Detecting phishing emails using machine learning techniques

العناوين الأخرى

الكشف عن رسائل البريد الإلكتروني الخداع عن طريق تقنيات تعليم الآلة

مقدم أطروحة جامعية

al-Sayidah, Said Abd Allah Abd

مشرف أطروحة جامعية

Viktorov, Oleg

أعضاء اللجنة

Alya, Muhammad Ahmad
Abu Shurayhah, Ahmad Adil
al-Jarrah, Muzaffar Munir Fahmi

الجامعة

جامعة الشرق الأوسط

الكلية

كلية تكنولوجيا المعلومات

القسم الأكاديمي

قسم علم الحاسوب

دولة الجامعة

الأردن

الدرجة العلمية

ماجستير

تاريخ الدرجة العلمية

2017

الملخص الإنجليزي

Phishing is a fraud technique used for identity theft where users receive fake e-mails from deceiving addresses that seem as belonging to legitimate and real business in an attempt to steel the receiver’s personal details.

This act endangers the privacy of many users and therefore, researchers work continuously on finding detection tools and developing existing ones.

Classification is one of the machine learning techniques that can be effectively used to detect received phishing emails.

Through this research, varied classification algorithms are discussed and compared, such as; Naïvebayes, Decision Tree (DT), Logistic Regression, Classification and Regression Trees and Sequential Minimal Optimization (SMO).

A new system was built to detect the phishing emails in an integrating between the supervised and unsupervised technique.

In addition, the study compares the manual and automated feature selection groups for the Email.

The experiment was executed using WEKA Tool on a dataset of 4800 Email, 2400 phishing emails and 2400 legitimate emails represented the 47 features of the email structure.

Indicated that the best manually selected groups achieved an equal accuracy level achieved by the automated features group of 98.25 percent.

Also the Decision Tree, J48 and SMO classifiers topped the previously-mentioned algorithms by providing the highest accuracy average in both manual and automated scenarios.

Moreover, an integrated system of multiple classifiers was constructed using the three top algorithms of SMO, Decision Tree, and J48 and the results showed that integrating unsupervised techniques with supervised ones before the testing provides more accurate results of detecting phishing emails with 98.37 for all the features.

التخصصات الرئيسية

تكنولوجيا المعلومات وعلم الحاسوب

الموضوعات

• التعلم الآلي
• البريد الإلكتروني
• التنقيب في البيانات

عدد الصفحات

59

قائمة المحتويات

• APA
• MLA
• AMA

لغة النص

الإنجليزية

نوع البيانات

رسائل جامعية

رقم السجل

BIM-762691