应用梯度提升树的小区域无线网络多标签流量预测

当前无线网络流量地理分布不均且可用网络资源有限，因而开展拆闲补忙工作极为必要。为合理投放无线网络资源以保证网络性能，提出了一种针对小区域范围的多标签流量预测算法。该算法结合历史流量信息，根据无线用户偏好特性建立多标签流量预测模型，并通过梯度提升树算法得到预测结果。仿真结果表明，相比于广泛应用的移动平均自回归(Autoregressive Integrated Moving Average, ARIMA)和神经网络预测方法，多标签预测模型在对小区域突发性流量的预测上具有很大的优越性。

Multi-label traffic prediction for small area wireless network via gradient boosting tree

Due to the uneven geographical distribution of wireless network traffic and the limited network resource,it is necessary to optimize IDLE-BUSY adjustment.Therefore,a multi-label traffic prediction algorithm is proposed to distribut rationally wireless network resource to ensure network performance.The multi-label traffic prediction model is established by considering the historical traffic and wireless user preference characteristics.According to the prediction model,gradient boosting tree scheme is used to obtain the prediction result.The simulation results show that,compared with Autoregressive Integrated Moving Average(ARIMA) and neural network prediction algorithm,the multi-label traffic prediction model has great advantage in predicting the burstiness of small area wireless network traffic.