Bai Wang

Overlapping Community Detection in Bipartite Networks

Researches have discovered that rich interactions among entities in nature and human society bring about complex networks with community structures. In this paper, we propose a novel algorithm BiTector (bi-community detector) to mine the overlapping communities in large-scale sparse bipartite networks. We apply the algorithm to various real-world datasets, showing that BiTector can identify the overlapping community structures in the bipartite networks efficiently and effectively.

Cell phone mini challenge award: Social network accuracy— exploring temporal communication in mobile call graphs

In the mobile call mini challenge of VAST 2008 contest, we explored the temporal communication patterns of Catalano/Vidro social network which is reflected in the mobile call data. We focus on detecting the hierarchy of the social network and try to get the important actors in it. We present our tools and methods in this summary. By using the visual analytic approaches, we can find out not only the temporal communication patterns in the social network but also the hierarchy of it.

Link Prediction Based on Local Information

Link prediction in complex networks is an important issue in graph mining. It aims at estimating the likelihood of the existence of links between nodes by the know network structure information. Currently, most link prediction algorithms based on local information consider only the individual characteristics of common neighbors. In this paper, first, we study the link prediction results as the change of the exponent on the degree of common neighbors, and find some regular pattern between different networks and different exponent. After that, we come up with a new algorithm exploiting the interactions between common neighbors, namely Individual Attraction Index. To reduce the time complexity, we design a simple edition, called Simple Individual Attraction Index. We compare nine well-known local information metrics on eight real networks. The result proves well the best overall performance of these two new algorithms.

Backbone Discovery in Social Networks

Recent years have seen a thriving development of the World Wide Web as the most visible social media which enables people to share opinions, experiences and expertise with each other across the world. People now get involved in many different social networks simultaneously, which are often large intricate web of connections among the massive entities they are made of. As a result, the challenge of collecting and analyzing large-scale data among social members has left most basic questions about the global composition and function of such networks largely unresolved: What is the essential organization of a social network? who are the influential individuals whose voice is echoed by others? To address these questions, this paper presents an algorithm called sketcher to discover and describe the overall backbone of a specific network. Experimental results on the American College Football, Scientific Collaboration, and Telecommunications Call networks show that sketcher can extract the essential composition of a social network both efficiently and intuitively.

Detecting Link Communities in Massive Networks

Most of the existing literature which has entirely focused on clustering nodes in large-scale networks. To discover multi-scale overlapping communities quickly, we propose a highly efficient multi-resolution link community detection algorithm to detect the link communities in massive networks based on the idea of edge labeling. First, we will get the node partition of the network based on a new multi-resolution node detection algorithm. After that, we can find the link community in a linear time by the labels of nodes. Its time complexity is near linear and its space complexity is linear. The effectiveness of our algorithm is demonstrated by extensive experiments on lots of computer generated artificial graphs and real-world networks. The results show that our algorithm is very fast and highly reliable. Tests on real and artificial networks also give excellent results comparing with the newly proposed link partition algorithm.

Detecting Communities in Massive Networks Based on Local Community Attractive Force Optimization

Currently, community detection has led to a huge interest in data analysis on real-world networks. However, the high computationally demanding of most community detection algorithms limits their applications. In this paper, we propose a heuristic algorithm to extract the community structure in large networks based on local community attractive force optimization whose time complexity is near linear and space complexity is linear. The effectiveness of our algorithm is demonstrated by extensive experiments on lots of computer generated graphs and public available real-world graphs. The result shows our algorithm is extremely fast, and it is easy for us to explore massive networks interactively.

Visual Analysis of a Co-authorship Network and Its Underlying Structure

An interesting property of network is that the information is not only contained in the entities, but also in the links between them. As the structure of the co-authorship network can greatly influence its function and reflect how the internal information is exchanged. We attempt to get deep insight of the features in a co-authorship network at a university. This is done by the following two steps. First, we will explore the basic statistical properties of the co-authorship network and try to present these properties by different graph drawing techniques. Second, to gain more insight of the co-authorship network, we will use a community detecting algorithm to find the clusters of the network. By filtering the unstable links and detecting the clusters, we find there are many stable links in these clusters and many unstable links take the role of bridges between these clusters. By labeling these communities with different department names, we can get an overview of the main research fields of the university and their relations.

Community Ranking in Social Network

Social network is one of the most important true-life networks in our real world scenarios. A typical feature of the social network is the dense sub-structure (quasi-clique or community) which is essential for understanding the networks internal structure and function. Traditional social network analysis usually focuses on the centrality and power of a single individual or entity, however, in peoples daily life, a group or an organization often holds a more influential position and plays a more important role. Therefore, in this paper, we first present a parallel algorithm for the detection of quasi-cliques, and then we describe the techniques that are useful for evaluating the centrality and significance of a quasi-clique. Computational results on a real call graph from a telecom career and a collaboration network of co-authors are given in the end.

Random Walk Based Resource Allocation: Predicting and Recommending Links in Cross-Operator Mobile Communication Networks

Link prediction is an important issue in social network analysis. It aims at estimating the likelihood of the existence of links between nodes by the known network information. Although this problem has been extensively studied, many significant factors about how to calculate the similarity between two nodes remains unexplored and largely open, especially for the mobile phone networks. We are first to take the link prediction task for missing calls in cross-operator communication networks. This task is of interest from not only a purely scientific point of view but also an excellent scenario for the marketing of telecom operators. We develop several algorithms based on a physical concept: resource allocation that treats all nodes in a network differently by allocating resource to each one. We use the call attributes of each node as resources which are owned by themselves. The process of resource allocation is guided by weighted random walk which naturally combines the information from network structure with call attributes of nodes and edges. Our experiments on the cross-operator networks show that our newly proposed algorithms outperform the other state-of-the-art unsupervised approaches as well as the supervised HPLP method.

VisCRM: A social network visual analytic tool to enhance customer relationship management

As acquiring and retaining the most profitable customers are challenging tasks of service providers, various CRM tools are used to support these processes. Traditional CRM methods focus on various customer profitability models in different scenarios based on their past profit contribution. Social network analysis provides a natural way to understand the relationships between customers; however, this method has seldom been used in customer relationship management. In this paper, we propose a visual analytic tool-VisCRM to analyze the graph features in customer relationship network. Based on the Visual Analytics Mantra, users can get the properties of the whole network as well as the communication patterns of certain customer with VisCRM. Furthermore, customers ranking scores could be got with different statistical algorithms, and a combined ranking score could also be got for each customer by setting different weights to these properties. After choosing a valuable customer, in order to get a deeper insight into customers communication patterns, people can explore the features of certain customers egocentric social network visually. To evaluate the effect of VisCRM, we will explore different mobile call networks in telecom services in the case studies.