Online data mining using evolving neural networks

Author

Nadeem Ahmed

Abstract

"The primary aim of this research is to develop an intelligent system for online data mining for classification problems. A smart system is developed by employing techniques such as neural networks and genetic algorithms. Evolving neural network architecture previously developed in the Smart Engineering Systems Laboratory at the University of Missouri--Rolla is used for developing this system. The architecture uses genetic algorithms for automatic generation of neural network, feature selection and aggregating the neural networks. Genetic algorithms are also used to optimize the neural network weights. Ensembles of classifiers are used to have diverse networks to handle noise and improve the performance of the system. The proposed system can be used as a real-time system, which typically learns from the existing data. As new information is available over a period of time, the system learns new information in addition to the previous information, thus tuning itself depending on how the system is used. The system was tested on various data sets like heart, liver-disorder and vehicle, and produced better results than traditional neural networks and evolving neural networks in batch training in which all the training data is presented at the same time. The system can be used for both offline and online training. A GUI is developed for ease of application. The main purpose of this research is to develop a smart real-time system that can adapt to a non-stationary environment on its own ability. The system can be used in many applications where user interaction always exists, like market basket analysis, intelligent user interface, web and intranet management, network management in telecommunications and process control in manufacturing"--Abstract, page iii.

Advisor(s)

Dagli, Cihan H., 1949-

Committee Member(s)

Enke, David Lee, 1965-
St. Clair, Daniel C.

Department(s)

Engineering Management and Systems Engineering

Degree Name

M.S. in Engineering Management

Publisher

University of Missouri--Rolla

Publication Date

Spring 2004

Pagination

viii, 55 pages

Note about bibliography

Includes bibliographical references (pages 53-54).

Rights

© 2004 Nadeem Ahmed, All rights reserved.

Document Type

Thesis - Restricted Access

File Type

text

Language

English

Subject Headings

Data mining
Neural networks (Computer science)

Thesis Number

T 8487

Print OCLC #

56479046

Link to Catalog Record

Electronic access to the full-text of this document is restricted to Missouri S&T users. Otherwise, request this publication directly from Missouri S&T Library or contact your local library.

Recommended Citation

Ahmed, Nadeem, "Online data mining using evolving neural networks" (2004). Masters Theses. 2499.
https://scholarsmine.mst.edu/masters_theses/2499