A Survey on Speech Emotion Based Music Recommendation System

Emotion-based music recommendation is a difficult task that has garnered a lot of interest from researchers lately. This research suggests a novel use of support vector machines (SVMs) for emotion-based music recommendation. In order to better properly capture the user’s emotional state, our method employs images as inputs rather than pronouns. Using convolutional neural networks (CNNs), we gathered a collection of photos linked to various emotional states and extracted features from the images. Next, using the input image as a basis, we trained an SVM model to predict the user’s emotional state. Lastly, we recommended music songs based on the user’s emotional state using the SVM model’s output.