Applying Principal Component Analysis for Categorized Dimensionality Reduction in DDoS Detection for Software-Defined Networks

<ns3:p>Background The explosive growth of Software-Defined Networks (SDN) has introduced unmatched scalability with increased flexibility, an essential component of this modern, complicated network infrastructure. While machine learning models promise to be a viable approach for detecting Distributed Denial of Service (DDoS) attacks, their efficiency relies on the quality of the engineered features. Methods In this study, an innovative approach for categorizing newly generated features based on domain-specific relevance is applied, followed by Principal Component Analysis (PCA) on each of the categories for dimensionality reduction. These new engineered features represent the originality of the features within the original dataset without losing their integrity by dropping multiple features from the original dataset. These PCA-transformed features, along with other individual features that were not used in the previous step, were merged into a single dataset for further processing using Machine Learning classifiers. This unique methodology not only addresses the curse of dimensionality but also ensures that the meaningful variance within the categories of features is retained. The CICDDoS2019 dataset was used to evaluate the developed model against features engineered from this dataset. The performance of the model was optimized by carefully and strategically selecting features and transforming them into appropriate new features. The algorithm is finely tuned with basic parameters for an effective outcome with reduced run times, and we used the following metrics for evaluation: accuracy, F1-score, precision, recall, and cross-validation to ensure robustness. Results This model achieved an overall accuracy score of 0.97, with a dataset of 50,000 values having a multi-class target column with eight different class for categorization. With an expanded dataset having 100,000 values and the same multi-class target column, the model maintained an accuracy of 0.97, proving the reliability and scalability of the model. Conclusion This planned and logically structured approach underscores the importance of domain-driven feature generation and categorization.</ns3:p>