Studi eksperimen metode regularized incremental simultaneous matrix factorization dalam sistem rekomendasi = Experiment Study of Regularized Incremental Simultaneous Matrix Factorization Method on Recommender System
| Main Authors: | Subian Saidi, author, Add author: Hendri Murfi, supervisor, Add author: Djati Kerami, examiner, Add author: Dian Lestari, examiner, Add author: Alhadi Bustamam, examiner |
|---|---|
| Format: | Masters Bachelors |
| Terbitan: |
, 2012
|
| Subjects: | |
| Online Access: |
https://lib.ui.ac.id/detail?id=20317264 |
Daftar Isi:
- <b>ABSTRAK</b><br> Sistem rekomendasi merupakan model penyelesaian masalah yang menerapkan teknik-teknik tertentu untuk memberikan rekomendasi suatu infomaasi, produk dan jasa. Salah sam pendekatan yang digunakan dalam sistem rekomendasi yaitu melalui pendekatan collaborative filtering dengan menggunakan teknik/metode faktorisasi matriks (matrix factorization). Seberapa baik suatu metode diterapkan dalam sistem rekomendasi diukur dari kinerja atau akurasi model tersebut. Penelitian ini menguji kinelja metode Regular-ized Incremental Simultaneous Matrix Factorization (RISMF) dalarn sistem rekomendasi melalui studi eksprimen. Eksperimen dilakukan melalui simulasi komputasi untuk mendapatkan parameter model yang optimal. Hasihmya menunjukkan bahwa akurasi model pada saat parameter mencapai optimal sebesar 0.93. Hasil tersebut membuktikan bahwa metode RISMF cukup baik digunakan dalam sistem rekomendasi. Kata Kunci: Sistem rekomendasi, Collaborarivejilrering, Faktodsasi matriks, akurasi. vi Universitas Indonesia <hr> <>Abstract</b><br> Recommendation System is a problem solving model by using methods to give recommendation some infomtations, products and services. Matrix Factorization for collaborative filtering is one of approach in recommendation system. How well a applied method in recommendation system measure from performance or accuration this model. This research examined performance of Regularized Incremental Simultaneous Matrix Factorization (RISMF) method on recommendation system by experimental study. Experiment was done for get optimum parameter model. The result shown that accuracy value on parameter optimum is 0.93. The result proof that RISMF method good enough used in recommendation system. Key words: Recommender System, Matrix factorization, Collaborative Filtering, Accuracy vii Universitas Indonesia