MAGNETO and DeepInsight: Extended Image Translation with Semantic Relationships for Classifying Attack Data with Machine Learning Models

The translation of traffic flow data into images for the purposes classification in machine learning tasks has been extensively explored recent years. However, method a significant impact on success such attempts. In 2019, called DeepInsight was developed to translate genetic information images. It then adopted 2021 purpose translating network images, allowing retention semantic about relationships between features, model MAGNETO. this paper, we explore and extend research, using MAGNETO algorithm three new intrusion detection datasets—CICDDoS2019, 5G-NIDD, BOT-IoT—and also realm multiclass first One versus Rest model, followed by full task, multiple classifiers comparison against CNNs implemented original model. We have undertaken comparative experiments datasets, CICIDS17, KDD99, UNSW-NB15, as well other state-of-the-art models NSL-KDD dataset. results show that method, without use augmentation, offer boost accuracy when classifying data. Our research shows effectiveness Decision Tree Random Forest type Further potential real-time execution is needed possibilities extending real-world scenarios.