Abstract—Detecting and classifying multi-brand emotions on Facebook, Twitter and other social media can be challenging work due to the general nature of the linguistics used in these styles of platforms. Most of the previous studies focused mainly on single-label emotion detection, which detected only one emotion in a very given text. But human expressions are multiple, they should have more emotions with different semantics. Thus, during this study, we mainly focus on multi-label emotion classification, which can classify the possible different emotions in a very given text or data. Multi-label categorization became the first solution among classification problems. Multi-label emotion categorization can be a supervised learning that focuses on multi-label emotion classification from given data, which has a wide selection of implementations in marketing, e-learning, education, and medical management, etc., to make the classification more effective. use hand-curated datasets labeled for the basic eight categories of emotions, these 8 basic emotions are based on Plutchik's model, which has a physiological purpose for each.

 Keywords – Emotion classification , Multi-label classification , NLP , Plutchik's wheel of Emotion.

| DOI: 10.17148/IARJSET.2023.10560