Evaluation of Diversity in Movie Recommendation Systems

There are many movie streaming applications today, such as Netflix, Disney Hotstar, Viu, WeTV, iQiYi, and others. The application offers a variety of movies that users can watch, thus confusing users. Each user has a different interest in the content provided. Thus, it is important to use a recommen...

Full description

Saved in:

Bibliographic Details
Published in:	2024 2nd International Conference on Software Engineering and Information Technology (ICoSEIT) pp. 64 - 69
Main Authors:	Chintya, Rere, Kusumo, Dana Sulistiyo, Gandhi, Arfive
Format:	Conference Proceeding
Language:	English
Published:	IEEE 28-02-2024
Subjects:	Clustering algorithms diversity Information technology Measurement Motion pictures Real-time systems recommendation system Recommender systems Testing user satisfaction
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	There are many movie streaming applications today, such as Netflix, Disney Hotstar, Viu, WeTV, iQiYi, and others. The application offers a variety of movies that users can watch, thus confusing users. Each user has a different interest in the content provided. Thus, it is important to use a recommendation system on platforms that provide services to users to provide satisfaction through recommendations so that users feel satisfied. Several studies have focused more on measuring the accuracy aspect of the recommendation system. In comparison, there is an aspect of the recommendation system besides accuracy, namely diversity. Diversity is an aspect that plays a role in providing various recommendations based on users and content on the platform. Evaluation of the diversity aspect of the recommendation system is carried out by implementing the K-means clustering and cosine similarity algorithms using a film dataset. Then, measure it using the Intra-list diversity metric and get a result of 0.65. Tests on users were also carried out regarding user satisfaction with the list of films using a questionnaire so that the median value was 4, and the mode value was between 4 and 5. From the results of the evaluation, it was found that the recommendations provided were diverse and satisfactory for users.
DOI:	10.1109/ICoSEIT60086.2024.10497505