Putra, David (2026) RANCANG BANGUN CHATBOT VIRTUAL ASSISTANT LAYANAN INFORMASI AKADEMIK MENGGUNAKAN METODE TF-IDF DAN COSINE SIMILARITY. S1 / D3 thesis, Universitas Kuningan.
ABSTRAK.pdf
Download (7MB)
BAB I.pdf
Download (489kB)
BAB II.pdf
Restricted to Repository staff only
Download (1MB) | Request a copy
BAB III.pdf
Restricted to Repository staff only
Download (3MB) | Request a copy
BAB IV.pdf
Restricted to Repository staff only
Download (1MB) | Request a copy
BAB V.pdf
Download (381kB)
DAFTAR PUSTAKA.pdf
Download (607kB)
LAMPIRAN.pdf
Restricted to Repository staff only
Download (2MB) | Request a copy
Abstract
Layanan informasi akademik di Program Studi Sistem Informasi Fakultas Ilmu Komputer Universitas Kuningan masih bergantung pada interaksi manual staf dengan keterbatasan jam operasional, memicu penumpukan pertanyaan repetitif dan inkonsistensi penyampaian SOP. Penelitian ini bertujuan membangun Chatbot Virtual Assistant berbasis web yang beroperasi 24 jam menggunakan metode Research and Development (R&D) dengan kerangka kerja Chatbot Development Life Cycle (CDLC). Sistem mengimplementasikan arsitektur hybrid yang menggabungkan Rule-based System sebagai lapisan penyaringan awal dengan Retrieval-based System berbasis algoritma TF-IDF untuk pembobotan kata dan Cosine Similarity untuk pencocokan kueri secara deterministik. Pengujian terhadap 53 skenario kueri menghasilkan akurasi 75,47%, presisi 81,39%, dan recall 87,50% dengan kecepatan komputasi di bawah 2 milidetik. Black-Box Testing menunjukkan seluruh fitur berjalan valid, sementara User Acceptance Testing mencapai 86,77% dengan predikat Sangat Layak. Sistem terbukti efektif menyajikan informasi akademik secara instan, akurat, dan deterministik tanpa risiko halusinasi data.
Academic information services at the Information System Study Program, Faculty of Computer Sciences, Universitas Kuningan still rely on manual staff interactions with limited operational hours, causing an accumulation of repetitive inquiries and inconsistent SOP delivery. This study aims to develop a standalone web-based Virtual Assistant Chatbot operating 24 hours, using the Research and Development (R&D) method with the Chatbot Development Life Cycle (CDLC) framework. The system implements a hybrid architecture combining a Rule-based System as an initial filtering layer with a Retrieval-based System utilizing TF-IDF for keyword weighting and Cosine Similarity for deterministic query matching. Testing on 53 query scenarios yields an accuracy of 75.47%, precision of 81.39%, and recall of 87.50%, with computational speed under 2 milliseconds. Black-Box Testing confirms all features performed validly, while User Acceptance Testing achieves 86.77%, classified as Highly Feasible. The system effectively delivers academic information instantly, accurately, and deterministically without the risk of AI data hallucination.
| Item Type: | Thesis (S1 / D3) |
|---|---|
| Uncontrolled Keywords: | Asisten Virtual, Chatbot, Layanan Akademik, TF- IDF, Academic Services, Cosine Similarity, |
| Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software |
| Divisions: | Fakultas Ilmu Komputer > S1 Sistem Informasi |
| Depositing User: | S.Kom David Putra |
| Date Deposited: | 11 Aug 2026 06:38 |
| Last Modified: | 11 Aug 2026 06:38 |
| URI: | https://rama.uniku.ac.id/id/eprint/6188 |

