The Dynamics of Actor Loyalty to Groups in Affiliation Networks.

Can Friends Be Trusted%3F Exploring Privacy in Online Social Networks.

Generating Abstract Networks Using Multi%2Drelational Biological Data.

Privately detecting bursts in streaming%2C distributed time series data.

Privacy%2DPreserving Collaborative Social Networks.

Structure%2DBased Hierarchical Transformations for Interactive Visual Exploration of Social Networks.

Software Engineering for Web Services Workflow Systems.

Privacy Preserving Burst Detection of Distributed Time Series Data Using Linear Transforms.
Abstract — In this paper, we consider burst detection within the context of privacy. In our scenario, multiple parties want to detect a burst in aggregated time series data, but none of the parties want to disclose their individual data. Our approach calculates bursts directly from linear transform coefficients using a cumulative sum calculation. In order to reduce the chance of a privacy breech, we present multiple data perturbation strategies and compare the varying degrees of privacy preserved. Our strategies do not share raw time series data and still detect significant bursts. We empirically demonstrate this using both real and synthetic distributed data sets. When evaluating both privacy guarantees and burst detection accuracy, we find that our percentage thresholding heuristic maintains a high degree of privacy while accurately identifying bursts of varying widths. I.
Measuring Topological Anonymity in Social Networks.
While privacy preservation of data mining approaches has been an important topic for a number of years, privacy of social network data is a relatively new area of interest. Previous research has shown that anonymization alone may not be sufficient for hiding identity information on certain real world data sets. In this paper, we focus on understanding the impact of network topology and node substructure on the level of anonymity present in the network. We present a new measure, topological anonymity, that quantifies the amount of privacy preserved in different topological structures. The measure uses a combination of known social network metrics and attempts to identify when node and edge inference SMPCes arise in these graphs. 1
Measuring Topological Anonymity in Social Networks.
While privacy preservation of data mining approaches has been an important topic for a number of years, privacy of social network data is a relatively new area of interest. Previous research has shown that anonymization alone may not be sufficient for hiding identity information on certain real world data sets. In this paper, we focus on understanding the impact of network topology and node substructure on the level of anonymity present in the network. We present a new measure, topological anonymity, that quantifies the amount of privacy preserved in different topological structures. The measure uses a combination of known social network metrics and attempts to identify when node and edge inference SMPCes arise in these graphs. 1
Detecting Aggregate Bursts from Scaled Bins within the Context of Privacy.
Abstract — In this paper, we consider burst detection within the context of privacy. In our scenario, multiple parties want to detect a burst in aggregated time series data, but none of the parties want to disclose their individual data. We introduce two data perturbation approaches that alter the local data so that raw time series data values are not shared and bursts can be identified using a Shewhart threshold. The first involves lossy data compression via windowing. Unfortunately, windowing alone does not guarantee enough privacy because the envelope of the time series can still be determined. Therefore, we introduce a second data perturbation approach that employs scaled binning. This method transmits values for each data point based on the distance of the data point to a local mean of the time series. The strength of this approach is its increased privacy. We empirically demonstrate the burst detection results using both real and synthetic distributed data sets. When attempting to optimize both privacy guarantees and burst detection accuracy, we find that a combined approach using both windowing and scaled binning balances burst accuracy and privacy better than either approach individually. I.
Detecting Aggregate Bursts from Scaled Bins within the Context of Privacy.
Abstract — In this paper, we consider burst detection within the context of privacy. In our scenario, multiple parties want to detect a burst in aggregated time series data, but none of the parties want to disclose their individual data. We introduce two data perturbation approaches that alter the local data so that raw time series data values are not shared and bursts can be identified using a Shewhart threshold. The first involves lossy data compression via windowing. Unfortunately, windowing alone does not guarantee enough privacy because the envelope of the time series can still be determined. Therefore, we introduce a second data perturbation approach that employs scaled binning. This method transmits values for each data point based on the distance of the data point to a local mean of the time series. The strength of this approach is its increased privacy. We empirically demonstrate the burst detection results using both real and synthetic distributed data sets. When attempting to optimize both privacy guarantees and burst detection accuracy, we find that a combined approach using both windowing and scaled binning balances burst accuracy and privacy better than either approach individually. I.
Visual Mining of Multi%2DModal Social Networks at Different Abstraction Levels.
Social networks continue to become more and more feature rich. Using local and global structural properties and descriptive attributes are necessary for more sophisticated social network analysis and support for visual mining tasks. While a number of visualization tools for social network applications have been developed, most of them are limited to uni-modal graph representations. Some of the tools support a wide range of visualization options, including interactive views. Others have better support for calculating structural graph properties such as the density of the graph or deploying traditional statistical social network analysis. We present Invenio, a new tool for visual mining of socials. Invenio integrates a wide range of interactive visualization options from Prefuse, with graph mining algorithm support from JUNG. While the integration expands the breadth of functionality within the core engine of the tool, our goal is to interactively explore multi-modal, multi-relational social networks. Invenio also supports construction of views using both database operations and basic graph mining operations. 1
Increasing the Predictive Power of Affiliation Networks.

