Pengmebangan model deteksi anomali berbasis alpha miner pada IoT untuk sistem keamanan pintu

Prabowo, Stefanus Pratama Adhi (2024) Pengmebangan model deteksi anomali berbasis alpha miner pada IoT untuk sistem keamanan pintu. Masters thesis, Universitas Pelita Harapan.

[thumbnail of Title] Text (Title)
Title.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (70kB)
[thumbnail of Abstract] Text (Abstract)
Abstract.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (140kB)
[thumbnail of Chapter1] Text (Chapter1)
Chapter1.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (308kB)
[thumbnail of Chapter2] Text (Chapter2)
Chapter2.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (539kB) | Request a copy
[thumbnail of Chapter3] Text (Chapter3)
Chapter3.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (224kB) | Request a copy
[thumbnail of Chapter4] Text (Chapter4)
Chapter4.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (657kB) | Request a copy
[thumbnail of Chapter5] Text (Chapter5)
Chapter5.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (118kB) | Request a copy
[thumbnail of Bibliography] Text (Bibliography)
Bibliography.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (228kB)
[thumbnail of Appendices] Text (Appendices)
Appendices.pdf
Restricted to Repository staff only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (6MB) | Request a copy

Abstract

IoT devices are increasingly being integrated into modern security infrastructures, creating an urgent need for robust systems capable of identifying and preventing unauthorized access. However, the dynamic and complex nature of IoT ecosystems, with the continuous introduction of new devices and behaviors, presents significant challenges for traditional anomaly detection methods. This research aims to address these challenges by adapting the Alpha Miner algorithm, initially used in process mining, to identify suspicious behavioral patterns within IoT log data. By focusing on improving the detection of abnormal access patterns, this study seeks to strengthen IoT-based security systems and address critical vulnerabilities in current security protocols. This study uses an experimental methodology that includes designing, implementing, and testing an anomaly detection model on log data from IoT system. The Alpha Miner algorithm is applied to extract process models from these logs and identify deviations that signal potential threats. Furthermore, the research explores the integration of the proposed model into existing security systems, ensuring smooth operation and adaptability to the dynamic IoT environment. The research findings demonstrate that the anomaly detection model based on the Alpha Miner algorithm significantly enhances the accuracy of threat detection in IoT-based door security systems. By analyzing 141,616 log entries, the model successfully distinguishes between normal and anomalous patterns, including activities such as normal punch open (84.01%), access denied (1.56%), and person not registered (0.31%). The false positive rate was reduced to 0.77%, while anomalies such as unauthorized access were identified in 0.09% and blank logs in 0.094% of the total recorded activities. Log data visualization generate a process model encompassing 10 main categories, facilitating the identification of risk patterns. Further analysis revealed that 56.03% of initiated activities were classified as access denied, highlighting the significant number of unauthorized access attempts successfully detected. Additionally, recurring patterns such as punch interval too short (0.77%) were identified. These contributions illustrate the model's significant advancements in detecting and preventing threats in IoT systems, while also providing a solid foundation for the development of more effective and adaptive security protocols in the future. / Perangkat IoT semakin banyak diintegrasikan ke dalam infrastruktur keamanan modern, menciptakan kebutuhan mendesak untuk sistem yang kuat dalam mengidentifikasi dan mencegah akses yang tidak sah. Namun, sifat ekosistem IoT yang dinamis dan kompleks, dengan diperkenalkannya perangkat dan perilaku baru secara terus-menerus, menghadirkan tantangan signifikan bagi metode deteksi anomali tradisional. Penelitian ini bertujuan untuk mengatasi tantangan tersebut dengan mengadaptasi algoritma Alpha Miner, yang awalnya digunakan dalam process mining, untuk mengidentifikasi pola perilaku mencurigakan dalam data log IoT. Dengan berfokus pada peningkatan deteksi pola akses abnormal, penelitian ini bertujuan untuk memperkuat sistem keamanan berbasis IoT dan mengatasi kerentanan kritis dalam protokol keamanan saat ini. Penelitian ini menggunakan metodologi eksperimental yang mencakup desain, implementasi, dan pengujian model deteksi anomali pada data log dari sistem IoT. Algoritma Alpha Miner diterapkan untuk mengekstrak model proses dari log ini dan mengidentifikasi penyimpangan yang mengindikasikan potensi ancaman. Selain itu, penelitian ini mengeksplorasi integrasi model yang diusulkan ke dalam sistem keamanan yang sudah ada, memastikan operasi yang mulus dan adaptabilitas terhadap lingkungan IoT yang dinamis. Hasil penelitian menunjukkan bahwa model deteksi anomali berbasis algoritma Alpha Miner secara signifikan meningkatkan akurasi deteksi ancaman pada sistem keamanan pintu berbasis IoT. Dengan menganalisis 141.616 entri log, model berhasil membedakan antara pola normal dan anomali, termasuk aktivitas seperti normal punch open (84,01%), access denied (1,56%), dan person not registered (0,31%). Tingkat false positive berkurang menjadi 0,77%, sementara anomali seperti unauthorized access teridentifikasi pada 0,09% dan log kosong sebanyak 0,094% dari total aktivitas yang tercatat. Visualisasi data log menghasilkan model proses yang mencakup 10 kategori utama, yang memfasilitasi identifikasi pola risiko. Analisis lebih lanjut mengungkapkan bahwa 56,03% aktivitas yang dimulai diklasifikasikan sebagai access denied, menunjukkan jumlah upaya akses tidak sah yang signifikan yang berhasil terdeteksi. Selain itu, pola berulang seperti punch interval too short (0,77%) teridentifikasi. Kontribusi ini menggambarkan kemajuan signifikan dari model ini dalam mendeteksi dan mencegah ancaman pada sistem IoT, sekaligus memberikan dasar yang kuat untuk pengembangan protokol keamanan yang lebih efektif dan adaptif di masa depan.
Item Type: Thesis (Masters)
Creators:
Creators
NIM
Email
ORCID
Prabowo, Stefanus Pratama Adhi
NIM01679230004
shadow.stefanus@gmail.com
UNSPECIFIED
Contributors:
Contribution
Contributors
NIDN/NIDK
Email
Thesis advisor
Murwantara, I Made
NIDN0302057305
made.murwantara@uph.edu
Uncontrolled Keywords: Anomaly detection ; Alpha Miner algorithm ; Door security systems ; IoT security ; Log data analysis
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: University Subject > Current > Faculty/School - UPH Karawaci > School of Information Science and Technology > Master of Informatics
Current > Faculty/School - UPH Karawaci > School of Information Science and Technology > Master of Informatics
Depositing User: Stefanus Pratama Adhi Prabowo
Date Deposited: 20 Feb 2025 09:14
Last Modified: 20 Feb 2025 09:14
URI: http://repository.uph.edu/id/eprint/67123

Actions (login required)

View Item
View Item