Song Sense: Delving into Music Moods Through Machine Learning

This article describes a machine learning model to identify the feelings generated by songs in a family group. Two models were compared: Support Vector Machines and Random Forest, to determine the precision in musical recommendation. A data set with songs of various genres and associated emotions was used. Natural language processing and feature extraction techniques were applied. The models were evaluated with metrics such as accuracy, precision, and F1 score. Both models performed well in sentiment identification, but Random Forest showed a slight advantage (82% of accuracy in testing). It is concluded that Random Forest is more effective for generating musical recommendations based on feelings. This study demonstrates the potential of machine learning in understanding musical emotions and provides a solid foundation for personalized recommendation systems in music.