Computer Science Faculty Publications

A Linear Transform Scheme for Building Weighted Scoring Rules

Hu Tianming et al. — Fri, 01 Mar 2013 12:35:40 PST

It is often necessary to combine multiple scores into a joint decision in many real-world applications. To that end, an essential challenge is to build a proper weighted scoring rule, which assigns more weight to the more important scores and derives a final scoring value. In this paper, we first introduce some desirable properties of weighted scoring rules, and then propose a scheme that can incorporate weights into scores in a natural way. In addition, we show our method has overcome some drawbacks of a classic weighted scoring method. Finally, as a case study, we apply these weighted scoring methods to outlier detection to evaluate how much they help solve the ultimate ranking problem.

Visual Analysis of a Cardiovascular System Based on ECG and ABP Signals Using Evolvable Hardware Design

Fan Xiong et al. — Fri, 01 Mar 2013 12:35:38 PST

In this paper, visual data analysis was applied to raw medical data using probability theory to provide valuable information for preliminary diagnosis. The evolvable hardware design approach combined with information theory was applied to model an adaptive cardiovascular system. The cardiovascular system is modelled by a digital logic circuit based on ECG and ABP signal samples as input and output respectively. In our experiments, five patients' ECG and ABP data was chosen for the visual analysis. A user friendly GUI was demoed and the correlation of patient data was analyzed in the space and time domain. The digital circuit model was extrinsically evolved using genetic programming as the evolutionary algorithm and mutual information as the fitness function. In our experiments using MATLAB, we demonstrated that the data analysis could provide valuable information for preliminary diagnosis, and the proposed method could fit the input-output relationship as recorded samples piece-wise in which each piece contains monotonic input data. The model we proposed is a self reconfigurable digital circuit model based on input and output information. It's safe to conclude that the model is adaptive to changes based on different patient's unique ECG and ABP signals since the I/O information is also changed. Furthermore, a "divide and conquer" method was employed to get a more accurate piece-wise model. Experimental results show that the method is feasible, scalable, and promising as a personalized medical simulation tool.

MIning Spatial Colocation Patterns: A Different Framework

Jin Soung Yoo et al. — Fri, 01 Mar 2013 12:35:37 PST

Recently, there has been considerable interest in mining spatial colocation patterns from large spatial datasets. Spatial colocation patterns represent the subsets of spatial events whose instances are often located in close geographic proximity. Most studies of spatial colocation mining require the specification of two parameter constraints to find interesting colocation patterns. One is a minimum prevalent threshold of colocations, and the other is a distance threshold to define spatial neighborhood. However, it is difficult for users to decide appropriate threshold values without prior knowledge of their task-specific spatial data. In this paper, we propose a different framework for spatial colocation pattern mining. To remove the first constraint, we propose the problem of finding N-most prevalent colocated event sets, where N is the desired number of colocated event sets with the highest interest measure values per each pattern size. We developed two alternative algorithms for mining the N-most patterns. They reduce candidate events effectively and use a filter-and-refine strategy for efficiently finding colocation instances from a spatial dataset. We prove the algorithms are correct and complete in finding the N-most prevalent colocation patterns. For the second constraint, a distance threshold for spatial neighborhood determination, we present various methods to estimate appropriate distance bounds from user input data. The result can help an user to set a distance for a conceptualization of spatial neighborhood. Our experimental results with real and synthetic datasets show that our algorithmic design is computationally effective in finding the N-most prevalent colocation patterns. The discovered patterns were different depending on the distance threshold, which shows that it is important to select appropriate neighbor distances.

3D Visualization in Elementary Education Astronomy: Teaching Urban Second Graders about the Sun, Earth, and Moon

Zeynep Isik-Ercan et al. — Tue, 07 Aug 2012 13:50:39 PDT

This research-in-progress hypothesizes that urban second graders can have an early understanding about the shape of Sun, Moon, and Earth, how day and night happens, and how Moon appears to change its shape by using three dimensional stereoscopic vision. The 3D stereoscopic vision system might be an effective way to teach subjects like astronomy that explains relationships among objects in space. Currently, Indiana state standards for science teaching do not suggest the teaching of these astronomical concepts explicitly before fourth grade. Yet, we expect our findings to indicate that students can learn these concepts earlier in their educational lives with the implementation of such technologies. We also project that these technologies could revolutionize when these concepts could be taught to children and expand the ways we think about children’s cognitive capacities in understanding scientific concepts.

