Abstract:
Fuzzy rule-based classification systems are known to make predictions based on the knowledge that is embedded in their rule bases. Whereas one-shot explanations extracted...Show MoreMetadata
Abstract:
Fuzzy rule-based classification systems are known to make predictions based on the knowledge that is embedded in their rule bases. Whereas one-shot explanations extracted from a rule base can be made interpretable, they suffer from a lack of interaction with the user leaving the social aspect of explanation unaddressed. In this paper, we present a dialogue system allowing for communication of automated textual explanations for the output of fuzzy classifiers to the end user in natural language. Importantly, the underlying explanation communication framework allows the user to iteratively explore the explanation space of the given fuzzy rule-based classifier, making the resulting explanations interactive in accordance with the recent considerations for automated explanations. In addition, we demonstrate the usage of the software with two illustrative use cases. It is worth noting that the software presented in this paper is made publicly available.
Date of Conference: 30 June 2024 - 05 July 2024
Date Added to IEEE Xplore: 05 August 2024
ISBN Information: