论文部分内容阅读
微博客消息中可能蕴含大量描述城市道路的交通信息,如交通状况、交通事件、交通管制等,提取这些交通信息能够为传统的固定式传感器和浮动车采集交通信息手段提供有效补充。然而,微博客消息描述的模糊性、差异性及非结构化特征,使得从海量微博客消息中快速准确地提取和甄别交通信息成为难题。提出一种从微博客消息中快速提取和融合交通信息的技术方法,首先对采集到的微博客消息进行分词解析和路网匹配,然后采用基于神经网络的模糊C聚类方法对描述路段交通状态的微博客消息定量化结果进行分析,获取各路段置信度最高的交通状态描述,最后得到各路段的交通畅通度水平。基于新浪微博客和北京路网的实验过程验证了本文技术方法的有效性。
The microblogging messages may contain a lot of traffic information that describes the urban roads, such as traffic conditions, traffic accidents, traffic control, etc. The extraction of these traffic information can effectively supplement the traditional stationary sensors and floating vehicles to collect traffic information. However, the vagueness, diversity and unstructured features of the microblogging message make it difficult to extract and identify traffic information quickly and accurately from the mass of microblog messages. This paper proposes a technique to quickly extract and integrate traffic information from microblog messages. Firstly, word segmentation and road network matching are adopted for the collected microblog messages. Then fuzzy C-clustering method based on neural network is used to describe the traffic status The micro-blog message quantification results to obtain the highest confidence level of each section of the traffic state description, and finally get the traffic smoothness of the various sections of the level. Based on the experiment of Sina Weibo and Beijing Road Network, this paper validates the effectiveness of this method.