Indonesian Journal of Electrical Engineering and Computer Science Vol. 40, No. 1, October 2025, pp. 490~498 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v40.i1.pp490-498  490 Journal homepage: http://ijeecs.iaescore.com An improved conversation emotion detection using hybrid f-nn classifier Abhishek A. Vichare 1,2 , Satishkumar L. Varma 3 1 Department of Computer, Mukesh Patel School of Technology Management and Engineering, SVKM's NMIMS, Mumbai, India 2 Department of Computer, Pillai College of Engineering, New Panvel, India 3 Department of Information Technology, Dwarkadas J. Sanghvi College of Engineering, Vile Parle, Mumbai, India Article Info ABSTRACT Article history: Received Oct 4, 2024 Revised Mar 19, 2025 Accepted Jul 4, 2025 Emotion recognition from text is a crucial task in natural language processing (NLP) with applications in sentiment analysis, human-computer interaction, and psychological research. In this study, we present a novel approach for text-based emotion recognition using a modified firefly algorithm (MFA). The firefly algorithm is a swarm intelligence method inspired by the bioluminescent communication of fireflies, and it is known for its simplicity and efficiency in optimization tasks. In this paper MFA- based model is evaluated on the international survey on emotion antecedents and reactions (ISEAR) dataset, which includes text entries categorized by various emotions. Experimental results indicate that our approach achieved promising outcomes. Specifically, the proposed method, which combines the firefly algorithm with a multilayer perceptron (MLP), attained an accuracy of 92.07%, surpassing most other approaches reported in the literature. Keywords: Emotion recognition Firefly algorithm Machine Learning NLP Swarm intelligence This is an open access article under the CC BY-SA license. Corresponding Author: Abhishek A. Vichare Department of Computer, Mukesh Patel School of Technology Management and Engineering SVKM's NMIMS Mumbai, India Email: vichare1@gmail.com 1. INTRODUCTION Emotion recognition from text is important research domain in natural language processing (NLP) due to its wide range of applications [1]-[3]. Understanding human emotions embedded in textual data can enhance the performance of various systems such as chatbots, social media analyzing tools, and mental health applications. Understanding emotion is considered as challenging task [1], [2]. Emotion recognition is important because emotions can appear in different ways, like stress. Health psychologists study how to identify emotions to help patients by understanding the connection between physical well-being, stress, and emotional state [3]. Among the numerous datasets available for emotion recognition, the International Survey on Emotion Antecedents and Reactions (ISEAR) dataset stands out due to its comprehensive collection of emotional responses across diverse scenarios [4]-[6]. Traditional ML and DL approaches have been extensively used for emotion recognition tasks. However, these approaches often require large amounts of labelled data and important computational resources, which may not always be feasible. Traditional algorithms have not provided good results on ISEAR dataset. As an alternative, swarm intelligence algorithms, influenced by the collective behaviour of social organisms, offer a promising solution [7], [8]. These algorithms are identified for their simplicity, flexibility, and ability to find feasible solutions in challenging search spaces with relatively low computational cost [9]. Our study aims to involve the Firefly Algorithm to improve how features are selected