IMPLEMENTASI METODE AGGLOMERATIVE HIERARCHICAL CLUSTERING UNTUK SISTEM REKOMENDASI FILM

Abstract

People can now watch movies on their cellphones or other devices using applications, in addition to watching them on television or in theaters. The user's entered keywords are used as the basis for a system that suggests movies from among the many that have appeared over time. Later, similarity between these keywords and text data, such as movie titles and descriptions, will be assessed. This recommendation system will include preprocessing, and the TF-IDF method will be used to determine the weight value. After the weight values have been determined, the grouping calculations will be performed using agglomerative hierarchical clustering. Previously, the Manhattan Distance method will be used to calculate the distance. After that, the distance that is closest can be determined. The data will be clustered according to the shortest distance once the distance calculation is complete. Following that, the system will display the grouping as a dendrogram. The data used was updated as of the date of scraping, which is November 25, 2022, and contains a total of 2467 data. The Agglomerative Hierarchical Clustering method yielded the best silhouette coefficient value, 0.5025559374455285, forming 20 clusters.