Research on 5G Network Slicing Type Prediction Based on Random Forest and Deep Neural Network

Network slicing is a new network architecture that provides multiple logical networks on the same shared network infrastructure, each serving a specific business type or industry user. The network slicing feature will provide end-to-end isolation between slices and the ability to customize each slice based on service requirements (bandwidth, coverage, security, latency, reliability, etc.). This paper uses random forest and deep neural network (DNN) algorithms to build intelligent models capable of proactively detecting and eliminating incoming connection-based threats to select the most appropriate network slices, even in the case of network failure. Simulation results show that the random forest algorithm always guarantees 100% prediction accuracy regardless of whether the dataset has multidimensional features, but the time complexity is approximately one order of magnitude more than that of the DNN algorithm; in addition, the use of the DNN algorithm can effectively improve the prediction accuracy after data preprocessing of the dataset.