Abstract:
Social network analyses and applications face the user alignment problem, which matches users across networks. State-of-the-art methods embed users into the low-dimensional representation space, where their features are preserved, and establish user correspondence based on their low-dimensional embedding similarities. Many embedding-based methods align latent spaces of two networks by learning a mapping function before computing similarities. Most of them learn the mapping function based on limited labeled aligned user pairs and ignore the distribution discrepancy of user representations from different networks, which may cause overfitting and performance issues. A cycle-consistent adversarial mapping model to establish user correspondence across social networks addresses the above issues. Adversarial training between mapping functions and discriminators and cycle-consistency training address the representation distribution discrepancy as the model learns mapping functions across latent representation spaces. The proposed model also trains with labeled and unlabeled users, which may reduce overfitting and the number of labeled users needed. Extensive experiments show that the proposed model aligns users on real social networks.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
