Sub-Project 2.3: Location Semantics

The widespread use of smart phone has led to the proliferation of location based social networks (LBSNs) such as the Foursquare, and the mobile sharing sites such as the Istagram and Twitter. The LBSNs enable people to announce their presence, while sharing their views, images and experience on the venues (see Figure 2.5). This sub-project focuses on the comprehensive location analytics research to better understand the functions, activities and user community of venues by:

(1) Discovering user's interest and social community. This can be used to generate user profile for venue recommendation within the same locality or globally.

(2) Generating venue ontology and venue semantics. The research aims to determine venue’s multimedia semantics by extracting the multimedia sub-topics of venue depicting its various functions and activities. The venue semantic can be used for venue classification and matching.

(3) Mining live events related to venue. One important aspect of location-based recommendation is to offer live and emerging events related to a venue.

(4) Performing the above analysis by utilizing a range of information and signals ranging from location meta data such as GPS, location-check-ins, images and texts, as well as ambient sounds.

From the derived location analytics, we can better understand what is happening and who are involved in any venue-based events, from which timely recommendation and various services can be provided to enhance users’ experience.

Figure 2.5: Users will check-in, upload photos, and wrrite comments at some location. First order analytics such as user flow, demographics, hotspots, and higher order analytics such as user community, user interest, events, location ontology can be mined from large scale location based user generated live data.


Representative Publication

[1] Xiangyu Chen, Weizhi Nie, Yiliang Zhao, Tat-Seng Chua: Venue Semantics: Multimedia Topic Modeling of Social Media Contents. Internal Report.

[2] Yan-Tao Zheng, Zheng-Jun Zha, Tat-Seng Chua: Mining Travel Patterns from Geotagged Photos. ACM Transactions on Intelligent Systems and Technology (ACM TIST) 3(3): 56 (2012)

[3] Yi-Liang Zhao, Yan-Tao Zheng, Xiangdong Zhou, Tat-Seng Chua: Generating Representative Views of Landmarks via Scenic Theme Detection. International Multimedia Modeling Conference (MMM 2011): 392-402.