Detection of Cyber Attacks with Machine Learning Algorithms

IMSS25 International Symposium, Düzce, Türkiye, 25 - 27 Eylül 2025, ss.226-239, (Tam Metin Bildiri)

ABSTRACT

Cyber-attack types and application areas have diversified considerably in the present era. The

diversification of cyberattacks has led to a continuous increase in their effects. Consequently,

detecting cyberattacks has become a crucial aspect of cybersecurity. Attacks such as Brute Force,

HTTP DDoS, ICMP Flood, Port Scan, and Web Crawling have gained significant prominence in the

cyber domain. The use of diverse techniques and methods has led to challenges in detecting these

attack types, which, in turn, have undermined the comprehensive security of cyber systems. This

study utilized the MSCAD data set, a frequently used dataset in intrusion detection systems, sourced

from the Kaggle platform. Initially, the data set underwent a series of preprocessing steps, including

data cleaning, removing low-variance features, and eliminating missing values. Subsequently, the

pre-processed data set was analyzed using machine learning algorithms, including Random Forest,

Decision Tree, Neural Network, Support Vector Machine (SVM), and K-Nearest Neighbours (KNN).

Furthermore, the chi-square feature selection algorithm was applied to the pre-processed dataset,

minimizing the number of features. Subsequently, the same machine learning algorithms were applied

to the dataset and analyzed. This study conducted experiments on Orange3 and JupyterLab, and the

results were compared. With the successful results obtained through machine learning algorithms,

improvements can be made to Intrusion Detection Systems...