Partha Basuchowdhuri

Detection of Communities in Social Networks Using Spanning Tree

Communities are inherent substructures present in social networks. Yet finding communities from a social network can be a difficult task. Therefore, finding communities from a social network is an interesting problem. Also, due to its use in many practical applications, it is considered to be an important problem in social network analysis and is well-studied. In this paper, we propose a maximum spanning tree based method to detect communities from a social network. Experimental results show that this method can detect communities with high accuracy and with reasonably good efficiency compared to other existing community detection techniques.

Finding Influential Nodes in Social Networks Using Minimum k-Hop Dominating Set

Challenges in social interaction networks are often modelled as graph theoretic problems. One such problem is to find a group of influential individuals of minimum size or the initial seed set in a social network, so that all the nodes in the network can be reached with only one hop from the seeds. This problem is equivalent to finding a minimum dominating set for the network. In this paper, we address a problem which is similar to finding minimum dominating set but differs in terms of number of hops needed to reach all the nodes. We have generalized the problem as k-hop dominating set problem, where a maximum of k hops will be allowed to spread the information among all the nodes of the graph. We show that the decision version of the k-hop dominating set problem is NP-complete. Results show that, in order to reach the same percentage of nodes in the network, if one extra hop is allowed then the cardinality of the seed set i.e. the number of influential nodes needed, is considerably reduced. Also, the experimental results show that the influential nodes can be characterized by their high betweenness values.

Spread of information in a social network using influential nodes

Viral marketing works with a social network as its backbone, where social interactions help spreading a message from one person to another. In social networks, a node with a higher degree can reach larger number of nodes in a single hop, and hence can be considered to be more influential than a node with lesser degree. For viral marketing with limited resources, initially the seller can focus on marketing the product to a certain influential group of individuals, here mentioned as core . If k persons are targeted for initial marketing, then the objective is to find the initial set of k active nodes, which will facilitate the spread most efficiently. We did a degree based scaling in graphs for making the edge weights suitable for degree based spreading. Then we detect the core from the maximum spanning tree (MST) of the graph by finding the top k influential nodes and the paths in MST that joins them. The paths within the core depict the key interaction sequences that will trigger the spread within the network. Experimental results show that the set of k influential nodes found by our core finding method spreads information faster than the greedy k -center method for the same k value.

Analysis of Product Purchase Patterns in a Co-Purchase Network

Real world social networks are often temporal in nature. They evolve with time as new nodes may appear, old nodes may cease to exist and the relationships between the entities may also change. In this paper, we have studied a publicly available dynamic network dataset, namely Amazon co-purchase network dataset. We have analyzed the network data to understand the significance of nodes with high in-degree and high out-degree. We subsequently analyze the evolution of communities in the network by observing the change of association among nodes and by observing communities to see how many members they could retain over time. We show some frequent item sets from the co-purchase market basket in terms of the item categories and subcategories. We finally select some central entities and group of such entities from the network to recommend how promoting some of the co-purchased items may increase the sales of the selected items.

Spanning tree-based fast community detection methods in social networks

In this paper, we address the problem of community detection in social networks. We present two maximum cost spanning tree-based community detection methods, namely P-SPAT and K-SPAT, for social networks. Communities are defined as dense subgraphs present in social networks. However, detecting communities in a social network can still be a challenging task, in terms of computational overheads and accuracy of the detected communities. Therefore, finding communities from a social network is considered to be an interesting problem. Again, due to practical applications of community detection techniques, it is a key area of research in social network analysis and is also well studied. Experimental results show that these methods can detect highly accurate communities faster than the state-of-the-art community detection techniques.

Analysis of Online Product Purchase and Predicting Items for Co-purchase

In recent years, online market places have become popular among the buyers. During this course of time, not only have they sustained the business model but also generated large amount of profit, turning it into a lucrative business model. In this paper, we take a look at a temporal dataset from one of the most successful online businesses to analyze the nature of the buying patterns of the users. Arguably, the most important purchase characteristic of such networks is follow-up purchase by a buyer, otherwise known as a co-purchase. In this paper, we also analyze the co-purchase patterns to build a knowledge-base to recommend potential co-purchase items for every item.

Fast detection of community structures using graph traversal in social networks

Finding community structures in social networks is considered to be a challenging task as many of the proposed algorithms are computationally expensive and does not scale well for large graphs. Most of the community detection algorithms proposed till date are unsuitable for applications that would require detection of communities in real time, especially for massive networks. The Louvain method, which uses modularity maximization to detect clusters, is usually considered to be one of the fastest community detection algorithms even without any provable bound on its running time. We propose a novel graph traversal-based community detection framework, which not only runs faster than the Louvain method but also generates clusters of better quality for most of the benchmark datasets. We show that our algorithms run in

Detecting Community Structures in Social Networks by Graph Sparsification

O(|V| + |E|)