Music Recommender System based on Sentiment Analysis Enhanced with Natural Language Processing Technics
Hugo D. Calderon-Vilca, Andres Leonardo Satornicio Medina, Reynaldo Sucari León
Abstract
In the field of computer science, many efforts have been made with respect to music recommendation in order to offer the user songs much more in line with his current context or tastes and thus also reduce the large number of musical pieces found on the web. However, there are few studies that take into account the user’s feelings for this task. In this paper we present a model and recommendation system that emphasizes sentiment analysis to make music recommendations using natural language processing, this is achieved by using different artificial intelligence tools such as Word2Vec to vectorize words and neural networks to recognize the sentimental information of the texts. In the results, we show that this approach improves the recommendation results obtained by 80% for the accuracy metrics.