Unsupervised Grouping of Configuration Management Parameters in Network Elements : Using Association Rule Mining

Abstract

Over the past few decades, the widespread adoption of communication devices in everyday life has greatly expanded the reach and complexity of telecommunication networks. This rapid expansion has also brought significant challenges in the deployment and maintenance of wireless systems. One key challenge is the identification and management of appropriate configuration parameters, particularly Logical Network Cells usually referred to as LNCELs. As the number of these cells continues to grow, configuring each network element has become increasingly time-consuming and costly.

At the same time, notable progress in data analysis, data mining, and data modeling has opened new opportunities for innovation in the telecommunications field. In line with these advancements, this study explores the use of association rule mining techniques on configuration management parameter data of LNCELs in Multi Radio Base Transceiver Stations usually referred to as MRBTS. The objective is to transform the large configuration management parameters dataset into smaller, more manageable configuration sets, thereby reducing complexity in LNCEL configuration management and supporting more efficient and reliable wireless network operations.