A New Method for Automating Voice Calls Routing Using Data Preprocessing Techniques for Supervised Learning

Abstract

In spite of the massive technological development in the telecommunications field globally, to transit and terminate international phone calls that are originated from one country to another is a process that requires much time and resources, which poses a great challenge. This paper proposes a new method for automated voice calls routing which automates the selection process for the appropriate carrier/channel of calls by creating a class label for quality statistics using data preprocessing for supervised learning while ensuring it complies with the requirements and conditions of the operator with its targeted quality of services. Data preprocessing tasks were performed on international calls duration records to automate the routing decision-making process in terms of sorting and prioritizing the routes of carriers who deliver the calls to the rightful destination. The methods used in the research were data cleaning to get rid of the noisy data, manual fill-in for the missing values, aggregation for data transformation, attribute subset selection in data reduction and finally creating a class label to determine the quality of each call and prepare the preprocessed data for supervised learning. The experimental results proved the efficiency of the proposed method and solved a great challenge that is facing telecom operators by saving operational costs and time.