Supervised Learning and the Indeterminacy of Disordered Genres

The poster puts forward the idea of interpreting accuracy scores of classification tasks as a measure of the psychological degree of looseness of historical genre concepts. Based on a more comprehensive corpus, this poster shall introduce an unconventional way of applying supervised machine learning to describe the peculiar order of disordered genres in three short steps: Firstly, the notion of ›disordered genres‹ has to be explicated. Secondly, results of classification tasks are presented, and, thirdly, the idea of interpreting strong as well as weak validation scores as a measure of the degree of the psychological manageability of genre concepts readers generate from reading is put forward. If historical reader responses correspond to the statistical results, validation scores of supervised learning tasks can be interpreted as a metric that measures the degree of looseness of historical genre concepts in general.