Community detection and observation in large-scale transaction-based networks

Understanding community structures in a graph give an insight into the fundamental properties of a network by observing the characteristics and relationship between the nodes. This thesis will explore the state of the art community detection algorithms for large-scale networks preferably in blockchain/distributed ledger domain. As the popularity for both community detection and blockchain grows, research interest in respective domain grows simultaneously. Detecting community in the large-scale network is challenging because of time and space complexity of the underlying algorithm and to observe the change in the network requires additional approaches. This thesis proposes a prototypical framework for detecting community structures in blockchain data and observing changes in communities afterward. All the implementation steps of the framework are clearly defined. It evaluates the framework with respect to time and space complexity with different well-known community detection algorithms. It also proposes additional steps to observe changes in the community at any given time-stamp. This framework can be easily modified to suit the need of observing changes in any blockchain network. Significant improvement of this prototypical framework can be done in processing time-stamped data-sets by using high-end systems, parallel or distributed computing.

Personalized Web Recommendation Combining User-centered Collaborative Technique with URL Weighting

Dalwar Hossain, A H M Sofi Ullah, K M Habibullah and Md Ali Al Mamun. Article: Personalized Web Recommendation Combining User-centered Collaborative Technique with URL Weighting. International Journal of Computer Applications 63(2):13-18, February 2013. Published by Foundation of Computer Science, New York, USA.