Student Presentations at the Creative Collaborations

Thursday, May 10, 2012

Creative Collaborations took place on April 19, 2012 in the Hahn University Center. Eight of our students and five faculty mentors showcased their collaborative research results in poster presentations described below. Congratulations to our student participants!

Student Presenter: Cameron Askew
Faculty Advisor: Eric Jiang, PhD
Title: A Study on How to Approach Data Mining Problems
Discipline: Computer Science
Data mining is a new and growing practice of Computer Science which helps discover previously unknown patterns from large data sets using a combination of artificial intelligence, statistics and database management. Due to the wide variety of data mining problems, there are few standards defining how to approach a problem. This project considers how to appropriately construct a data mining solution for a specific problem. Previously proven algorithms including random forests, neural networks, multiple regressions, and gradient boosting will be tested to evaluate their effectiveness in different scenarios. More refined techniques may be tested for less common problems. As a result, we will have a better understanding of why different techniques are more appropriate for different problems.

Student Presenter: Jun Ryan Delacruz
Faculty Advisor: Drew Talley, PhD and Simon Koo, PhD
Title: Biodex: The Index of Life
Discipline: Marine Biology and Computer Science
Field guides have a long history of usefulness to biologists as a way to record and view data about specific organisms. The goal of this project was to develop an interactive field guide for mobile devices to provide a cleaner and more efficient means of recording and viewing entries. To accomplish this goal, the author created Biodex, an Objective-C based mobile app, and used MySQL and PHP to manage the organism database on the back-end. Biodex will contain an interactive field guide for organism querying, as well as several additional features such as a data entry form, a documents viewer, and a map that displays the user’s current location. This app will be tested by the USD Marine Biology department during their field studies. It is anticipated that Biodex can be utilized by other field sciences.

Student Presenter: Theresa Chadwick
Faculty Advisor: Lynn McGrath, PhD and Jane Friedman, PhD
Title: Investigating Communication Networks via Expander Graphs
Discipline: Mathematics
A graph G can be used to model of a communications network. The vertices of a graph can represent entities that communicate with each other and edges represent lines of direct communication. H(G) (the expansion constant of the graph) is one way to measure how good the graph is a communication network, taking into account speed and cost. We investigate the expansion constant for 2-bicycle graphs.

Student Presenter: Megan Mahoney
Faculty Advisor: Luby Liao, PhD
Title: MegaCMS: A Better Way to Create, Manage and Customize Websites
Discipline: Computer Science
A web content management system (CMS) is a powerful tool to help users with little or no knowledge of web programming create and manage their website with ease. We have observed through substantial use of popular CMSs that many of them are not customizable and as a result they fail to meet customer needs. These CMSs lack the ability to create a completely custom theme, and the learning curve for many is steep. Thus we have designed and built megaCMS, a robust CMS, customizable to various user needs, such as the ability to manage documents, layouts, themes, and widgets while maintaining a user-friendly interface. Through the use of PHP and MySQL, this CMS allows customers to easily create the website they have always dreamed of.

Student Presenter: Michael Nichols
Faculty Advisor: Simon Koo, PhD
Title: Zoo Inspector for iPad
Discipline: Computer Science
This project is a collaborative effort on behalf of the San Diego Zoo and the University of San Diego to develop an iPad app to help exceed the national requirements for animal habitats and enclosures. The objective is to provide a simple and efficient way to fill out and submit inspection forms to the proper administrators. This will minimize any additional administrative effort that would normally be involved in a paper based system. The app also allows for the ability to easily append information to a form, such as images, specific location information, and additional free form notes. The app has been designed to integrate into the current work flow. It is suggested that the San Diego Zoo invest in their mobile strategy, such as wireless infrastructure, security, and content management, in order to realize the full potential of the system.

Student Presenters: Nicholas Nobles and Ryan Yandel
Faculty Advisor: Eric Jiang, PhD
Title: A Clustering Based Approach to Network Intrusion Detection
Discipline: Computer Science
With the rapid growth of online connectivity, sophisticated network intrusion detection techniques are absolutely crucial to maintaining privacy and security. Network connections have data associated with them that can be used to model the legitimate and intrusive behaviors. In this project, we apply clustering techniques to detect intrusive network events that deviate sufficiently from the expected legitimate behavior. Presumably, the network data that reflect legitimate connections share some common attributes which also applies to the data with intrusions. Thus, clustering can produce clusters or groups of data instances that contain either mostly valid connections or mostly intrusive connections. Our clustering based intrusion
detection approach builds a boundary, based on a given set of network connection data, between legitimate and intrusive activities, and can be used to monitor and analyze new network data and determine the legitimacy of the associated connections.

Student Presenter: David Yandell
Faculty Advisor: Simon Koo, PhD
Title: Wise Gremlin: Algorithmic Stock Trader
Discipline: Computer Science
In 2009, the TABB Group estimated that annual aggregate profits of high-frequency arbitrage strategies exceeded $21 billion. In light of this, Wise Gremlin was created as a high-frequency stock trading software package that focuses on the implementation of an effective statistical arbitrage algorithm. The software continuously gathers real-time price information, makes financial decisions, and updates visual graphics for any number of stocks. The strength of
Wise Gremlin is a simple, yet powerful decision-making algorithm that takes advantage of predictable, temporary fluctuations in data rather than using news analysis or traditional long-term investment plans. Performance is measured by the return on investments (ROI) on a
simulated portfolio using real, live data.


Contact Us Email

Phone: (619) 260-4600
Fax: (619) 260-4162

Visit Campus Map

Founders Hall
5998 Alcalá Park
San Diego, CA 92110