Authors: Xujiao Wang*, Texas State University, Yihong Yuan, Texas State University
Topics: Geographic Information Science and Systems, Spatial Analysis & Modeling, Temporal GIS
Keywords: spatial-temporal data analytics, LBSM, data mining, OPTICS, Vector Space Model
Session Type: Paper
Start / End Time: 8:00 AM / 9:40 AM
Room: Washington 1, Marriott, Exhibition Level
Presentation File: No File Uploaded
Location-Based Social Media (LBSM) provides individuals’ thoughts, geo-referenced footprints, and the corresponding temporal information. The data collected by these types of services can be analyzed to discover human activity patterns. However, LBSM has various data issues such as uneven sampling frequency over space and time, which inevitably affects the reliability of results when deriving human activity patterns from these datasets. To address this issue, this research proposes a measurement to compare LBSM user activity patterns. First, we cluster user check-in locations based on the density-based spatial clustering of applications with noise (DBSCAN). Appropriate clustering parameters are chosen with the help of Elbow method. Second, we then calculate the spatial-temporal similarity of user check-in points using a spatial-temporal vector space model (STVSM). Each vector represents a user’s spatial-temporal pattern based on the frequency of visiting different clusters within different time periods. Third, measure LBSM users’ activity similarity by applying an extended cosine similarity analysis. The results demonstrated the effectiveness of the proposed approach in measuring the spatial-temporal similarity of LBSM user activities.