Dokument: Machine-assisted Text Classification of Public Participation Contributions
| Titel: | Machine-assisted Text Classification of Public Participation Contributions | |||||||
| URL für Lesezeichen: | https://docserv.uni-duesseldorf.de/servlets/DocumentServlet?id=64622 | |||||||
| URN (NBN): | urn:nbn:de:hbz:061-20240201-105333-5 | |||||||
| Kollektion: | Dissertationen | |||||||
| Sprache: | Englisch | |||||||
| Dokumententyp: | Wissenschaftliche Abschlussarbeiten » Dissertation | |||||||
| Medientyp: | Text | |||||||
| Autor: | Romberg, Julia [Autor] | |||||||
| Dateien: |
| |||||||
| Beitragende: | Prof. Dr. Conrad, Stefan [Gutachter] Jun. Prof. Dr. Escher, Tobias [Gutachter] | |||||||
| Dewey Dezimal-Klassifikation: | 000 Informatik, Informationswissenschaft, allgemeine Werke » 004 Datenverarbeitung; Informatik | |||||||
| Beschreibung: | Engaging citizens in decision-making processes is a widely implemented instrument in democracies. Such public participation processes serve the goal of achieving a more informed procedure to potentially improve the process outcome and increase the public acceptance of decisions made. As public officials try to evaluate the often large quantities of citizen input by hand, they regularly face challenges due to restricted resources. For textual contributions, the most common form of citizen input, natural language processing offers the prospect of automatic support for evaluation. Still, many methods are inadequate due to insufficient accuracy, lack of robustness across datasets, or a neglect of important aspects of practical application. This thesis explores how existing research gaps can be overcome with text classification methods, focusing on the tasks of thematic structuring and argument analysis in the manual evaluation cycle.
We start with a systematic literature review of previous approaches to the machine-assisted evaluation of textual contributions. Given the identified shortage of language resources, we subsequently create a multidimensionally annotated corpus to facilitate the development of text classification models for German-language public participation. Once the groundwork is laid, our initial focus is on the thematic structuring of public input, particularly considering the uniqueness of many public participation processes in terms of content and context. To make customized models for automation worthwhile, we leverage the concept of active learning to reduce manual workload by optimizing training data selection. In a comparison across three participation processes, we show that transformer-based active learning can significantly reduce manual classification efforts for process sizes starting at a few hundred contributions while maintaining high accuracy and affordable runtimes. We then turn to the criteria of practical applicability that conventional evaluation does not encompass. By proposing measures that reflect class-related demands users place on data acquisition, we provide insights into the behavior of different active learning strategies on class-imbalanced datasets, which is a common characteristic in collections of public input. Afterward, we shift the focus to the analysis of citizens’ reasoning. Our first contribution lies in the development of a robust model for the detection of argumentative structures across different processes of public participation. Our approach improves upon previous techniques in the application domain for the recognition of argumentative sentences and, in particular, their classification as argument components. Following that, we explore the machine prediction of argument concreteness. In this context, we account for the subjective nature of argumentation by presenting a first approach to model different perspectives in the input representation of machine learning in argumentation mining. | |||||||
| Lizenz: |  Dieses Werk ist lizenziert unter einer Creative Commons Namensnennung 4.0 International Lizenz | |||||||
| Fachbereich / Einrichtung: | Mathematisch- Naturwissenschaftliche Fakultät » WE Informatik | |||||||
| Dokument erstellt am: | 01.02.2024 | |||||||
| Dateien geändert am: | 01.02.2024 | |||||||
| Promotionsantrag am: | 17.08.2023 | |||||||
| Datum der Promotion: | 22.11.2023 |
