Utilizing Data Mining Technology to Identify the Causes of Mobile Call Drops.

ABSTRACT

Mobile call drop occurs when calls are abruptly disconnected before the conversation is completed. This research aims to design a predictive model that can determine the occurrence of mobile call drops using ethio telecom mobile network data. To overcome the limitations of simple statistical methods, data mining techniques, methods, and methodologies were proposed and employed in this study. Approximately 20,000 records of Fault Management data collected over one year and six months were selected for analysis. After eliminating irrelevant and unnecessary data, a total of 16,996 datasets with 8 attributes were used for the study. Data preprocessing was performed to clean and prepare the datasets for data mining tasks. The study utilized WEKA software version 3.8.1 and four classification techniques, namely J48 and Random Forest algorithms from decision trees, as well as PART and JRIP algorithms from rule induction. Among these techniques, the J48 decision tree algorithm with 10-fold cross-validation demonstrated the best performance and processing speed, achieving an accuracy of 95.43% in 0.06 seconds. This algorithm utilized 8 attributes for experimentation in the research. One of the major challenges during the study was the lack of related works on telecommunication mobile call drop analysis. Additionally, the FMS mobile network server could only store data for up to two years, limiting access to data beyond that period. The research concludes by recommending ethio telecom apply data mining techniques on mobile network data to effectively identify the reasons for call drops. By doing so, the telecom operator can take proactive measures to improve service quality and reduce call drop incidents, leading to enhanced customer satisfaction and a more reliable communication experience

Utilizing Data Mining Technology to Identify the Causes of Mobile Call Drops. GET MORE COMPUTER SCIENCE PROJECT TOPICS AND MATERIALS