Exploring Digital Libraries through Visual Interfaces

Beomjin Kim et al. — Tue, 07 Aug 2012 13:50:38 PDT

Programming Embedded Computing Systems Using Static Embedded SQL

Lubomir Stanchev — Tue, 07 Aug 2012 13:50:36 PDT

In this paper we show how parts of the software of an embedded system can be written in a declarative programming language, such as SQL. To facilitate this process, we propose the following techniques: (1) We limit the input query language to queries for which we can guarantee efficient execution, (2) All data operations are pre-compiled in order to avoid performing expensive query optimization during runtime, (3) Only the data that is required for efficient execution of the predefined operations is stored on the embedded device, and (4) All search structures are implemented using a novel physical design structure that reduces the need for storing redundant data. We experimentally validate our approach by implementing five popular UNIX commands of the MINIX operating system using static embedded SQL.

Synergistic Validation Methodology for Knowledge-based Engineering Systems

U John Tanik — Tue, 07 Aug 2012 13:50:34 PDT

A methodology for validation is provided for the architecture of Knowledge based Engineering (KBE) applications. A standard approach to validation is introduced, followed by the Synergistic Validation Methodology (SVM), which addresses four key aspects of KBE system validation: a) software architecture, b) product design process, c) artificial intelligence, and d) knowledge acquisition. The synergistic methodology is introduced as a comprehensive approach to validation of Knowledge-based engineering systems. This SVM approach is shown to be divided into four target divisions addressing multiple areas each, which correspond to well-defined methodologies in terms of general standards, techniques, and methods utilized. The five-tier pre-validation and post-validation process is intended to verify and validate reconfigurable, modular, and scalable KBE systems.

Temporal Data Mining: Similarity-Profiled Association Pattern

Jin Soung Yoo — Tue, 07 Aug 2012 13:50:32 PDT

Using a Website and Wiki as a Support tool for Teaching Assistants

Alexander Ganago et al. — Mon, 06 Feb 2012 05:53:24 PST

Visual Search Interfaces on Digital Libraries

Beomjin Kim et al. — Mon, 06 Feb 2012 05:53:23 PST

Mining Spatial Colocation Patterns: A Different Framework

Jin Soung Yoo et al. — Mon, 06 Feb 2012 05:53:22 PST

Bag Relational Algebra with Grouping and Aggregation over C-Tables with Linear Conditions

Lubomir Stanchev — Mon, 21 Feb 2011 06:54:30 PST

Efficient Access to Non-Sequential Elements of a Search Tree

Lubomir Stanchev — Mon, 21 Feb 2011 06:54:29 PST

This article describes how a search tree can be extended in order to allow efficient access to predefined subsets of the stored elements. This is achieved by marking some of the elements of the search tree with marker bits. We show that our approach does not affect the asymptotic logarithmic complexity for existing operations. At the same time, it is beneficial because the modified search tree can now efficiently support requests on predefined subsets of the search elements that it previously could not.

Saving Space and Time Using Index Merging

Lubomir Stanchev et al. — Mon, 21 Feb 2011 06:54:29 PST

The Promise of 3D Visualization: Urban Second Graders Exploring the Sun, Earth, and Moon

Z Isik-Ercan et al. — Mon, 21 Feb 2011 06:54:28 PST

On Solving Some Heterogeneous Problems of Healthcare Information Sharing and Interoperability Using Ontology Computing

C S. Wei et al. — Mon, 21 Feb 2011 06:54:26 PST

Computer Fluency and the National Academies' Report

Kurt F. Lauckner et al. — Tue, 25 May 2010 08:09:46 PDT

Panel discussion report.

On Approaches to Congestion Control Over Wireless Networks

David Q. Liu et al. — Fri, 09 Apr 2010 07:25:55 PDT

Similarity-Profiled Temporal Association Mining

Jin Soung Yoo et al. — Fri, 09 Apr 2010 07:25:54 PDT

Systematic Component-Oriented Development with Axiomatic Design

Cengiz Togay et al. — Fri, 09 Apr 2010 07:25:53 PDT