CAJ | 학술논문

识别海量手机数据中蕴含的行为模式，是地理学的一个研究热点与难点。目前，较多研究针对手机用户移动特征开展，而对停留及其模式的研究则相对较少；其时空分异规律对理解城市人群动态，甚至优化城市系统至关重要。本文根据人们日常时空约束条件定义了手机用户停留，提出了基于海量手机位置数据的手机用户停留模式的提取方法，以深圳市约790万个匿名手机用户一天的海量手机位置数据为例，识别出了覆盖约98%用户的典型停留模式，并结合该城市土地利用的空间分布与分异特征，剖析不同停留模式的手机用户空间分异特征和城市不同区域停留次数的时段分异特征。研究发现：（1）15种停留模式可覆盖约98%的手机用户，而且其一天不同的停留位置数量不超过4个；（2）15种停留模式手机用户在城市区域空间上的分布存在分异现象，严重受制于土地利用的空间分布；（3）城市不同区域停留次数的时段分异特征与该区域常住人口、人口密度，以及区域主要职能和性质存在较强的相关性。研究结论对理解城市手机用户行为模式的群体特征有积极的意义，对城市土地利用的科学决策和城市交通规划与预测有重要参考价值。
식별해량수궤수거중온함적행위모식，시지이학적일개연구열점여난점。목전，교다연구침대수궤용호이동특정개전，이대정류급기모식적연구칙상대교소；기시공분이규률대리해성시인군동태，심지우화성시계통지관중요。본문근거인문일상시공약속조건정의료수궤용호정류，제출료기우해량수궤위치수거적수궤용호정류모식적제취방법，이심수시약790만개닉명수궤용호일천적해량수궤위치수거위례，식별출료복개약98%용호적전형정류모식，병결합해성시토지이용적공간분포여분이특정，부석불동정류모식적수궤용호공간분이특정화성시불동구역정류차수적시단분이특정。연구발현：（1）15충정류모식가복개약98%적수궤용호，이차기일천불동적정류위치수량불초과4개；（2）15충정류모식수궤용호재성시구역공간상적분포존재분이현상，엄중수제우토지이용적공간분포；（3）성시불동구역정류차수적시단분이특정여해구역상주인구、인구밀도，이급구역주요직능화성질존재교강적상관성。연구결론대리해성시수궤용호행위모식적군체특정유적겁적의의，대성시토지이용적과학결책화성시교통규화여예측유중요삼고개치。
Identifying human behavior patterns embedded in massive mobile phone data records is a research hot-spot and difficulty of study in geography. While lots of researches are aiming at investigation of the mobility characteristics of mobile phone users, studies focusing on the stay of mobile phone users and their stay patterns are relatively little. The spatio-temporal heterogeneous regularity is vital for understanding the urban human dy-namic and optimizing urban system. Therefore, this paper explored the stay patterns of mobile phone users in or-der to understand the human mobility patterns and their features. According to the spatial and temporal con-straints of human routine and travel behavior, this paper defined the stay conditions and the stay patterns of ur-ban mobile phone users, and come up with methods of identifying and extracting stay patterns from massive mo-bile phone data records. Then, we took about 7.9 million mobile phone users’one day data records in Shenzhen City in China as an example, processed them with the designed method, and identified some typical stay patterns which covered about 98%of the urban mobile phone users. In addition, we conducted statistical analysis for ur-ban mobile phone users’stay patterns and analyzed their spatial distribution with respect to different administra-tive districts combining with the city’s land use pattern. Furthermore, we dissected the heterogeneous character-istics of the spatial distribution of different stay pattern’s users and the stay frequencies against different administra-tive districts over a 22-hour period. The study shows:(1) 15 types of stay patterns can cover 98%urban mobile phone users’behavior and their stay locations are no more than 4 in one day. The stay patterns with 2 stay loca-tions shows the highest probability that nearly half of the mobile phone users have this tendency. (2) The spatial distribution of the 15 stay patterns in the city is evidently heterogeneous, which is subjected to the urban land use. We found that the higher proportion of leisure and entertainment land in a certain administrative district, the stay patterns with more stay locations in this region are likely to occur. (3) The heterogeneity of the stay frequen-cy for different urban administrative district is strongly related to its permanent resident population, population density and major function of the district. The study conclusions are useful in understanding the population char-acteristics of urban phone users’behavior, the activity patterns and urban residents’commuting behavior. More-over, it can provide important references for scientific decisions of urban land use, urban transportation planning and prediction.