Intan Febyola Putri Dwina Sidabutar, . (2026) SISTEM PENDETEKSIAN JARINGAN BOT PADA PLATFORM TIKTOK MENGGUNAKAN CANOPY CLUSTERING DAN ALGORITMA KRUSKAL. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.
![[img]](http://repository.upnvj.ac.id/style/images/fileicons/text.png) |
Text
Abstrak.pdf Download (215kB) |
|
Text
AWAL.pdf Download (863kB) |
|
Text
BAB I.pdf Restricted to Repository UPNVJ Only Download (307kB) |
|
Text
BAB II.pdf Restricted to Repository UPNVJ Only Download (597kB) |
|
Text
BAB III.pdf Restricted to Repository UPNVJ Only Download (795kB) |
|
Text
BAB IV.pdf Restricted to Repository UPNVJ Only Download (858kB) |
|
Text
BAB V.pdf Download (251kB) |
|
Text
DAFTAR PUSTAKA.pdf Download (239kB) |
|
Text
RIWAYAT HIDUP.pdf Restricted to Repository staff only Download (126kB) |
|
Text
LAMPIRAN.pdf Restricted to Repository UPNVJ Only Download (5MB) |
|
Text
HASIL PLAGIARISME.pdf Restricted to Repository staff only Download (27MB) |
|
Text
ARTIKEL KI.pdf Restricted to Repository staff only Download (338kB) |
Abstract
The rise of coordinated bot activity on TikTok poses a serious threat of public opinion manipulation that cannot be detected using conventional individual-account analysis methods. Supervised learning approaches have limitations in the context of this study because they require labeled data, which is difficult to obtain on TikTok, and some modern models may also demand high computational resources. This study aims to design an unsupervised Graph-based bot network detection system for TikTok by integrating Canopy Clustering and Kruskal's algorithm (Minimum Spanning Tree) into a single modular pipeline. This study employs Design Research Methodology (DRM) combined with an SDLC Prototyping model. Data were collected from public TikTok video comments using TikTokApi across four dataset sizes (100–1,000 comments). The system was built as a six-module pipeline using FastAPI and Cytoscape.js. Comments are represented as TF-IDF vectors, then Canopy Clustering using HNSW is applied as soft pre-filtering before a weighted graph is constructed from mention, reply, content similarity, and co-thread relations. Kruskal MST forms community clusters by cutting weak edges based on the median MST edge weight per component. Each cluster is scored using five indicators: cluster density, content repetition, temporal burst, high frequency, and account age. Evaluation is conducted through functional, integration, and performance testing, as well as Modularity, Silhouette Score, and Conductance metrics.
| Item Type: | Thesis (Skripsi) |
|---|---|
| Additional Information: | [No.Panggil: 2210511065] [Pembimbing 1: Jayanta] [Pembimbing 2: Muhammad Panji Muslim] [Penguji 1: Didit Widiyanto] [Penguji 2: Nurhuda Maulana] |
| Uncontrolled Keywords: | Bot Detection, Canopy Clustering, Graph Analysis, Kruskal MST, TikTok. |
| Subjects: | Q Science > Q Science (General) |
| Divisions: | Fakultas Ilmu Komputer > Program Studi Informatika (S1) |
| Depositing User: | INTAN FEBYOLA PUTRI DWINA SIDABUTAR |
| Date Deposited: | 08 Sep 2026 06:00 |
| Last Modified: | 08 Sep 2026 06:00 |
| URI: | http://repository.upnvj.ac.id/id/eprint/51989 |
Actions (login required)
 |
View Item |