基于合作对策的交通量组合预测模型

Combined Forecasting Model of Traffic Volume Basedon Cooperative Game

以高速公路交通量为研究对象,建立一种基于合作对策法的交通量线性组合预测模型。以K近邻非参数回归模型、门控循环单元神经网络模型和支持向量回归机模型为基础,从对策论的角度将模型的组合看成一个合作对策。把组合中的三个单项模型看成三个独立局中人,用组合预测的误差平方和来测定合作的预测性能,根据各模型对合作的贡献大小确定它们在组合模型中的权重,实现了对高速公路交通量的组合预测。实验表明:组合预测模型的性能整体上优于各单项预测模型,权值变化的组合预测模型的性能优于权值固定的组合预测模型。

Taking expressway traffic volume as the research object and based on cooperative game method.,a linear combined traffic volume forecasting model of traffic volume was established.Based on the K nearest neighbors non parametric regression model,the Gated Recurrent Unit network model and the support vector regression model,the combination of the model was considered as a cooperative game from the perspective of game theory.Three single forecasting models were considered as three independent players in the cooperative game,and the predictive performance of the cooperation was measured with the sum of error square of combined prediction.The weights of the models were determined according to the contribution of each model to the combination model to realize the combination forecast of the traffic volume of the expressway.The experimental results show that the performance of the combined prediction model was better than that of the single prediction model,and the performance of the combination prediction model with varying weights is better than that of the combination prediction model with fixed weights.