Perancangan Aplikasi Berbasis Android Untuk Pendeteksi Pesan Berbahaya Phishing (Smishing) Pada Instant Messenger WhatsApp Berbahasa Indonesia Menggunakan Pendekatan Deep Learning

Kustiawan, Muhammad Firdho (2024) Perancangan Aplikasi Berbasis Android Untuk Pendeteksi Pesan Berbahaya Phishing (Smishing) Pada Instant Messenger WhatsApp Berbahasa Indonesia Menggunakan Pendekatan Deep Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 5027201005_Undergraduate_Thesis.pdf] Text
5027201005_Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2026.

Download (14MB) | Request a copy

Abstract

Phishing, sebagai bentuk serangan siber, semakin berkembang dan merambah ke perangkat smartphone dan aplikasi instant messenger. Instant messenger, seperti WhatsApp, menjadi platform yang rentan terhadap serangan phishing, khususnya smishing. Smishing ialah jenis phishing yang memanfaatkan pesan teks untuk menipu pengguna. Smishing yang terjadi pada instant messenger sering kali tidak dapat dideteksi dengan mudah sehingga terdapat sebuah celah keamanan. Dalam menanggapi tantangan ini, tugas akhir ini mengusulkan sebuah solusi berupa perancangan aplikasi berbasis Android yang memanfaatkan metode pendekatan deep learning untuk mendeteksi pesan berbahaya phishing (smishing) berbahasa Indonesia dan URL yang disertakan atau dikirim terpisah oleh penyerang pada instant messenger WhatsApp. Arsitektur model deep learning yang dipilih terdiri dari Convolution Neural Network (CNN) untuk menangkap fitur spasial dari data. Kemudian Long Short-Term Memory (LSTM) dan Bidirectional LSTM (BiLSTM) untuk menangkap pola urutan, dan mekanisme attention untuk memberi bobot lebih pada fitur penting. Perancangan aplikasi ini memanfaatkan API WhatsApp WAHA (WhatsApp HTTP API) yang tersedia secara publik untuk mengambil pesan setiap kali ada yang pesan masuk menggunakan webhook. Pengujian terhadap model mencakup pengujian seperti akurasi dan kecepatan model dalam mendeteksi. Pengujian terhadap aplikasi mencakup pengujian seperti fungsional dan kinerja dari fitur deteksi pesan smishing dan URL phishing. Uji coba model didapatkan akurasi, presisi, recall, dan F1 score untuk pesan smishing masing-masing 1 dan URL phishing masing-masing 0,97 dalam tahap pelatihan model. Sedangkan pada tahap uji data tes didapatkan masing-masing 0,94, 0,89, 1,00, dan 0,94 untuk pesan smishing serta 0,90, 0,83, 1,00 , dan 0,91 untuk URL phishing. Kemudian pada uji aplikasi didapatkan rata-rata durasi waktu deteksi aplikasi dalam keadaan foreground sebesar 2,58 detik, background sebesar 2,68 detik, layar kunci sebesar sebesar 2,63 detik dan tidak bisa deteksi dalam keadaan terminate.
========================================================================================================================
Phishing, as a form of cyberattack, is growing and penetrating smartphone devices and instant messenger applications. Instant messengers, such as WhatsApp, are platforms that are vulnerable to phishing attacks, especially smishing. Smishing is a type of phishing that utilizes text messages to deceive users. Smishing that occurs on instant messengers often cannot be detected easily so there is a security gap. In response to this challenge, this final project proposes a solution in the form of designing an Android-based application that utilizes a deep learning approach method to detect malicious phishing (smishing) messages in Indonesian and URLs that are included or sent separately by attackers on the WhatsApp instant messenger. The architecture of the chosen deep learning model consists of a Convolution Neural Network (CNN) to capture the spatial features of the data. Then Long Short-Term Memory (LSTM) and Bidirectional LSTM (BiLSTM) to capture sequence patterns, and an attention mechanism to give more weight to important features. The design of this application utilizes the publicly available WhatsApp WAHA API (WhatsApp HTTP API) to retrieve messages whenever there are incoming messages using a webhook. Testing of the model includes tests such as accuracy and speed of the model in detecting. Testing of the application includes tests such as functional and performance of the smishing message and URL phishing detection features. Model testing obtained accuracy, precision, recall, and F1 score for smishing messages of 1,00 each and phishing URLs of 0,97 each in the model training stage. While in the test data stage, 0,94, 0,89, 1,00, and 0,94 were obtained for smishing messages and 0,90, 0,83, 1,00, and 0,91 for phishing URLs, respectively. Then in the application test, the average duration of application detection time in the foreground state is 2,58 seconds, background is 2,68 seconds, lock screen is 2,63 seconds and cannot be detected in the terminate state.

Item Type: Thesis (Other)
Uncontrolled Keywords: Phishing, Smishing, Instant Messenger, Deep Learning, Text classification, WhatsApp
Subjects: Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Q Science > QA Mathematics > QA336 Artificial Intelligence
Q Science > QA Mathematics > QA76.585 Cloud computing. Mobile computing.
Q Science > QA Mathematics > QA76.76.A63 Application program interfaces
Q Science > QA Mathematics > QA76.76.A65 Application software. Enterprise application integration (Computer systems)
Q Science > QA Mathematics > QA76.774.A53 Android
Q Science > QA Mathematics > QA76.9.U83 Graphical user interfaces. User interfaces (Computer systems)--Design.
R Medicine > R Medicine (General) > R858 Deep Learning
Z Bibliography. Library Science. Information Resources > ZA Information resources > ZA4450 Databases
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User: MUHAMMAD FIRDHO KUSTIAWAN
Date Deposited: 26 Jul 2024 01:00
Last Modified: 26 Jul 2024 01:00
URI: http://repository.its.ac.id/id/eprint/108539

Actions (login required)

View Item View Item