Predicting Quality of Service of ethiotelecom GSM Mobile Network using Machine Learning algorithms

No Thumbnail Available

Date

2023-01

Journal Title

Journal ISSN

Volume Title

Publisher

Addis Ababa University

Abstract

Global System for Mobile Communication (GSM) is globally accepted stand ard for di gital mobile communications. Ethiotel ecom is one of the oldest telecom serv ice providers in Africa, which offer telecommunication services in Ethiopia. GSM ce llular mobile ervice is one of the various telecom services, which has millions of customers in Ethiopia. However, ethiotelecom has done many remarkable works in deve loping information and communications technology network, customers still have been complaining about the poor quality in GSM ce llul ar mobile service. The prime objective of this study is building a pred ictive model using machine-l ea rning techniques to determine the quality of service of GSM network for ethiotelecom Addis Ababa region, which helps to optimize quali ty of service in the area. The data used in this study was obtained from Ethiopian Communications Authority qua lity of service department. For the aim of constructing the machine learning model , a total of 2294 data sets with 6 attributes are employed before preprocessing. After co mpil ati on of the primary dataset preprocessing task is undertaken to make su itab le for the ML task li ke cleaning and attribute selection. Strictl y, following the experimental resea rch process, various experiments are conducted using python as a tool. This is done to find out the best model that clas i fies the KP I data by applyi ng the best classification models by comparing the performance of the models developed using KNN, SVM, as well as the Logistic regress ion learning methods. According to experimental results, logistic regressio n classification algo rithm outperfo rms the I other two classification algorithms with an accuracy of 99.854%. The fi nding indicates Call Setup Success Rate, Handover success rate, Dropped call rate and call attempt are the major determinant factors of QoS of GSM Network. The study also indicates that a GSM mobile network located in Addis Ababa and its surrounding are susceptible to fa ilure in network quality. Finall y, the study is limited to build a predictive machine-learning model for class ificati on of the dataset into the right level of QoS data. Through the res ults found in this study, we recommend ethiotelecom to implement such mobile network quality prediction techniques and avoid inegularities throughout the network that will improve customer satisfaction. Key words: Quality of service, Machine learning, Mobile network, GSM, Key Peljorman ce Indicators.

Description

Keywords

Quality of service, Machine learning, Mobile network, GSM, Key Peljorman ce Indicators.

Citation