关键词: 无线局域网业务流, 预报, 时间序列, 人工神经网络, 小波

Abstract: A number of traffic prediction methods based on time series as Artificial Neural Networks (ANN) and wavelet were briefly described. Their performances were compared using three actual wireless network traffic traces. Wireless traffic with time series model, artificial neural network, and wavelet-base method were separately predicted. The experimental results show the significant advantages of the ANN technique and AutoRegression Integrated Moving Average (ARIMA) predictor performs relatively the worst.

Key words: Wireless Local Area Network (WLAN) traffic, prediction, time series, Artificial Neural Network (ANN), wavelet