Chaivin, Pamela (2018) Sistem rekomendasi drama Korea dengan teknik content based method = Korean drama recommendation system with content based method. Bachelor thesis, Universitas Pelita Harapan.
Full text not available from this repository.Abstract
Ketertarikan pencarian drama Korea cenderung meningkat dari tahun 2004 sampai sekarang, begitu juga dengan peningkatan jumlah drama korea per tahunnya. Karena itu penonton tidak pernah kehabisan pilihan tetapi justru sebaliknya, penonton kesulitan memilih. Selain itu, pembahasan topik rekomendasi semakin meningkat. Dengan teknik content based method, penelitian ini membuat sistem rekomendasi drama korea.
Data yang digunakan adalah data drama tahun 2016 sebanyak 134 drama korea yang diambil dari wiki.d-addicts.com. Rekomendasi dihasilkan dari perhitungan bobot term frequency - inverse document frequency dan perhitungan kemiripan cosine similarity. Untuk meningkatkan akurasi sistem, feedback user diperlukan dan akan diproses dengan Rocchio relevance feedback. Setelah itu sistem akan diuji dengan presisi dan recall. Dari pengujian hasil awal sistem terhadap 50 user didapatkan hasil presisi awal sistem sebesar 82,69% dan hasil recall awal sistem sebesar 82,26%. Kemudian, sistem diuji lagi terhadap 30 user untuk mengetahui pengaruh Rocchio relevance feedback terhadap hasil rekomendasi. Hasil pengujian menunjukkan nilai presisi sistem meningkat sebesar 32,38% pada hasil kedua dan meningkat lagi sebesar 3,198% pada hasil ketiga.
/
Korean drama search interests tend to increase from 2004 until now, as well as the increase in the number of Korean dramas per year. Hence the audience never runs out of options but quite the opposite, the audience has difficulty choosing. In addition, the discussion of recommendations topics is increasing. With contentbased method technique, this research makes the recommendation system of Korean
drama. The data used are drama from year 2016 in total of 134 Korean dramas taken from wiki.d-addicts.com. Recommendations calculation is based on the term frequency - inverse document frequency and cosine similarity. To improve system accuracy, user feedback is required and will be processed with Rocchio relevance feedback. After that the system will be tested with precision and recall. From the test to 50 users, system obtained the initial precision result of 82.69% and the initial recall result of 82.26%. Then, the system is tested again to 30 users to determine the effect of Rocchio relevance feedback on the recommendation system. The test results showed the system precision value increased by 32.38% in the second result and increased again by 3.198% in the third result.
Item Type: | Thesis (Bachelor) |
---|---|
Creators: | Creators NIM Email ORCID Chaivin, Pamela NIM00000009681 UNSPECIFIED UNSPECIFIED |
Contributors: | Contribution Contributors NIDN/NIDK Email Thesis advisor Saputra, Kie Van Ivanky NIDN0401038203 kie.saputra@uph.edu Thesis advisor Ferdinand, Ferry Vincenttius NIDN0323059001 ferry.vincenttius@uph.edu |
Additional Information: | SK 112-14 CHA s 2018 |
Uncontrolled Keywords: | sistem rekomendasi berbasis konten; term frequency - inverse document frequency; cosine similarity; rocchio relevance feedback |
Subjects: | Q Science > QA Mathematics |
Divisions: | University Subject > Current > Faculty/School - UPH Karawaci > Faculty of Science and Technology > Mathematics Current > Faculty/School - UPH Karawaci > Faculty of Science and Technology > Mathematics |
Depositing User: | Mrs Veronica Fitri Astuti |
Date Deposited: | 12 May 2021 14:21 |
Last Modified: | 31 Oct 2023 05:30 |
URI: | http://repository.uph.edu/id/eprint/18420 |