基于K-Means交通流量空间特征分析方法

K-Means Traffic Flow Prediction Method Based on Spatial Features

摘要: 随着汽车保有量的上升，交通流量预测已经成为当前研究的一个重点。针对交通中空间特征利用率点的问题，本文针对道路具有的空间特征，提出将空间路段分成四种上下路段关系，并对不同路段上的路段ID、路段长度宽度特征，利用K-means对空间交通流量存在相互的内在影响进行空间特征分析。本方法充分考虑交通流量数据在空间上的特性，利用K-means聚类深度分析空间特征因素及其之间的关系，完成交通流量空间特征分析，为更精准的时空联合预测处理提供更有效的数据。

Abstract: With the increase in car ownership, traffic flow forecasting has become a focus of current research. Aiming at the problem of utilization points of spatial features in traffic, this paper proposes to di-vide road sections into four kinds of relations between upper and lower sections and analyze the spatial characteristics of road ID, road length, and width on different sections by using K-means to analyze the internal influence of spatial traffic flow. This method fully considers the spatial charac-teristics of traffic flow data, uses K-means clustering to analyze the spatial feature factors and the relationship between them in-depth, completes the spatial feature analysis of traffic flow, and pro-vides more effective data for more accurate spatial-temporal joint prediction processing.