International Conference on Computer and Knowledge Engineering

Home / 14th International Conference on Computer and Knowledge Engineering

Designing an IT2 Fuzzy Rule-based System for Emotion Recognition Using Biological Data

Authors :

Mahsa Keshtkar¹ Hooman Tahayori²

1- Department of Computer Science and Engineering and IT, School of Electrical and Computer Engineering, Shiraz University, Shiraz, Iran. 2- Department of Computer Science and Engineering and IT, School of Electrical and Computer Engineering, Shiraz University, Shiraz, Iran.

Keywords :

emotion recognition،rule-based fuzzy system،effective computing،Physiological signals،FOU،interval type two fuzzy system

Abstract :

Abstract— In recent years, significant efforts have been dedicated to detect human emotions. This interest primarily stems from the fact that emotions influence individuals' reactions and behaviors. Understanding these impacts can be beneficial in various fields, ranging from developing healthcare surveillance systems for the elderly to identifying adverse emotions in those engaged in high-stakes occupations, such as pilots, sailors, and long-haul truckers. This paper presents a solution based on the application of interval type two fuzzy sets and systems to detect individuals' emotions using their physiological signals. The physiological signals on the same emotion varies from person to person and even for the same person in different situations. This variability has led us to employ interval type-2 fuzzy sets. We defined five footprints of uncertainty for each physiological signal to cover its whole range and employed instance selection technique to optimize the rule base. We develop a rule-based fuzzy system that employs the fewest feasible number of rules for the purpose of emotion recognition. In order to enhance the comprehensibility of the system, we have also utilized a minimal set of features in the construction of the rules. This study implemented on physiological signals from the K-EmoCon dataset. Obtained results shows that the proposed method outperform other existing method with lower computational complexity.