Please use this identifier to cite or link to this item: http://dspace.iua.edu.sd/handle/123456789/5738
Title: كشف البريد الإلكتروني المزعج بإستخدام أله المتجهات الداعمه = Detection spam email using Support Vector machines
Authors: أسماء محمد أحمد محمد خوجلي
Keywords: علم الحاسوب
البريد الإلكتروني
Issue Date: 2022
Publisher: جامعة إفريقيا العالمية
Citation: جامعة إفريقيا العالمية ـ عمادة الدراسات العليا و البحث العلمي والنشرـ كلية دراسات الحاسوب ـ قسم علوم الحاسوب
Abstract: Today e-mail messages have become one of the most popular and efficient types of correspondence in the world of the Internet. Because of its importance and the degree of dependence on it, its use is fraught with many risks because it contains important information or personal data. Many problems may occur as a result of using e-mail without sufficient experience. The main problem occurs when spam and malicious emails cannot be identified from the start. Machine learning in this field has proven to play an important role in classifying e-mail messages and detecting annoying and harmful messages. The methods of classification and data identification must be effective to classify the data in a good way, especially in the systems-based computer world. In this research, classification techniques were used. Using the SVM support algorithm on the data set of email messages from the Center for Machine Learning and Smart Systems at the University of California. The main objective is to evaluate the accuracy of data classification with respect to the efficiency and effectiveness of the SVM algorithm in terms of accuracy, recall, precision, specificity and F1 degree. The experimental results showed that (accuracy = 98.7%), (precision = 99.8%), (recall = 98.6%), (Specificity = 99.2%) and (F1 score = 99.1%).
URI: http://dspace.iua.edu.sd/handle/123456789/5738
Appears in Collections:أطروحات الماجستير

Files in This Item:
File Description SizeFormat 
Research.pdf1.26 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.