive
power, led to four experiments on personal
correspondence and poetry from Edna St Vincent Millay and
William Butler Yeats. The methods used were stepwise
multiple linear regression, regression trees, ordinal
logistic regression and artificial
neural networks. Prediction was found influenced by
method, authorship, and genre whereas classification
was found sensitive to all factors. Given the current
experiments, generalizable conclusions for the wider
author population have been avoided.