With increased relevance, emotion detection using electroencephalogram (EEG) signals is expected to grow considerably in the number of applications and impact on the society’s understanding of human emotions. The present article offers a systematic approach for emotion recognition using EEG data, which include a dataset expanded into positive, negative, and neutral emotive categories. It is important to note that we don’t only focus on the end-result, but also integrate the exploratory data analysis with careful preprocessing so as to maintain the level of relevance of the data. It is in this context that we managed to construct a functional model for classification of emotional states using ensemble learning algorithms with the aim of maximizing accuracy and visual representation of decision boundaries. This framework puts emphasis on the usefulness of ensemble learning in performing emotion recognition through the application of EEG signals.