Automatic Identification of Texts Written by Authors with Alzheimer's Disease
- Juan Soler-Company, TALN, DTIC, Universitat Pompeu Fabra, Barcelona, Barcelona, Spain
- Leo Wanner, ICREA and Pompeu Fabra University, Barcelona, Spain
AbstractAlzheimer's disease leads to a degradation of vocabulary and communication skills. This makes us assume that writings of individual authors can be classified automatically into "pre-Alzheimer's period'' and "Alzheimer's period''. But the writing style of an author is highly individual. Can we still detect whether any given novel is written by an author who suffers from Alzheimer's? To assess this, we use a corpus of novels by three well-known writers who were diagnosed with Alzheimer's: Iris Murdoch, Terry Pratchett and Agatha Christie. Using a mostly stylistic set of features we are able to distinguish between novels written under the influence of the disease and novels written by healthy writers with more than 82% accuracy. The classification of the novels of a given author into "pre-Alzheimer's period'' and "Alzheimer's period'' is accomplished with more than 86% accuracy.
Return to previous page