Developed a machine learning pipeline using Python. I began with data exploration and pre-processing to prepare the dataset for analysis. I then applied supervised learning models for classification to detect DDoS attacks, followed by unsupervised learning techniques for clustering to group similar attack types. The final step was analyzing and explaining the clusters to gain deeper insights into the attack patterns.
-
1-Data_exploration_and_pre-processing.ipynb: The first task of the project is to present the dataset through various data visualization techniques and statistical analysis.
-
2-Supervised_learning.ipynb: The second task consists on classifying the flows according to the attack – supervised classification.
-
3-Clustering_GMM.ipynb & 3-Clustering_KMEANS_DBSCAN.ipynb: In this task, I group flows that produce similar, correlated, or coordinated patterns. I perform the clustering in an unsupervised manner. The goal is to determine whether there are similar "families" of attacks.
-
ML_project.pdf: This file is the project report, which contains detailed instructions and all the key results extracted from the analyses performed during the project. It provides a comprehensive overview of the methodologies used and the findings obtained throughout the study.