基于LWSVR的繁忙机场航班滑出时间预测

针对繁忙机场航班滑出时间预测准确率低的问题,结合局部回归和加权支持向量回归,提出基于局部加权支持向量回归的离港航班滑出时间预测模型。该模型采用K最近邻方法,减小训练样本集容量,并为每个预测样本构建一个预测模型。通过计算训练样本与预测样本间的马氏距离,来优化加权支持向量回归中高斯核加权函数的带宽参数,获得加权系数。结合某机场离港航班数据仿真分析,实验结果表明模型在误差允许范围内的预测准确率达到83.33%,模型更加稳定。

Prediction of Flight Taxi-out Time in A Busy Airport Based on LWSVR

Aiming at improving the accuracy of predicting the flight taxi-out time in a busy airport, based on the local regression and weighted support vector regression, a prediction model of the locally weighted support vector regression is proposed. The model uses the K nearest neighbor method to reduce the capacity of the training sample set and build a predictive model for each predicted sample. The bandwidth parameter of the Gaussian weighting function is optimized with the Mahalanobis distance between the forecast sample and training samples, and the weighting coefficients are obtained. Combining the airport departure flight data in simulation analysis, the experimental results show that the accuracy of LWSVR within the error range is 83.33%, and the model is more stable.