Prediction of missing or future link is an arduous task in complex networks especially in the current scenario of big data where networks are growing at a high speed. We investigate into both the supervised and unsupervised learning approaches to solve this problem. Supervised approaches use the latent representation of nodes (representation learning) while unsupervised approaches work on the heuristic score given to each node pair having no edge in between them. In this work, Deep learning concept is explored to predict the missing links in the network as a part of the supervised classification. Our experiment on four real-world datasets represents that deep learning approach outperforms some existing supervised learning methods like the Random forest (RF) and the Logistic Regression (LR).
CITATION STYLE
Dadu, A., Kumar, A., Shakya, H. K., Arjaria, S. K., & Biswas, B. (2019). A study of link prediction using deep learning. In Communications in Computer and Information Science (Vol. 955, pp. 377–385). Springer Verlag. https://doi.org/10.1007/978-981-13-3140-4_34
Mendeley helps you to discover research relevant for your work.