Application Research On Network Traffic Prediction Method Based On Improved Temporal Convolutional Network

With the development of the times,the quality of network service is facing a great challenge,accurate prediction of port network traffic data plays an important role in improving the quality of network service.The current artificial neural network has some difficulty in further improving the predictive performance of the network because of the randomness of some network parameter initialization.It is also difficult to assign the weight of the integrated network.On the basis of the original algorithm,the improved grey wolf optimizer algorithm improves the algorithm's optimization mechanism,which makes the algorithm perform better in local and global optimization.Therefore,this thesis tries to use the improved grey wolf optimizer algorithm for the optimization of the initial network parameters of the temporal convolutional network and the optimization of integrated network weight,and applies the proposed algorithm to the solution of the problem of network traffic prediction,the main work is as follows:1.The optimization method of temporal convolutional network parameters based on improved grey wolf optimizer algorithm is realized.In view of the problem of randomness in the initialization of temporal convolutional network parameters,this method optimizes the initial weight and bias of its network by introducing the improved grey wolf optimizer algorithm,finds the optimal weight and bias,so improved temporal convolutional network is proposed.By testing on the port network traffic sample,comparing the classical machine learning algorithm and the time series prediction algorithm,the improved algorithm is verified to improve the prediction performance.2.A weight optimization method for network integration based on improved grey wolf optimizer algorithm is studied.Due to ITCN and LSTM exist significant differences in the network structure,introducing the idea of integrated learning,will improve ITCN and LSTM as an integration of individual learning,searching for optimal weighting combination scheme of integrated network by improved grey wolf optimizer algorithm,to get integrated model has better prediction performance in network.3.The algorithm proposed in this paper is applied to solve the problem of port network traffic prediction,and the applicability of machine learning prediction algorithm based on improved grey wolf optimizer algorithm in the field of network traffic prediction is explored.Overall,this thesis applies the improved grey wolf optimizer algorithm to solve the problem of the initialization parameter optimization of the temporal convolutional network and weight optimization in the integrated network model,and use the port network traffic datasets to test it.The experimental results show that the proposed algorithm can effectively improve the prediction accuracy and generalization ability,which has a certain potential in the performance optimization of machine learning,and a broad application prospect in the field of network traffic data prediction.