Aktuelles

Ein Papier akzeptiert für die RATIO 2024

[23.04.2024]

Das Papier "Cluster-Specific Rule Mining for Argumentation-Based Classification" von Jonas Klein, Isabelle Kuhlmann und Matthias Thimm wurde für die RATIO 2024 akzeptiert.


Inhalt des Papiers

"Cluster-Specific Rule Mining for Argumentation-Based Classification"

von Jonas Klein, Isabelle Kuhlmann und Matthias Thimm

We present a multi-step classification approach that combines classical machine learning methods with computational models for argumentation. In the first step, the dataset is divided into different groups using a clustering algorithm. In the second step, we employ rule-learning algorithms to extract frequent patterns and rules from each resulting cluster. In the last step, we interpret the rules as the input for structured argumentation approaches. Given a new observation, we first assign it to one of the previously generated clusters.
Subsequently, the classification of the observation is determined by formulating arguments based on the respective cluster-specific rules for the different classes. Finally, the justification status of the arguments is determined using the argumentative inference method of the structured argumentation approach.


Weitere Informationen zur RATIO 2024

10.05.2024