Sistem Klasifikasi Email Phishing Berbasis Fitur Stylometric dengan Support Vector Machine dan Shapley Additive Explanations

Ningrum, Lintang Syawalani Hawa (2026) Sistem Klasifikasi Email Phishing Berbasis Fitur Stylometric dengan Support Vector Machine dan Shapley Additive Explanations. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002221058-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Perkembangan teknologi informasi telah mempermudah komunikasi melalui email, namun juga meningkatkan ancaman *email phishing* yang bertujuan memperoleh informasi sensitif pengguna. Metode deteksi *phishing* yang umum digunakan masih mengandalkan analisis URL, domain, dan *blacklist*. Pendekatan tersebut memiliki keterbatasan ketika *email phishing* memiliki isi dan gaya penulisan yang menyerupai email resmi. Oleh karena itu, penelitian ini membangun model klasifikasi *email phishing* menggunakan fitur *stylometric*, yaitu fitur yang merepresentasikan karakteristik gaya penulisan email. Penelitian ini menggunakan 60 fitur *stylometric* yang diekstraksi dari isi email sebagai masukan pada algoritma *Support Vector Machine* (SVM). Untuk menjelaskan hasil klasifikasi yang dihasilkan model, penelitian ini menggunakan *Shapley Additive Explanations* (SHAP) sebagai pendekatan *Explainable Artificial Intelligence* (XAI). Model yang dibangun kemudian diimplementasikan ke dalam aplikasi berbasis web menggunakan Streamlit yang menyediakan dua mode penggunaan, yaitu input manual dan integrasi Gmail melalui API dengan hak akses *read-only*. Berdasarkan hasil evaluasi, model memperoleh nilai *accuracy* sebesar 88,87%, *precision* sebesar 87,93%, *recall* sebesar 91,11%, *F1-score* sebesar 89,49%, dan *ROC-AUC* sebesar 95,09%. Hasil analisis SHAP membantu menunjukkan fitur-fitur yang paling berpengaruh terhadap hasil klasifikasi model. Model yang dibangun berhasil diimplementasikan ke dalam aplikasi berbasis web sehingga pengguna dapat melakukan klasifikasi *email phishing* dan *legitimate email* serta melihat hasil interpretasi SHAP secara langsung.
=======================================================================================================================================
Advances in information technology have facilitated communication through email but have also increased the threat of phishing emails aimed at obtaining users' sensitive information. Existing phishing email detection methods generally rely on URL analysis, domain analysis, and blacklists. However, these approaches have limitations when phishing emails closely resemble legitimate emails in both content and writing style. Therefore, this study develops a phishing email classification model using *stylometric* features, which represent the writing style characteristics of email content. A total of 60 *stylometric* features extracted from email bodies were used as input to the *Support Vector Machine* (SVM) algorithm. To explain the model's classification results, *Shapley Additive Explanations* (SHAP) was employed as an *Explainable Artificial Intelligence* (XAI) approach. The developed model was subsequently implemented in a web-based application using Streamlit, providing two usage modes: manual input and Gmail integration through a read-only API. Based on the evaluation results, the model achieved an *accuracy* of 88.87%, *precision* of 87.93%, *recall* of 91.11%, an *F1-score* of 89.49%, and a *ROC-AUC* of 95.09%. SHAP analysis identified the most influential features contributing to the model's classification decisions. The developed model was successfully deployed as a web-based application, enabling users to classify both phishing and legitimate emails while providing real-time SHAP-based explanations for the classification results.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Phishing, Stylometric, Support Vector Machine, Explainable AI, SHAP, Phishing, Stylometric, Support Vector Machine, Explainable AI, SHAP
Subjects:	Q Science > QA Mathematics > QA336 Artificial Intelligence T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Lintang Syawalani Hawa Ningrum
Date Deposited:	04 Aug 2026 02:02
Last Modified:	04 Aug 2026 02:02
URI:	http://repository.its.ac.id/id/eprint/142696

Actions (login required)

View Item