Can I read A Novel Hybrid Approach for Improving the Accuracy of the Supervised Link Prediction Based on Graph Structure Features in Social Networks on EtoBox?

A Novel Hybrid Approach for Improving the Accuracy of the Supervised Link Prediction Based on Graph Structure Features in Social Networks by Mohamed Badiy; Fatima Amounas; Moha Hajar is a book available to read on EtoBox.

What is A Novel Hybrid Approach for Improving the Accuracy of the Supervised Link Prediction Based on Graph Structure Features in Social Networks about?

Now a day's, the analysis of social networks has grown signifficantly. Link prediction is a challenging issue in the social network analysis area, which uses existing network information in order to predict the future link. Many different link prediction techniques have been proposed in order to predict the future link in the network. Recently, supervised link prediction has become a growing field in which several research efforts have been made. In this context, this paper deals with a new hybrid approach to supervised link prediction based on the merging of local and global similarity methods. It attempts to improve the performance of supervised link prediction by combining various similarity measures. This research considers multiple topological features for training supervised machine learning classifiers. Practical implementation proved that the proposed approach revealed satisfactory results as compared to the existing methods. Our results show that using both local and global features outperforms similarity features when applied individually. Further, the hybridization of multiple features can achieve the highest accuracy.

Author

Mohamed Badiy; Fatima Amounas; Moha Hajar

Publisher

Springer International Publishing : Imprint : Springer

Published

2022

Language

EN

ISBN

9783031064579

Subjects

Management, Computer Science, Business