Hanseung Lee

Temporal Event Sequence Simplification

Electronic Health Records (EHRs) have emerged as a cost-effective data source for conducting medical research. The difficulty in using EHRs for research purposes, however, is that both patient selection and record analysis must be conducted across very large, and typically very noisy datasets. Our previous work introduced EventFlow, a visualization tool that transforms an entire dataset of temporal event records into an aggregated display, allowing researchers to analyze population-level patterns and trends. As datasets become larger and more varied, however, it becomes increasingly difficult to provide a succinct, summarizing display. This paper presents a series of user-driven data simplifications that allow researchers to pare event records down to their core elements. Furthermore, we present a novel metric for measuring visual complexity, and a language for codifying disjoint strategies into an overarching simplification framework. These simplifications were used by real-world researchers to gain new and valuable insights from initially overwhelming datasets.

iVisClustering: An Interactive Visual Document Clustering via Topic Modeling

Clustering plays an important role in many large‐scale data analyses providing users with an overall understanding of their data. Nonetheless, clustering is not an easy task due to noisy features and outliers existing in the data, and thus the clustering results obtained from automatic algorithms often do not make clear sense. To remedy this problem, automatic clustering should be complemented with interactive visualization strategies. This paper proposes an interactive visual analytics system for document clustering, called iVisClustering, based on a widely‐used topic modeling method, latent Dirichlet allocation (LDA). iVisClustering provides a summary of each cluster in terms of its most representative keywords and visualizes soft clustering results in parallel coordinates. The main view of the system provides a 2D plot that visualizes cluster similarities and the relation among data items with a graph‐based representation. iVisClustering provides several other views, which contain useful interaction methods. With help of these visualization modules, we can interactively refine the clustering results in various ways. Keywords can be adjusted so that they characterize each cluster better. In addition, our system can filter out noisy data and re‐cluster the data accordingly. Cluster hierarchy can be constructed using a tree structure and for this purpose, the system supports cluster‐level interactions such as sub‐clustering, removing unimportant clusters, merging the clusters that have similar meanings, and moving certain clusters to any other node in the tree structure. Furthermore, the system provides document‐level interactions such as moving mis‐clustered documents to another cluster and removing useless documents. Finally, we present how interactive clustering is performed via iVisClustering by using real‐world document data sets.

Trust management and adversary detection for delay tolerant networks

Delay Tolerant Networks (DTNs) have been identified as one of the key areas in the field of wireless communications. They are characterized by large end-to-end communication latency and the lack of end-to-end path from a source to its destination. These characteristics pose several challenges to the security of DTNs. Especially, Byzantine attacks give serious damages to the network in terms of latency and data availability. Using reputation-based trust management systems is shown to be an effective way to handle the adversarial behavior in Mobile Ad-Hoc Networks (MANETs). However, because of the unique characteristics of DTNs, the techniques to build a trust mechanism for MANETs do not apply to DTNs. Our main objective in this paper is to develop a robust trust mechanism and an efficient and low cost malicious node detection technique for DTNs. Inspired by our recent results on reputation management for online systems and e-commerce, we developed an iterative malicious node detection mechanism for DTNs which is far more effective than existing techniques. Our results indicate the proposed scheme provides high data availability and packet-delivery ratio with low latency in DTNs under adversary attacks.

An iterative algorithm for trust and reputation management

Trust and reputation play critical roles in most environments wherein entities participate in various transactions and protocols among each other. The recipient of the service has no choice but to rely on the reputation of the service provider based on the latters prior performance. This paper introduces an iterative method for trust and reputation management referred as ITRM. The proposed algorithm can be applied to centralized schemes, in which a central authority collects the reports and forms the reputations of the service providers as well as report/rating trustworthiness of the (service) consumers. The proposed iterative algorithm is inspired by the iterative decoding of low-density parity-check codes over bipartite graphs. The scheme is robust in filtering out the peers who provide unreliable ratings. We provide a detailed evaluation of ITRM via analysis and computer simulations. Further, comparison of ITRM with some well-known reputation management techniques (e.g., Averaging Scheme, Bayesian Approach and Cluster Filtering) indicates the superiority of our scheme both in terms of robustness against attacks (e.g., ballot-stuffing, bad-mouthing) and efficiency. Furthermore, we show that the computational complexity of the proposed ITRM is far less than the Cluster Filtering; which has the closest performance (to ITRM) in terms of resiliency to attacks. Specifically, the complexity of ITRM is linear in the number of clients, while that of the Cluster Filtering is quadratic.