Visual analysis of dynamic group membership in temporal social networks.
C-Group is a tool for analyzing dynamic group membership in temporal social networks over time. Unlike most network visualization tools, which show the group structure within an entire network, or the group membership for a single actor, C-Group allows users to focus their analysis on a pair of individuals. While C-Group allows for viewing the addition and deletion of nodes (actors) and edges (relationships) over time, its major contribution is its focus on changing group memberships over time. By doing so, users can investigate the context of temporal group memberships for the pair. C-Group provides users with a flexible interface for defining (and redefining) groups interactively, and supports two novel visual representations of the evolving group memberships. This flexibility gives users alternate views that are appropriate for different network sizes and provides users with different insights into the grouping behavior. We demonstrate the utility of the tool on a scientific publication network.
Pruning Social Networks Using Structural Properties and Descriptive Attributes.
Scale is often an issue with understanding and making sense of large social networks. Here we investigate methods for pruning social networks by determining the most relevant relationships. We measure importance in terms of predictive accuracy on a set of target attributes of the social network. Our goal is to create a pruned network that models only the most informative affiliations and relationships. We present methods for pruning networks based on both structural properties and descriptive attributes demonstrate it on a network of NASDAQ and NYSE businesses and on a bibliographic network. 1
An Algorithm for Constrained Association Rule Mining in Semi%2Dstructured Data.
. The need for sophisticated analysis of textual documents is becoming more apparent as data is being placed on the Web and digital libraries are surfacing. This paper presents an algorithm for generating constrained association rules from textual documents. The user specifies a set of constraints, concepts and/or structured values. Our algorithm creates matrices and lists based on these prespecified constraints and uses them to generate large itemsets. Because these matrices are small and sparse, we are able to quickly generate higher order large itemsets. Further, since we maintain concept relationship information in a concept library, we can also generate rulesets involving concepts related to the initial set of constraints. 1 Introduction Finding global patterns or generalizations can provide insight into subsets of a textual data set. However, because of the large number of patterns that typically exist in a textual database, preconstraining the data space and searching for target...
A Robust System Architecture for Mining Semi%2DStructured Data.
The value of extracting knowledge from semi-structured data is readily apparent with the explosion of the WWW and the advent of digital libraries. This paper proposes a versatile system architecture for text mining that maintains structured data components in a relational database and unstructured concepts in a concept library. After a detailed explanation of our system architecture, we briefly describe IRIS, our prototype rule generation system
Generating Association Rules from Semi%2DStructured Documents Using an Extended Concept Hierarchy.
Most data mining research has focused on generating rules within databases containing structured values while essentially ignoring the potentially valuable information that exists in the unstructured blocks of text. This paper suggests an approach for generating association rules that relates structured data values to concepts extracted from unstructured data. Our approach involves the use of an extended concept hierarchy (ECH) to maintain parent, child, and sibling relationships between concepts. This structure allows us to generate rules that relate a given concept in the ECH and a given structured attribute value to the neighbors of the given concept in the ECH. We also describe an efficient implementation of the ECH that keeps track of concepts and pointers to documents associated with them. Experimental results on documents from the ABI/Inform Information Retrieval System are presented. 1
