CAHYO, DWI NUR (2026) IMPLEMENTASI ALGORITMA LOGISTIC REGRESSION UNTUK SISTEM REKOMENDASI DENGAN PERBANDINGAN ALGORITMA SVM DAN RANDOM FOREST PADA APLIKASI ABSENSI MAGANG UNIVERSITAS MERCU BUANA. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
COVER.pdf Download (1MB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
BAB I.pdf Restricted to Registered users only Download (477kB) |
|
|
Text (BAB II)
BAB II.pdf Restricted to Registered users only Download (226kB) |
|
|
Text (BAB III)
BAB III.pdf Restricted to Registered users only Download (124kB) |
|
|
Text (BAB IV)
BAB IV.pdf Restricted to Registered users only Download (1MB) |
|
|
Text (BAB V)
BAB V.pdf Restricted to Registered users only Download (52kB) |
|
|
Text (DAFTAR PUSTAKA)
DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (338kB) |
|
|
Text (LAMPIRAN)
LAMPIRAN.pdf Restricted to Registered users only Download (938kB) |
Abstract
The eligibility evaluation process for internship participants and freelance workers at Universitas Mercu Buana is still conducted manually without adequate digital support, making contract extension decisions dependent on the subjective assessment of department heads and potentially resulting in inconsistent outcomes. This condition drives the need for a data-driven system capable of providing eligibility recommendations automatically and objectively. This study aims to develop a web-based internship attendance application integrating a recommendation system using the Logistic Regression algorithm to classify contract extension eligibility based on four criteria: performance, responsibility, discipline, and communication. The research employs a Research and Development method with a quantitative approach, in which the assessment attributes were determined based on field observation and system requirements discussions. Since the developed system is entirely new with no historical data, a synthetic Dataset of 1,000 records was used with an eligibility Threshold of 24 out of a maximum total of 40. The model was then compared against two benchmark algorithms, namely Random Forest and Support Vector Machine. Evaluation results using the synthetic Dataset show that Logistic Regression achieved an accuracy of 80.5% with the highest AUC value among all tested algorithms at 0.911, demonstrating excellent discriminative capability in distinguishing between eligible and ineligible classes. System testing using Blackbox Testing across 13 scenarios and User Acceptance Testing involving 10 respondents confirmed that all functionalities performed as expected and in accordance with user needs, making this system a viable solution to support more objective and data-driven contract extension decision-making at Universitas Mercu Buana. Keywords: Recommendation System, Logistic Regression, Machine Learning, Internship Attendance, Performance Evaluation Proses evaluasi kelayakan peserta magang dan pekerja lepas di Universitas Mercu Buana masih dilakukan secara manual tanpa sistem digital yang memadai, keputusan perpanjangan kontrak bergantung pada penilaian subjektif kepala bagian dan berpotensi menghasilkan keputusan yang kurang konsisten. Kondisi ini mendorong kebutuhan sistem berbasis data yang mampu memberikan rekomendasi kelayakan secara otomatis dan objektif. Penelitian ini bertujuan membangun aplikasi absensi magang berbasis web yang mengintegrasikan sistem rekomendasi menggunakan algoritma Logistic Regression, untuk mengklasifikasikan kelayakan perpanjangan kontrak berdasarkan empat kriteria yaitu kinerja, tanggung jawab, kedisiplinan, dan komunikasi. Penelitian menggunakan metode Research and Development dengan pendekatan kuantitatif, di mana atribut penilaian ditetapkan berdasarkan hasil observasi lapangan dan diskusi kebutuhan sistem. Karena sistem yang dibangun merupakan sistem baru tanpa data historis, penelitian menggunakan Dataset sintetis sebanyak 1.000 data dengan Threshold kelayakan sebesar 24 dari total maksimal 40. Kemudian model dibandingkan dengan dua algoritma pembanding yaitu Random Forest, dan Support Vector Machine. Hasil evaluasi dengan Dataset sintetis menunjukkan bahwa Logistic Regression menghasilkan akurasi sebesar 80,5% dengan nilai AUC tertinggi di antara seluruh algoritma yang diuji yaitu 0,911, membuktikan kemampuan model yang sangat baik dalam membedakan kelas layak dan tidak layak. Pengujian sistem menggunakan Blackbox Testing terhadap 13 skenario dan User Acceptance Testing dengan 10 responden menunjukkan seluruh fungsionalitas berjalan sesuai yang diharapkan dan sesuai dengan kebutuhan, sehingga sistem ini dapat menjadi solusi nyata dalam mendukung pengambilan keputusan perpanjangan kontrak yang lebih objektif dan berbasis data di Universitas Mercu Buana. Kata kunci: Sistem Rekomendasi, Logistic Regression, Machine Learning, Peserta Magang, Sistem Pendukung Keputusan.
Actions (login required)
 |
View Item |