An interactive visual testbed system for dimension reduction and clustering of large-scale high-dimensional data

Many of the modern data sets such as text and image data can be represented in high-dimensional vector spaces and have benefited from computational methods that utilize advanced computational methods. Visual analytics approaches have contributed greatly to data understanding and analysis due to their capability of leveraging humans’ ability for quick visual perception. However, visual analytics targeting large-scale data such as text and image data has been challenging due to the limited screen space in terms of both the numbers of data points and features to represent. Among various computational methods supporting visual analytics, dimension reduction and clustering have played essential roles by reducing these numbers in an intelligent way to visually manageable sizes. Given numerous dimension reduction and clustering methods available, however, the decision on the choice of algorithms and their parameters becomes difficult. In this paper, we present an interactive visual testbed system for dimension reduction and clustering in a large-scale high-dimensional data analysis. The testbed system enables users to apply various dimension reduction and clustering methods with different settings, visually compare the results from different algorithmic methods to obtain rich knowledge for the data and tasks at hand, and eventually choose the most appropriate path for a collection of algorithms and parameters. Using various data sets such as documents, images, and others that are already encoded in vectors, we demonstrate how the testbed system can support these tasks.

VisIRR: Visual analytics for information retrieval and recommendation with large-scale document data

We present VisIRR, an interactive visual information retrieval and recommendation system for large-scale document data. Starting with a query, VisIRR visualizes the retrieved documents in a scatter plot along with their topic summary. Next, based on interactive personalized preference feedback on the documents, VisIRR collects and visualizes potentially relevant documents out of the entire corpus so that an integrated analysis of both retrieved and recommended documents can be performed seamlessly.

PIVE: Per-Iteration visualization environment for supporting real-time interactions with computational methods

A main bottleneck in integrating computational methods with visual analytics is their significant computational cost, which hinders real-time interactive visualization with them. To solve this, we present PIVE (Per-Iteration Visualization Environment), which visualizes intermediate results from algorithm iterations, thus allowing users to efficiently perform multiple interactions in real time.

Data ingestion and evidence marshalling in Jigsaw VAST 2010 Mini Challenge 1 award: Good support for data ingest

This article describes the sense-making process we applied to solve the VAST 2010 Mini Challenge 1 using the visual analytics system Jigsaw. We focus on Jigsaws data ingest and evidence marshalling features and discuss how they are beneficial for a holistic sense-making experience.

GeneTracer: Gene sequence analysis of disease mutations VAST 2010 mini challenge 3 award: Excellent process explanation

Our visual analytics tool GeneTracer, developed for the VAST 2010 genetic sequence mini challenge, visualizes gene sequences of current outbreaks and native sequences along with disease characteristics. We successfully used GeneTracer in combination with data mining techniques to solve the challenge.

VisIRR: A Visual Analytics System for Information Retrieval and Recommendation for Large-Scale Document Data

In this article, we present an interactive visual information retrieval and recommendation system, called VisIRR, for large-scale document discovery. VisIRR effectively combines the paradigms of (1) a passive pull through query processes for retrieval and (2) an active push that recommends items of potential interest to users based on their preferences. Equipped with an efficient dynamic query interface against a large-scale corpus, VisIRR organizes the retrieved documents into high-level topics and visualizes them in a 2D space, representing the relationships among the topics along with their keyword summary. In addition, based on interactive personalized preference feedback with regard to documents, VisIRR provides document recommendations from the entire corpus, which are beyond the retrieved sets. Such recommended documents are visualized in the same space as the retrieved documents, so that users can seamlessly analyze both existing and newly recommended ones. This article presents novel computational methods, which make these integrated representations and fast interactions possible for a large-scale document corpus. We illustrate how the system works by providing detailed usage scenarios. Additionally, we present preliminary user study results for evaluating the effectiveness of the system.