Spam Detection in Emails Using Machine Learning Techniques: A Review

Spam Detection in Emails Using Machine Learning Techniques: A Review

Spam, Class Imbalance, SMOTE, Feature Selection, Ensemble Learning

Despite the vast amounts of data available within email communication systems, spam remains a persistent issue, posing challenges for both users and organizations. Analyzing this data holds the potential to develop more effective methods for detecting and mitigating spam emails. However, extracting actionable insights from this data and leveraging them to construct robust spam detection systems presents a significant challenge. Traditional approaches to combating spam, such as rule-based filtering and heuristic methods, have become increasingly inadequate due to the evolving tactics of spammers. Machine learning techniques offer a promising solution by enabling the training of predictive models using historical email data. However, the effectiveness of these models is influenced by factors such as class imbalance and the identification of relevant features essential for spam detection. This paper provides a comprehensive review of various machine learning techniques employed in spam detection within email communication systems. By examining the strengths and weaknesses of different approaches, we aim to identify strategies for improving the efficiency and accuracy of spam detection. Additionally, we propose a spam detection framework centered around ensemble learning models trained on balanced datasets using techniques like SMOTE, and featuring only the most relevant features. This approach is intended to enhance detection performance while reducing false positives, thereby offering a more effective solution to the challenge of spam detection in email systems.

Stanley Munga Ngigi1, Richard Mathenge3, Josphat Karani2, Nicholus Muriithi4

www.ijcit.com

September 2024

peri irawan

pdf

english

text