Retention of customers is essential in any industry but a vital factor in telecommunications. It is required due to increased competition and the availability of competitors in the industry. Identifying customers unsatisfied with the current services and assisting them by creating new offers is an essential function of organizations. Acquiring a new customer costs more than retaining an old customer. Maintaining an existing customer is more important than customer acquisition. Customer churn/attrition is one of the most critical metrics for a growing business to evaluate. It measures the customer that leaves the company/service. Customer churn has drawn high business attention, particularly in the telecoms sector. The churn prediction models have been developed, which are heavily based on data mining principles and use machine learning and meta-heuristic algorithms. This paper examines some of the most significant churn prediction approaches created in recent years with the support of Machine Learning (ML) and Deep Learning (DL) techniques. The effective preprocessing methods employed by various researchers are studied and combined to create a preprocessing workflow for processing the data for the ML algorithms. These models aim to predict customer preference in the telecom sector.