APPLICATION OF NEURAL NETWORK TECHNIQUE TO TELECOMMUNICATION FAULT DOCKET SYSTEM

¹Okpeki U.K and ²Adebari F.A

¹Department of Electrical/Electronics/Computer Engineering, Delta State University, Oleh

 ²Department of Computer Engineering, Yaba College of Technology, Yaba Lagos

email: omakazeem@yahoo.com

ABSTRACT

This paper concerns the application of neural network technique to automatic telecommunication fault docket system. Neural networks have a unique set of characteristics. They are not programmed; instead they are trained by being repeatedly shown large numbers of examples for the problem under consideration. As a result of this, they can provide good results in relatively short timescales but for only certain types of problem, and then a great deal of care is taken over the collection of the data, pre-processing of the data and the design of the network. In this research the feed-forward neural network strategy was adopted with mapping innovation called Kohonen self organizing mapping. It was implemented using Microsoft visual basic 6.0 as front end and Microsoft access 2000 as the back end. The results shows that, by observing great care in data collection, processing and network design, faults are diagnosed and solutions automatically proffered within a short time. This approach results in significant cost saving by improving the availability and efficiency of telecommunication systems.

Keywords: service records, customer service database, intelligent fault docket system, neural process & data mining.