Emotion Recognition System Using Deep Learning Algorithm

Emotions form a vital part of communication between humans. Being able to express emotions and the ability to recognize the emotions of others is essential for communication. Unfortunately, not everybody can express emotions in the most obvious ways, and this makes it difficult for others to recognize and understand how these people feel at given points in time. Some autistic patients and people with severe cases of Parkinson's disease are unable to express emotions through regular facial expressions, hindering communication with health practitioners and close relatives. Machine Learning, a subset of artificial intelligence, has over the years proven to be efficient over traditional software development methods in dealing with the dynamic nature of real-world problems, and given the right datasets, machine learning models can be trained to learn unique patterns and traits specific to different models which can ultimately lead to the prediction of emotions. This paper presents a machine-learning model for emotion recognition using brain signals collected through electroencephalography (EEG). The study utilizes the GAMEEMO dataset, which includes EEG data collected from 28 subjects playing different video games designed to evoke various emotions. Five models were imple-mented-deep Learning, Support Vector Machines, K-nearest neighbors, Decision Trees, and multiple linear regression-where the Deep Learning model achieved the best performance with a 96% testing accuracy. The results were validated using 10-fold cross-validation, confirming the robustness of the model.