a pca-enhanced deep convolutional model for cyber intrusion detection in networks

With the increasing complexity of cyberattacks, such as fifth-generation multi-vector attacks, traditional intrusion detection systems (ids) have limited effectiveness in complex networks like industrial internet of things (iiot) and internet of medical things (iomt). this research proposes a novel deep learning model for intrusion detection that comprises six one-dimensional convolutional layers, batch normalization, dropout layers, and the mish activation function. to enhance computational efficiency, principal component analysis (pca) is applied for dimensionality reduction of the input features. experiments were conducted using the cic-ids2017 dataset in an environment equipped with an intel core i9-13700f processor and an nvidia geforce rtx 4090 gpu. the results indicate that the proposed model achieves an accuracy of 99.96%, a recall of 99.85%, a precision of 98.99%, and a kappa coefficient of 0.998, demonstrating performance comparable to advanced networks like inception v2 and vgg19-bn, while significantly reducing detection time to only 8 milliseconds. compared to cnn and lstm methods, the proposed model offers a better balance between accuracy and speed, making it suitable for real-time applications in sensitive environments.