Conceptualisation of Cyberattack Prediction using a Deep Learning Model. – A. E. Ibor, F. A. Oladeji, O. B. Okunoye and B I. Ele
ABSTRACT
The state of the cyberspace portends uncertainty for the future Internet and its accelerated number of users. New paradigms add more concerns with big data collected through device sensors divulging large amounts of information, which can be used for targeted attacks. Though a plethora of extant approaches, models and algorithms have provided the basis of cyberattack predictions, there is the need to consider new models and algorithms, which dwell more on data representations other than task-specific techniques. Deep learning, which is underpinned by representation learning, has found widespread relevance in computer vision, speech recognition, natural language processing, audio recognition, and drug design. However, its non-linear information processing architecture can be adapted towards learning the different data representations of network connection vectors to classify benign and malicious network packets. In this paper, we model cyberattack prediction as a classification problem. Furthermore, the deep learning architecture was co-opted into a new model using rectified linear units (ReLU) as the activation function in the hidden layers of a deep feed forward neural network to realise a greedy layer-by-layer learning process that best represents the features useful for predicting cyberattacks in a dataset of connection vectors. The underlying algorithm of the model also performs feature selection, dimensionality reduction, and clustering at the initial stage, to generate a set of input vectors called hyper-features. The model is evaluated using CICIDS2017 dataset on a Python environment test bed. The results obtained show a 99.99% accuracy for the modeled attack types, and a very low false positive rate of 0. 00001. The proposed model is therefore fit for purpose in the prediction of cyberattacks.