Klasifikasi Multi Label untuk Deteksi Keseimbangan Emosi Pengguna Media Sosial Menggunakan K-Fold Cross Validation

Social media has grown in popularity, with millions of people using it to engage with and share information worldwide. Social media, in addition to serving as a communication tool, are crucial for expressing the emotions and feelings of users. The widespread use of social media has had a significant impact on people's emotions. In particular, negative emotions are frequently experienced and can have a significant impact on mental health. This study aimed to analyze multiple classification models to discover the optimal model for detecting emotional balance among social media users. The classification models utilized in this study include the K-Nearest Neighbor, Random Forest, Support Vector Machine, Decision Tree, and AdaBoost to identify the best classification model capable of detecting the emotional balance of social media users. Several classification models are applied and compared with the aim of evaluating model performance. This research project employed K-fold cross-validation to evaluate the categorization model by comparing various k values. The Random Forest algorithm achieved the greatest accuracy of 99.90% at a K-Fold cross validation value of 10 and an Area Under the Curve (AUC) value of 100%. Thus, this study successfully found a reliable model for accurately detecting emotions of social media users, which is expected to contribute to the development of mental well-being monitoring systems on social media platforms.