Analysis of existing algorithms of music recommendation systems

We live in a time characterized by excessive information overload. For example, a user looking for music, goods or videos does not intend to spend a lot of time and delve into the complexities of the search process. In such situations, it is advisable to use recommender systems. Such systems provide a personalized list of items that best meet the user's information needs. Music is one of the most popular areas where the recommender system is used. When visiting any music streaming service (such as Spotify), the user immediately sees a list of recommended songs that they might like. Such systems are quite common on the Web and allow users to save time and effort when searching for the necessary material. Under the hood of such recommender systems, different algorithms can be used, which can be divided into three large groups: collaborative filtering, content filtering, and hybrid filtering. Each algorithm has its own features and use cases, which are considered in a more detailed description of these methods. After a detailed analysis, we designed our own music recommendation system. An algorithm for creating the system has been developed, which is based on a preliminary analysis of already existing algorithms. To describe the entities of the music recommendation system and their relations, we designed a corresponding class diagram.