Citation
Downloads
Abstract
Dramatic characters frequently fill out different role types and act ac- cording to traits conventionally attributed to their role. One of these role types is the “schemer,” characterized by intervening in a play’s main plot and driving forward the plot’s main conflicts. In our study, we utilized secondary literature to identify 50 characters as schemers and extracted a wide range of features which are likely to distinguish “schemers” from “non-schemers.” Using machine learning, we trained a model to automatically classify characters according to these two classes and performed a number of analyses in order to identify the most contributing features. Our model is able to reliably detect schemers, utilizing features that cover information about stage presence and content of character speech, but exhibits a rather low precision. We show that this can partially be attributed to the heterogeneous nature that characterizes the group of schemers.
BibTeX
@incollection{krautter2024a,
address = {Berlin},
author = {Benjamin Krautter and Janis Pagel},
booktitle = {{Computational Drama Analysis. Reflecting on Methods and Interpretations}},
doi = {10.1515/9783111071824-007},
editor = {Melanie Andresen and Nils Reiter},
month = {6},
pages = {123--148},
publisher = {De Gruyter},
title = {{The Schemer in German Drama. Identification and Quantitative Characterization}},
year = {2024},
}
RIS
TY - CHAP TI - The Schemer in German Drama. Identification and Quantitative Characterization AU - Benjamin Krautter AU - Janis Pagel ED - Melanie Andresen ED - Nils Reiter PY - 2024 CY - Berlin J2 - Computational Drama Analysis. Reflecting on Methods and Interpretations DO - 10.1515/9783111071824-007 PB - De Gruyter SP - 123 EP - 148 ER -