Abstract:
Many successful methods have been proposed for learning low-dimensional representations on large-scale networks, but almost all are designed in inseparable processes, learning embeddings for entire networks even when only a small percentage of nodes are of interest. This is inconvenient, especially on large or dynamic networks where these methods are nearly impossible to implement. This paper formalizes separated matrix factorization and proposes a novel objective function that preserves local and global information. SMF captures more information when factorizing a matrix than approximate SVD algorithms. We also propose SepNE, a simple and flexible network embedding algorithm that learns representations for different subsets of nodes in separate processes. Our algorithm scales to large networks by separability, reducing redundant efforts to embed irrelevant nodes. We discuss ways to leverage high-order proximities in large networks to incorporate complex information into SepNE. SepNE works on several real-world networks of different scales and subjects. On large networks, our approach runs faster than state-of-the-art baselines with comparable accuracy.
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
