结合TF-IDF算法和信息熵的城市热点探索--以新加坡为例

Urban Hot Spots Exploration Based on TF-IDF Algorithm and Information Entropy:A Case Study on Singapore

城市热点分布决定了城市的空间结构,为对城市热点进行分析,提出一种基于社交媒体签到数据的城市空间热点分析方法。首先,针对离散化的签到数据进行了时间单元和空间单元的划分,然后对划分后的签到数据进行自相关分析,检验数据的空间聚集特征;之后采用TF-IDF算法挖掘具有空间显著性的高频POI点,以确定该区域的功能划分;最后利用信息熵的概率衡量区域内人群的流动程度。以新加坡2011年1月1日至2012年1月13日的签到数据集为例,进行试验分析。结果表明,新加坡居民的日常活动和新加坡商业中心分结果符合客观事实,反应了居民日常活动及商业经济分布具有高度的相关性。

The distribution of urban hot spots determines the spatial structure of cities.In order to mine urban hot spots,this paper proposes an urban spatial hot spots analysis method based on social media check-in data.Firstly,we divide the social media check-in data into time unit and space unit,and use the autocorrelation analysis to indicate its significant spatial clustering characteristics.Then,we utilize TF-IDF algorithm to mine the high-frequency POI points with spatial significance to determine the functional division of this region.Finally,we use information entropy to measure the degree of crowd flow in the region.This paper takes the Singapore sign-in data from January 1,2011 to January 13,2012 as an example to conduct an experimental analysis.The results show that the daily activities of Singapore residents and the sub-results of Singapore’s commercial centers are consistent with the objective facts,reflecting the high correlation between the daily activities of residents and the distribution of commercial economy.