IS&T Insider:
Tell us about yourself.
Dr. Zhu:
I am a Professor and Chairperson of the Computer Science department. I received my Ph.D. in Computer and Systems Engineering from the Rensselaer Polytechnic Institute (RPI) in Troy, New York in 1986. I have been at UNO for 19 years.
IS&T Insider:
Did you have a specific mentor who has influenced you in this field?
My Ph.D. advisor, Dr. Herbert Freeman, an IEEE fellow who is well-known for his pioneer work in and contribution to pattern recognition and computer engineering, had great influence for me to be in my research field.
IS&T Insider:
Why did you choose the area of data mining?
The primary area of my research is Pattern recognition, which is fundamental to a number of research areas such as computer vision, machine intelligence, automation, robotics, and data mining. There are many data mining problems where pattern recognition techniques are needed to solve them. Many practical data mining problems are just the problem of pattern recognition.
With the ever-growing World Wide Web, there is an overwhelming amount of information available to users. The Web is, in essence, a giant data repository. This information is typically spread across different websites and may exist in various structured, semi-structured, or unstructured formats. The need to collect domain-specific information presents a formidable problem and calls for information retrieval systems which can be customized to user preferences. The concept of “What You Get Is What You Need (WYGIWYN)”, which we coined to describe our research on intelligent information services, is designed to provide users with an information set that is individually selected and organized, less redundant, more adaptable to the tasks in hand, and less confusing. .
IS&T Insider:
Briefly describe research accomplishments and papers you published and journals, if related to data mining, can you tell us briefly what that fields relates to?
I have been awarded three U.S. patents, with another one pending. My significant publications within the last five years include two papers published in IEEE Transactions on Systems, Man, and Cybernetics titled “Hierarchical Kohonen Net for Anomaly Detection in Network Security,” in April 2005, and “Min-Max Hyper-Ellipsoidal Clustering for Anomaly Detection in Network Security,” in August 2006. In fact, these two papers present a sequence of research in applying pattern recognition techniques to detect Internet intrusions. These are also examples of how data mining techniques can be applied to help solve network security problems. A number of researchers from worldwide have contacted me for potential collaborations in this field and in the specific subject area after reading these papers, some whom I am currently working with.
IS&T Insider:
Do you have any grand project that you are working on right now?
Our research group is concentrating on the research and development of concepts, methodologies, and software systems for the acquisition, modeling, representation, and management of knowledge critical to situational awareness and decision making. The group focuses on applying the Knowledge Engineering (KE) techniques to Web Intelligence (WI) tasks for efficient information retrieval, analysis, mining, and effective utilization. The theme of our projects, called WYGIWYN What You Get Is What You Need aims at providing web users ontime, onpurpose, and ontarget information services. Many challenges exist in defining the type of information, interinformation relationships, and the correlations between information object attributes so as to meet the needs of a specific user. One such challenge is the creation of a common format that allows the data to be easily analyzed and shared with other systems. In addition, the information retrieval system can be automated to support an ongoing process of updating the knowledge base, based on user preferences. There is a growing and recognized need for a mechanism by which the information structure can be stored in a way as specified by user defined ontology. Ontology Based Information Systems can clearly identify the semantics of data located within databases, text, or web pages. By adjusting the ontology after each information retrieval process, the user, or ontology engineer, can incrementally build and refine their domain specific ontology. As a result, an ontology based knowledge portal is created in a standardized format that facilitates greater interoperability with other internet-based applications. To assist in ontology building operations, we are developing a TIDS (Tailored Information Delivery Services) system to provide user selected information and reduce the amount of irrelevant information. Our TIDS system is based on semantic matching of concepts selected by users that are compared against the open source data tagged by using the same concepts. These concepts are part of a shared common ontology that can be extended through a collaborative effort of all users. For more information on the research, please visit the website: kewi.unomaha.edu.
IS&T Insider: What are your greatest accomplishments?
They are yet to come.
-Dr. Qiuming Zhu, CS Chairperson
