Skripsi

SISTEM REKOMENDASI DOSEN PEMBIMBING SKRIPSI MENGGUNAKAN PENDEKATAN HYBRID FILTERING

---

Choosing the right academic supervisor is an important part of writing a thesis, but many students struggle to find the best fit. This study introduces a recommendation system that combines Content-Based Filtering (CBF) and Collaborative Filtering (CF). As part of Natural Language Processing (NLP), the CBF component applies TF-IDF and cosine similarity to process and measure how well a student’s research topic, represented through the title and abstract, aligns with a lecturer’s expertise. Meanwhile, CF uses a user–item matrix to capture historical assignment patterns. The two methods are integrated through a weighted combination. Experimental results show that the hybrid model achieved its best performance when wCBF = 0.2 and wCF = 0.8, reaching Hit@5 of 0.8788, F1 Score of 0.3030, and AUC of 0.8611. Compared to using either CBF or CF alone, the hybrid approach delivers more accurate and balanced recommendations by leveraging both topic relevance and historical assignment patterns. These findings highlight the potential of the hybrid model, particularly its NLP-based component, to support students in selecting suitable supervisors more effectively.

---

Availability

Detail Information

Series Title

-

Call Number

T1897652025

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2025

Collation

xiv, 110 hlm.; ilus.; tab.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

006.307

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Teknik Informatika
Pendekatan Hybrid Filtering

Specific Detail Info

-

Statement of Responsibility

MI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment