PROJECT TITLE :

Microscopic and Macroscopic spatio-temporal Topic Models for Check-in Data - 2017

ABSTRACT:

Twitter, along with alternative on-line social networks, like Facebook, and Gowalla have begun to gather hundreds of millions of check-ins. Check-in data captures the spatial and temporal information of user movements and interests. To model and analyze the spatio-temporal aspect of check-in knowledge and see temporal topics and regions, we have a tendency to initial propose a spatio-temporal topic model, i.e., Upstream Spatio-Temporal Topic Model (USTTM). USTTM will discover temporal topics and regions, i.e., a user's alternative of region and topic is full of time in this model. We have a tendency to use continuous time to model check-in information, rather than discretized time, avoiding the loss of data through discretization. In addition, USTTM captures the property that user's interests and activity area can amendment overtime, and users have different region and topic distributions at completely different times in USTTM. However, both USTTM and other connected models capture “microscopic patterns” at intervals one city, where users share POIs, and can't discover “macroscopic” patterns during a world space, where users check-in to totally different POIs. Therefore, we have a tendency to conjointly propose a macroscopic spatio-temporal topic model, MSTTM, using words of tweets that are shared between cities to be told the topics of user interests. We tend to perform an experimental analysis on Twitter and Gowalla information sets from New York Town and on a Twitter US data set. In our qualitative analysis, we have a tendency to perform experiments with USTTM to get temporal topics, e.g., how topic “tourist destinations” changes over time, and to demonstrate that MSTTM indeed discovers macroscopic, generic topics. In our quantitative analysis, we evaluate the effectiveness of USTTM in terms of perplexity, accuracy of POI recommendation, and accuracy of user and time prediction. Our results show that the proposed USTTM achieves better performance than the state-of-the-art models, confirming that it's additional natural to model time as an upstream variable affecting the other variables. Finally, the performance of the macroscopic model MSTTM is evaluated on a Twitter US dataset, demonstrating a considerable improvement of POI recommendation accuracy compared to the microscopic models.