Rootkit Detection Using Hybrid Machine Learning Models and Deep Learning Model: Implementation

Rootkits are a type of malicious software designed to exploit system vulnerabilities and evade detection by traditional security mechanisms. This study proposes a comprehensive approach for rootkit detection by leveraging machine learning algorithms, including Logistic Regression, RandomForest, KNearest Neighbors, Support VectorMachine, Naive Bayes, Model Linear Discriminant Analysis, Multi-layer Perceptron, and Deep Neural Network. In addition, hybrid models combining Logistic Regression & Random Forest, K Nearest Neighbors & Support VectorMachine, and NaiveBayes & LDA are introduced to enhance detection accuracy. The Logistic Regression model provides interpretability, while Random Forest offers robust classification performance. K Nearest Neighbors and Support Vector Machine leverage proximity-based classification and hyperplane separation, respectively. Naive Bayes and LDA focus on probabilistic modeling and dimensionality reduction, respectively. Hybrid models aim to capitalize on the strengths and complementarity of individual algorithms for more effective rootkit detection. Through experimentation and evaluation, this research contributes to the development of advanced cybersecurity solutions for combating sophisticated cyber threats.