论文部分内容阅读
比较分析神经网络和粗糙集在数据处理过程中的各自优缺点,提出一种基于二者强耦合集成方式的短时交通流预测模型。首先利用粗集对获取的交通流数据进行预处理,简化神经网络训练样本数据集并通过粗集属性约简提取决策规则;其次,利用所提取的规则直接确定神经网络的隐层数、隐层节点数及节点的相互关系;最后训练神经网络用于短时交通流预测。通过与单纯利用神经网络预测的结果进行比较,发现该模型降低了网络训练时间,提高了预测精度。
By comparing the advantages and disadvantages of neural networks and rough set in data processing, a short-term traffic flow prediction model based on the strong coupling and integration of both is proposed. Firstly, rough set is used to preprocess the obtained traffic flow data to simplify the neural network training sample data set and to extract the decision rules by rough set attribute reduction. Secondly, the extracted rules are used to directly determine the hidden layer number and the hidden layer of the neural network The number of nodes and the relationship between nodes; and finally trained neural network for short-term traffic flow prediction. Compared with the results predicted by the neural network alone, the model is found to reduce the network training time and improve the prediction accuracy.