Construction and Adaptability Analysis of User's Preference Models Based on Check-in Data in LBSN
This work addresses the need for personalized services in location-based social networks by improving preference modeling, but it appears incremental as it builds on existing methods with a new validation approach.
The study tackled the problem of discovering and validating user preference models from location-based social network check-in data, proposing a multi-channel convolutional neural network to characterize model applicability and showing effectiveness across three datasets.
With the widespread use of mobile phones, users can share their location anytime, anywhere, as a form of check-in data. These data reflect user preferences. Furthermore, the preference rules for different users vary. How to discover a user's preference from their related information and how to validate whether a preference model is suited to a user is important for providing a suitable service to the user. This study provides four main contributions. First, multiple preference models from different views for each user are constructed. Second, an algorithm is proposed to validate whether a preference model is applicable to the user by calculating the stability value of the user's long-term check-in data for each model. Third, a unified model, i.e., a multi-channel convolutional neural network is used to characterize this applicability. Finally, three datasets from multiple sources are used to verify the validity of the method, the results of which show the effectiveness of the method.