Academic Journal
Overlapping Aspect-based Argument Cluster Analysis including Cluster Labelling for Opinion Formation with Argumentative Dialogue Systems ...
| Title: | Overlapping Aspect-based Argument Cluster Analysis including Cluster Labelling for Opinion Formation with Argumentative Dialogue Systems ... |
|---|---|
| Authors: | Schindler, Carolin |
| Publisher Information: | Universität Ulm |
| Publication Year: | 2024 |
| Collection: | DataCite Metadata Store (German National Library of Science and Technology) |
| Subject Terms: | Computational argumentation, Aspect-based similarity, Identification of frames, Argument clustering, Naming of frames, Opinion formation, Argumentative dialogue system, DDC 004 / Data processing & computer science, Hierarchical clustering Cluster analysis, Meinungsbildung |
| Description: | In order to form a well-informed opinion on a controversial topic, it is important to consider every aspect of the topic. Manual identification and naming of frames, which are grouping aspects, is an effortful task. Hence, in the herein work, we propose an approach to automate this task in a bottom-up manner. Given a set of argumentative sentences, we first create embeddings in a space that is capable of capturing aspect-based similarity. Afterwards, we group the arguments into frames by utilizing clustering algorithms. Thereby, we consider that an argumentative sentence can belong to more than one frame, which is often ignored or circumvented by other works. Finally, the subtopic represented by each cluster should be named succinctly in a human-understandable way. Moreover, we investigate the employment of an automated identification and naming of frames in applications for opinion formation, such as the argumentative dialogue system BEA. We provide methodological approaches and first results, but there are ... |
| Document Type: | text |
| Language: | English |
| DOI: | 10.18725/oparu-52286 |
| Availability: | https://doi.org/10.18725/oparu-52286 https://oparu.uni-ulm.de/xmlui/handle/123456789/52362 |
| Rights: | Creative Commons Attribution Non Commercial No Derivatives 4.0 International ; https://creativecommons.org/licenses/by-nc-nd/4.0/legalcode ; cc-by-nc-nd-4.0 |
| Accession Number: | edsbas.5E121FA4 |
| Database: | BASE |
| Description not available. |