A machine-learning model has been designed to predict emerging computer network security threats with greater reported accuracy than support vector machine systems, according to research in the International Journal of Intelligent Information and Database Systems.