Implementasi Sistem Rekomendasi Content-Based Filtering Video Game Menggunakan Algoritma Self-Organizing Map
Total View This Week0
Total View This Week0
Institusion
Institut Teknologi Telkom Purwokerto
Author
Nahdi, Duta Ahmad
Subject
T Technology (General)
Datestamp
2021-09-24 02:59:38
Abstract :
Video game is a type of entertainment that is very popular within all communities. That statement are proven by 192% increase in video game installations in 2016 compared to the previous year. The high growth of the video game industry and the types of communities playing it, as well as the high number of different video game’s genres encouraging the need for video game recommendations according to the behavior of the people within the video game community itself. A content-based recommendation system can solve this problem using the Self-Organizing Maps (SOM) algorithm. In this study, a recommendation system was made to recommend video games according to user behavior using SOM which was tested with 3 (three) different scenarios which also resulted in 3 (three) different recommendation systems. Each scenario was tested for its clustering quality by calculating the silhouette score of each cluster for each scenario. The results of the scenario test are 0,46554 for the results of scenario 1, -0.1104 for the results of scenario 2 and -0.01692 for the results of scenario 3.Based on the results obtained, SOM can be used as a cluster algorithm which the result itself are able to be used as the basis for a video game recommendation system. . The best results are shown in scenario 1 with a silhouette score of 0.05392 without applying dataset normalization. Keywords: Clustering, Recommender System, Self-Organizing Maps, Silhouette score, Video Game
Video game is a type of entertainment that is very popular within all communities. That statement are proven by 192% increase in video game installations in 2016 compared to the previous year. The high growth of the video game industry and the types of communities playing it, as well as the high number of different video game’s genres encouraging the need for video game recommendations according to the behavior of the people within the video game community itself. A content-based recommendation system can solve this problem using the Self-Organizing Maps (SOM) algorithm. In this study, a recommendation system was made to recommend video games according to user behavior using SOM which was tested with 3 (three) different scenarios which also resulted in 3 (three) different recommendation systems. Each scenario was tested for its clustering quality by calculating the silhouette score of each cluster for each scenario. The results of the scenario test are 0,46554 for the results of scenario 1, -0.1104 for the results of scenario 2 and -0.01692 for the results of scenario 3.Based on the results obtained, SOM can be used as a cluster algorithm which the result itself are able to be used as the basis for a video game recommendation system. . The best results are shown in scenario 1 with a silhouette score of 0.05392 without applying dataset normalization. Keywords: Clustering, Recommender System, Self-Organizing Maps, Silhouette score, Video Game
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
