基于卷积门控循环网络的居民出行模式识别

Residential Travel Pattern Recognition based on Convolutional Gated Recurrent Network

关注城市居民出行的特征及其交通方式选择行为,对于实现绿色交通具有重要的现实意义.该文基于GeoLife数据集,利用卷积神经网络(CNN)、门控循环单元(GRU)和混合模型(CNN-GRU)对居民的出行特征及规律进行研究.通过测试和对比分析表明:CNN-GRU模型具有较好的识别效果,且优于单独采用CNN和GRU分类方法的识别性能.

Paying attention to the travel characteristics of urban residents and their transportation mode choice behavior has important practical significance for realizing the development of green transportation.Based on the GeoLife data set,this article uses Convolutional Neural Network(CNN),Gated Recurrent Unit(GRU)and Hybrid Model(CNN-GUR)to study residents'travel characteristics and rules.Tests and comparisons show that the CNN-GRU model has a better recognition effect,and is better than the recognition performance of the CNN and GRU classification methods.