International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 07 | Jul 2023
p-ISSN: 2395-0072
www.irjet.net
Music Recommendation System using Euclidean, Cosine Similarity, Correlation Distance Algorithm and Flask Web Application Abhimanyu Umrani1, Vedant Satpute2, Aditya Chandsare3, Yash Umadi4 1Student, School of Computer Science Engineering and Technology, MIT-WPU, Pune, Maharashta, India 2Student, School of Computer Science Engineering and Technology, MIT-WPU, Pune, Maharashta, India 3Student, School of Computer Science Engineering and Technology, MIT-WPU, Pune, Maharashta, India 4Student, School of Computer Science Engineering and Technology, MIT-WPU, Pune, Maharashta, India
---------------------------------------------------------------------***--------------------------------------------------------------------distance, and cosine similarity. The paper focuses on the Abstract - This project's goal was to create a music
incorporation of the web framework Flask in addition to the recommendation algorithms to link the music recommendation model with an intuitive web interface. By serving as a bridge and facilitating smooth communication between the model and the web page, Flask enables users to interact with the system without any hassle. The report details the setup procedure and the integration of the music recommendation model with Flask, offering insights into the system architecture. In order to produce an interesting and user-friendly web page, it highlights how HTML, CSS, and JavaScript are used. The paper also discusses the difficulties encountered during the development process and the methods used to overcome them, as well as the testing and deployment phases. Utilising evaluation metrics, which offer a breakdown of the accuracy and relevance of the generated recommendations, allows for the performance of the recommendation system to be evaluated.
recommendation system that offers consumers suitable tracks depending on their tastes. Different distance and similarity algorithms, such as correlation, Euclidean distance, correlation distance, and cosine similarity, were used in the project. A user-friendly online interface was connected to the recommendation model using the web framework Flask. The architecture of the music recommendation system is described in the report, with an emphasis on the incorporation of Flask. Using HTML, CSS, and JavaScript, it describes the setup procedure, model integration, and user interface design. The phases of testing and deployment are also covered, along with the difficulties encountered and their remedies. In order to evaluate the effectiveness of the recommendation system, evaluation measures were used. The study's findings compare how well various algorithms do at producing precise and pertinent song recommendations. In summary, the project created a music recommendation system that makes use of distance and similarity metrics and uses Flask for easy web interface integration. The results emphasise the system's advantages and disadvantages and offer ideas for future study, such as looking into other metrics and incorporating user input for advancements.
1.2 Background Effective music recommendation systems are necessary due to the fast expansion of digital music platforms and the enormous volume of music that is accessible to users. These systems make use of algorithms and data analysis techniques to provide users with personalised song recommendations, improving their music discovery experience. The development of machine learning, data mining, and collaborative filtering methods has led to an increase in the use of music recommendation systems in recent years. These algorithms can produce recommendations that suit specific interests and preferences by examining user preferences, behaviour, and music attributes. Systems for recommending music heavily rely on distance and similarity measurements. Based on many characteristics like audio qualities, genre, artist, and lyrics, these metrics assess the degree of similarity across songs. These metrics, which gauge the proximity of songs, make it possible to recognise musically related tunes and make precise recommendations. Correlation, Euclidean distance, correlation distance, and cosine similarity are common distance and similarity measures in music recommendation systems. Each metric presents a distinct method for calculating the degree of similarity between songs, and it can be used in accordance with the demands and peculiarities of
Key Words: Music recommendation system, Distance metrics, Similarity metrics, Correlation, Euclidean distance, Correlation distance, Cosine similarity, Flask
1.INTRODUCTION 1.1 Introduction Our lives are not complete without music, since it offers us amusement, inspiration, and a means of self-expression. Nowadays, there is an enormous amount of music available, making it difficult to identify songs that suit a person's tastes. By providing individualised song recommendations based on user interests and behaviour, music recommendation systems have become effective solutions to address this difficulty. In order to provide consumers with useful song recommendations, this study details the creation and deployment of a music recommendation system. The algorithm evaluates the similarity of songs and generates recommendations using a variety of distance and similarity measures, such as correlation, Euclidean distance, correlation
© 2023, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 81