DEEP HYBRID CONTENT-BASED FILTERING FOR INDONESIAN MOVIE RECOMMENDATIONS BASED ON SYNOPSES AND AGE RATINGS FROM THE FILM CENSORSHIP BOARD (LSF)

Aulia, Nawal (2026) DEEP HYBRID CONTENT-BASED FILTERING FOR INDONESIAN MOVIE RECOMMENDATIONS BASED ON SYNOPSES AND AGE RATINGS FROM THE FILM CENSORSHIP BOARD (LSF). Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 22081010097-Cover.pdf Download (701kB)
	Text (Bab I) 22081010097-Bab I.pdf Download (191kB)
	Text (Bab II) 22081010097-Bab II.pdf Restricted to Repository staff only until 20 July 2029. Download (662kB)
	Text (Bab III) 22081010097-Bab III.pdf Restricted to Repository staff only until 20 July 2029. Download (696kB)
	Text (Bab IV) 22081010097-Bab IV.pdf Restricted to Repository staff only until 20 July 2029. Download (912kB)
	Text (Bab V) 22081010097-Bab V.pdf Download (178kB)
	Text (Daftar Pustaka) 22081010097-Daftar Pustaka.pdf Download (183kB)
	Text (Lampiran) 22081010097-Lampiran.pdf Restricted to Repository staff only until 20 July 2029. Download (1MB)

Abstract

The rapid growth of the Indonesian film industry has led to a continuous increase in the number of available films, making it difficult for users to find films that match their preferences and age categories. Recommendation systems can help address this issue; however, most content-based filtering approaches still rely on keyword-based methods that are unable to deeply understand the semantic meaning of synopses. This study aims to develop an Indonesian film recommendation system using a Deep Hybrid Content-Based Filtering approach by leveraging film synopses and age ratings from the Film Censorship Board (LSF). The study uses a dataset of Indonesian films obtained from IMDb and supplemented with LSF age rating information. The research stages include data preprocessing, synopsis feature extraction using the IndoBERT model, combining synopsis features and LSF age ratings through a feature augmentation approach, and calculating similarity levels using Cosine Similarity. The system was developed as a web-based application and generates recommendations based on the theme preferences selected by the user while considering the appropriateness for the viewer’s age. The evaluation was conducted using the Precision@K, Recall@K, Hit Rate@K, and NDCG@K metrics across several variations of the K value. The test results showed the best performance at K=10, with a Precision of 0.504, a Recall of 0.015, a Hit Rate of 0.914, and an NDCG of 0.741. These results indicate that the system is capable of generating relevant recommendations and placing suitable films at the top of the rankings. Thus, the use of IndoBERT and the integration of LSF age ratings have proven capable of improving the quality of Indonesian film recommendations, making them more relevant, contextual, and appropriate for the user’s age category.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Puspaningrum, Eva Yulia	NIDN0005078908	evapuspaningrum.if@upnjatim.ac.id
Thesis advisor	Putra, Chrystia Aji	NIDN0008108605	ajiputra@upnjatim.ac.id

Subjects:

T Technology > T Technology (General) > T58.6-58.62 Management Information Systems

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Nawal Aulia

Date Deposited:

21 Jul 2026 01:41

Last Modified:

21 Jul 2026 02:18

URI:

https://repository.upnjatim.ac.id/id/eprint/56503