MUKTI, UMAR HADI (2025) IMPLEMENTASI K-NEAREST NEIGHBORS DENGAN METODE CONTENT-BASED FILTERING UNTUK OPTIMASI PENUGASAN REVIEWER PROGRAM DANA PADANAN. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (426kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (41kB) |
|
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (103kB) |
|
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (45kB) |
|
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (693kB) |
|
|
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (31kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (159kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (1MB) |
Abstract
The manual reviewer assignment process in the 2025 Matching Fund Program creates problems including skill mismatches, slow processes, and lack of transparency. This research develops an automatic recommendation system using K-Nearest Neighbors (KNN) algorithm with Content-Based Filtering method to address these issues. Research data includes 3,154 reviewer histories and 541 proposals. The system employs text preprocessing and TF-IDF feature extraction yielding 4,191 unique features. Optimal parameters obtained are K=3 with average similarity score of 0.6750 and threshold of 0.6. System evaluation shows excellent performance with 71.3% precision, 97.7% recall, and 82.5% F1-Score. The system successfully generates 1,623 reviewer recommendations with 1,117 True Positives and only 26 False Negatives. The system has been integrated into a web application using Laravel and PostgreSQL. Results prove that KNN algorithm with ContentBased Filtering can significantly optimize reviewer assignment, improving efficiency and matching accuracy based on relevant expertise. Keywords: K-Nearest Neighbors, Content-Based Filtering, Machine Learning, Matching Fund, Ministry of Research Technology and Higher Education. Proses penugasan reviewer pada Program Dana Padanan 2025 yang dilakukan dengan manual menimbulkan beberapa masalah seperti ketidaksesuaian keahlian, proses yang lambat, dan kurang transparan. Penelitian ini mengembangkan sistem rekomendasi otomatis menggunakan algoritma K-Nearest Neighbors (KNN) dengan metode Content-Based Filtering untuk mengatasi permasalahan tersebut. Data penelitian meliputi 3.154 histori reviewer dan 541 proposal. Sistem menggunakan preprocessing teks dan ekstraksi fitur TF-IDF yang menghasilkan 4.191 fitur unik. Parameter optimal yang diperoleh adalah K=3 dengan rata-rata skor kemiripan 0,6750 dan threshold 0,6. Evaluasi sistem menunjukkan performa sangat baik dengan precision 71,3%, recall 97,7%, dan F1-Score 82,5%. Sistem berhasil menghasilkan 1.623 rekomendasi reviewer dengan 1.117 True Positive dan hanya 26 False Negative. Sistem telah diintegrasikan ke aplikasi web menggunakan Laravel dan PostgreSQL. Hasil penelitian membuktikan algoritma KNN dengan Content-Based Filtering dapat mengoptimalkan penugasan reviewer secara signifikan, meningkatkan efisiensi dan akurasi pencocokan berdasarkan keahlian yang relevan. Kata kunci: K-Nearest Neighbors, Content-Based Filtering, Program Dana Padanan, Kemendikbudristek, Machine Learning.
Actions (login required)
 |
View Item |
