Two SMU computer science students, John Forrest and Vladimir Jovanovic, represented SMU at the first Texas Undergraduate Research Day on February 14, 2011, at the state Capitol building.
More than 140 undergraduate students from 51 colleges and universities made poster presentations at the event, which was designed “to showcase the experiences of undergraduate students engaged in research for Texas legislators and the public.” During the day, state legislators and other officials visited the poster session.
Both Forrest and Jovanovic have been supported by a National Science Foundation Research Experiences for Undergraduates grant awarded to Professor Margaret Dunham and Visiting Assistant Professor Michael Hahsler in the department of Computer Science and Engineering at the Lyle School of Engineering.
Forrest is a senior majoring in computer science who will graduate in May with the first departmental distinction granted by the department. He has accepted a position at Microsoft after graduation.
Forrest is working with Hahsler, his adviser, in the area of data stream clustering. Data streams constantly produce typically high-dimensional data that need to be processed in real-time. Many modern applications generate such data streams (for example, web click-stream data, computer network monitoring data, telecommunication connection data, readings from sensor nets and stock quotes).
The title of Forrest’s research is “ClusterDS: A Framework For Clustering Data Streams in R.” R is a free programming environment widely used for data analysis and data mining. Although a commercial high-performance version of R is available, currently no standard exists to process data streams. By creating a flexible framework to represent data streams and data stream mining algorithms, Forrest hopes to make a significant contribution to the data mining community.
Jovanovic is a junior majoring in both computer science and psychology in Dedman College. He also has been supported by an SMU Undergraduate Research Award from Associate Physics Professor Robert Kehoe in Dedman College. Jovanovic currently is working on a research project through the Innovation Competition at the Caruth Institute for Engineering Education in bioinformatics under the direction of Dunham.
The title of Jovanovic’s research is “Unique Base Pair Count Vectors in Analysis of Nucleotide Sequences.” Secondary and tertiary structures in single-stranded nucleotide sequences are important for DNA and RNA sequences when binding to particular molecules. Jovanovic has developed a novel method of base pair counting that can take into account these structures. The base pair counter looks at complementary base pairings of a folding nucleotide sequence.
This method thus creates a unique vector for each nucleotide sequence that leads to a simple but effective summarization of the possible secondary and tertiary structures that the sequence can have. The method can be used in a variety of applications in genetics and bioinformatics and has been shown to be useful in at least one case of aptamer classification. Specifically, using the novel method with data mining classifiers, a truth positive rate of 94.4 percent to 100 percent was achieved, depending on the classifier used.