Published Paper Details:

The proposed project is a Ransomware Detection System developed using Python, TensorFlow/Keras, Pandas, and NumPy. It employs lightweight deep learning models, specifically a Multi-Layer Perceptron (MLP) and a Convolutional Neural Network (CNN), to accurately detect and classify Ransomware samples while remaining suitable for resource-limited environments. The MLP model captures nonlinear feature interactions, while the CNN model processes feature vectors as sequences to extract local patterns. Together, these architectures enable robust Ransomware detection with high accuracy and efficiency. To enhance performance, the system incorporates feature selection using mutual information, reducing input dimensionality while preserving critical information. The models are trained and evaluated on a balanced Ransomware dataset, with results measured using Accuracy, Precision, Recall, F1-score, and ROC-AUC. Additionally, a generalization test is performed using a hold-out subset of unseen Ransomware samples, demonstrating the models' ability to adapt to novel attack variants. By combining lightweight deep learning architectures with effective preprocessing, this proposed system provides practical and efficient solution for Ransomware detection in resource-constrained environments such as IoT devices, mobile systems, and embedded platforms.