Year: 2022
Type: conference-paper
Abstract: Controversy analysis is a broad topic where the opinions of different stakeholders are analyzed to identify the various arguments that are stated and classify the positions taken on the subject. Gathering this data can be very time-consuming to do manually and usually is error-prone and not exhaustive. Automated text classification can enhance this process and make it possible to analyze controversial topics by analyzing lists of relevant documents in a short timeframe. In this paper, we propose a 2-step approach to optimize the extraction and classification of arguments in textual data from controversial topics. First, we extract the most relevant paragraphs for the controversy with a retrieval model and then we use an argument mining model to find and classify the relevant arguments. With this method, we are able to successfully characterize long documents and understand the various opinions that are recorded.
Source: HAL
Language: French
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4319784466