INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 10, ISSUE 01, JANUARY 2021 ISSN 2277-8616 388 IJSTR©2021 www.ijstr.org Customer Churn Prediction in Telecom Sector: A Survey and way a head Ibrahim AlShourbaji, Na Helian, Yi Sun, Mohammed Alhameed Abstract—The telecommunication (telecom)industry is a highly technological domain has rapidly developed over the previous decades as a result of the commercial success in mobile communication and the internet. Due to the strong competition in the telecom industry market, companies use a business strategy to better understand their customers’ needs and measure their satisfaction. This helps telecom companies to improve their retention power and reduces the probability to churn. Knowing the reasons behind customer churn and the use of Machine Learning (ML) approaches for analyzing customers' information can be of great value for churn management. This paper aims to study the importance of Customer Churn Prediction (CCP) and recent research in the field of CCP. Challenges and open issues that need further research and development to CCP in the telecom sector are explored. Index Terms— Customer churn, prediction, machine learning, churn management, telecom ———————————————————— 1 INTRODUCTION Customers seek good service quality and competitive pricing factors in telecom sector. However, when these factors are missing, then they can easily leave to another competitor in the market [1]. This has led telecom to offer some incentives to customers to encourage them to stay [2]. The movement of customers (i.e., subscribers) from one service provider or carrier to another is called customer churn, It has been recognized that long-standing consumers are more lucrative in the long term, as new clients are engrossed by persuasive offers and incline to switch to an alternative competitor in the market at the moment they obtain a better concession [3-7], and therefore it is vital for companies to consider churn management as a part of Companies use CRM as a strategy to modify their process management, to improve their revenues and to find new approaches by primarily focusing on customers’ needs to avoid losing them rather than a product [13, 14]. These specifics have led competitive companies to capitalize on CRM to up-hold their customers, and thus helping to increase customer strength. Figure 1 shows the main sections of CRM [15].  Collaborative CRM: It aims to establish customized relationships with customers using several ways such as emails, telephone, websites, call centers, face-to-face contact, etc.  Operative CRM: This type offers services for the organizations to increase the efficiency of CRM processes  Analytical CRM: focuses on data collection and analysis to help the management build strategic decisions and plan for the future. Fig 1.CRM areas The data of customers are stored in such CRM systems which can then be transformed into valuable information with the help of ML techniques which aid telecom companies to formulate new polices, develop campaigns for existing clients and figure out the main reasons behind customer churn. In this way, companies can easily observe their customer’s behavior from time to time and manage them effectively. Therefore, ML approaches are needed in telecom sectors which remain the corner-stone of customer churn control and can play a fundamental role in decreasing the probability of churners. Due to the increased amount of data collection, organization and companies can store vast amount of data and information using several types of storage technologies at low cost. However, the challenge is to analyze, summarize and discover knowledge from these stored data. ML and statistics aiming at automatically discovering useful information and identifying hidden patterns in large data warehouses. ML involves few phases from raw data collection to some of the interesting patterns and this process includes data cleaning, transformation, selection and evaluation It has been reported that attracting new customers’ in the telecom sector, costs six times more than retaining existing ————————————————  Ibrahim AlShourbaji, Na Helian and Yi Sun, School of Computer Science, University of Hertfordshire, Hatfield, U.K,  Ibrahim AlShourbaji, Department of Computer and Network Engineering, Jazan University, 82822-6649 Jazan, Kingdom of Saudi  Mohammed Alhameed, Department of Computer Science, Jazan University, 82822-6649 Jazan, Kingdom of Saudi Arabia  Email: alshourbajiibrahim@gmail.com