Asee peer logo

Neural Network Processing Using Database Management Systems

Download Paper |


2001 Annual Conference


Albuquerque, New Mexico

Publication Date

June 24, 2001

Start Date

June 24, 2001

End Date

June 27, 2001



Page Count


Page Numbers

6.747.1 - 6.747.6

Permanent URL

Download Count


Request a correction

Paper Authors

author page

Bruce Segee

author page

Michael D. Amos

Download Paper |

NOTE: The first page of text has been automatically extracted and included below in lieu of an abstract

Session 2358

Neural Network Processing using Database Management Systems Michael Amos, Dr. Bruce Segee

University of Maine Department of Electrical and Computer Engineering Instrumentation Research Laboratory


Database Management Systems (DBMS) have become an integral part of any data storage, processing, or retrieval system. They are uniquely suited to manipulate large amounts of data while maintaining a level of data integrity and security. Meanwhile, Neural Networks are being used to perform operations on data to discover the unknown underlying input/output relationships. A significant part of Neural Network processing is reducing the input space to a manageable size. These algorithms utilize some of the same techniques that databases use (such as hashing functions) to retrieve or store large amounts of data quickly. This project utilizes a popular DBMS to build a Neural Network application that resides inside the database. This will yield many benefits that weren’t previously attainable, e.g., the data and the Neural Network are both located in the same physical location. There is no need to export the data from the database, manipulate it into an appropriate format, and then use a separate Neural Network application to process the data. Another benefit is that there is no need for database users to learn external applications to process the data from their database. They simply issue an appropriate SQL statement and the Neural Network will perform the calculations and give a result. If the data is changed, the user simply re-issues the SQL statement to recalculate the outputs of the system. Utilizing this system, database users will be able to perform high level data processing tasks using the tools with which they are familiar, while appreciating significant performance gains over other implementations.

I. Introduction

Problem Description

Neural Networks have been a staple of the engineering community for decades6. They are useful for many engineering applications where an adaptive learning tool proves useful, such as classification and approximation of unknown functions. Unfortunately, most applications that perform Neural Network calculations are cryptic at best and are only understood by someone with years of experience and an engineering degree. If Neural Networks were able to be used by someone with little or no engineering expertise, then more people could use these powerful tools.

Since many Neural Network applications are written for a specific purpose, they expect a certain format for input data. The application is designed for that particular input format and if that format happens to change, then the application will more than likely need major changes as well. Of course, this is quite difficult if the person that wrote the original application is not still part of the group needing these modifications. Even if the person is still available to make

Segee, B., & Amos, M. D. (2001, June), Neural Network Processing Using Database Management Systems Paper presented at 2001 Annual Conference, Albuquerque, New Mexico.

ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2001 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015