Emotion State Detection: A Federated Learning Approach

This study presents the development of a global model based on federated learning for detecting emotional states from textual data. The proposed approach leverages federated learning to ensure privacy by processing data locally on user devices, with only aggregated model updates shared for training. The model has been trained and evaluated using a publicly available Twitter dataset, demonstrating its ability to analyze text and identify emotional states effectively. While this model offers robust emotion detection capabilities, its integration with specific applications or systems will depend on developers, making it adaptable to various use cases. By safeguarding user privacy and providing accurate emotion classification, this work contributes to advancements in sentiment analysis and mental health monitoring.

• Supervision