Telecommunication subscribers' churn prediction model using machine learning

With the ongoing economic growth and the rise of new companies, customer churn has become a renewed focus for many businesses, particularly in the telecommunications and SaaS sectors. Accurately identifying lost customers using artificial intelligence technology is crucial for helping companies develop effective strategies to retain these valuable clients. This paper provides a comprehensive overview of recent research on the application of various machine learning models, including logistic regression, random forests, decision trees, XGBoost, and advanced deep learning models like Deep Neural Networks (DNN) and Convolutional Neural Networks (CNN), to different types of datasets in the telecommunications and SaaS sectors. The study examines the workflow of each model, offering an in-depth analysis of their effectiveness and performance in predicting customer churn. In addition to evaluating the strengths and weaknesses of these models, the paper also addresses key challenges in machine learning, such as interpretability, applicability, and privacy concerns. It highlights current solutions employed by researchers, including SHapley Additive exPlanations (SHAP) for model interpretability, transfer learning, domain adaptation, and federated learning, to overcome these challenges. Furthermore, this paper emphasizes the ongoing need for further research to enhance the robustness and practical deployment of churn prediction models in these industries. In conclusion, this paper offers a comprehensive overview of current research on customer churn prediction models, highlighting the application of various machine learning techniques and addressing key challenges and solutions within the telecommunications and SaaS sectors.

Churn prediction is an important task for Customer Relationship Management (CRM) in telecommunication companies. Accurate churn prediction helps CRM in planning effective strategies to retain their valuable customers. However, churn prediction is a complex and challenging task. In this paper, a hybrid churn prediction model is proposed based on combining two approaches; Neighborhood Cleaning Rules (NCL) and Particle Swarm Optimization (PSO). NCL is applied in the preprocessing stage for handling the imbalanced churn data; and eliminating outliers and unrepresentative data. In the next stage, a Constricted PSO is applied for developing the final prediction model. The developed model is evaluated and compared with a baseline PSO model. The proposed hybrid model is compared also with Artificial Neural Networks (ANN) and Decision trees (DT) models which are traditional and common approaches used in the literature for churn prediction. The experimental results show that the proposed hybrid model outperforms the baseline PSO model, ANN and DT in terms of accuracy and actual churn rate.

The principle objective of this chapter is to build up a churn prediction model which helps telecom administrators to foresee clients who are no doubt liable to agitate. Many studies affirmed that AI innovation is profoundly effective to anticipate this circumstance as it is applied through training from past information. The prediction procedure is involved three primary stages: normalization of the data, then feature selection based on information gain, and finally, classification utilizing different AI methods, for example, back propagation neural network (BPNNM), naïve Bayesian, k-nearest neighborhood (KNN), support vector machine (SVM), discriminant analysis (DA), decision tree (DT), and extreme learning machine (ELM). It is shown from simulation study that out of these seven methods SVM with polynomial based kernel is coming about 91.33% of precision where ELM is at the primary situation with 92.10% of exactness and MLANN-based CCP model is at third rank with 90.4% of accuracy. Similar observation is noted for 10-fold cross validation also.

Incorporating textual information in customer churn prediction models based on a convolutional neural network

With increasing number of mobile operators and competitive market, churn prediction has become a major challenge for telecom companies. The telecom sector is highly effected due to rival competition and companies face loss of revenue when customer switch to competitor. For retaining the customer, an exact or accurate customer churn prediction model is required to estimate the number of customer who will churn out of current service provider which will be referred as churners. With the significant boom in machine learning research in recent years and having both power and high dimensional data, machine learning techniques can be used to predict those churners. This paper provides information on challenges and way of building the churn prediction model. This also provides a roadmap for the researchers to tackle supervised learning techniques and applying those to solve the real world problems.

In present business competition, organizations are emphasizing on optimal decision support systems (DSS) to enhance its growth-oriented decision support system process. To meet these requirements, the enterprise solutions facilitate certain robust and productive business intelligence (BI) applications. On the other hand, to retain market, organizations are emphasizing on service innovation discovery and its redesign (SIDRD) paradigm using BI applications. To accomplish SIDRD, DSS-oriented BI applications require huge feedback datasets to process where the data could be retrieved from certain feedback channel or data warehouses (DWs). The data warehouses encompass datasets from varied resources, and to ensure optimal data extraction and retrieval, these DWs need robust data mining schemes compatible with multidimensional data model (MDDM). On the other hand, to provide most precise and accurate DSS applications, the data security and its privacy preservation are inevitable. Considering all these requirements, in this paper, a robust privacy-preserved mining model called Commutative RSA (CRSA) has been developed and implemented with C5.0 decision tree algorithm to achieve SIDRD objectives with BI utilities. The developed paradigm has exhibited optimal performance for optimized accuracy, overheads, and secure rule set generation for BI. The performance for the developed SIDRD has been analyzed in terms of coverage accuracy and F1 scores, and the overall system has performed optimal.

Globally, enterprises are undergoing significant transformation in line with developments based on industrial revolution by leveraging extensive computing resources, data capture technologies, information processing systems, and advanced data science models that span analytics, optimization, and algorithmic intelligence. The global market's increasing demand, diverse resource supply options, global competition, and environmental protection needs are driving organizations to adopt sustainable strategies. These involve utilizing information technology and analytics more effectively, innovating manufacturing and service support systems, and employing novel problem-solving methods. The awareness of these challenges and opportunities inspired the theme of the joint event: 56th Annual Convention of ORSI (2023-ORSI) and the 10th International Conference on Business Analytics and Intelligence (2023-ICBAI), held at Indian Institute of Science, Bangalore, India, from December 18 to 20, 2023. The Operational Research Society of India (ORSI) Karnataka, the Department of Management Studies, IISc Bangalore, and the Analytics Society of India (ASI), DCAL, IIM Bangalore jointly organized this event. The joint event aimed to establish a premier platform for knowledge sharing among distinguished practitioners, academics, and researchers from industry and academia, focusing on the current applications of Operations Research (OR), Business Analytics (BA), and Business Intelligence (BI). The conference received over 655+ paper submissions, with 455 selected for presentation. 66 papers were deemed particularly interesting, and authors of 13 promising articles were invited to submit extended versions for a special issue of International Journal of Mathematical, Engineering and Management Sciences (IJMEMS). After rigorous peer review, eight papers were accepted for publication in this special issue, addressing common challenges in Operations Research, Business Analytics, and Business Intelligence. This Special Issue of the IJMEMS explores recent developments in Operations Research, Business Analytics, and Business Intelligence. It presents cutting-edge trends and substantial contributions to key areas such as Scheduling problems, Transshipment problems, E-commerce, Nanofluids, Blockchain Technology, Generative AI, Augmented Analytics, Machine Learning, and real-time Anomaly Detection. This Special Issue delves into following eight topics: Unmasking Content Clarity: Advancements in Defining, Measuring and Enhancing Readability: The authors present a novel method using natural language processing and Generative AI to quantitatively evaluate readability and comprehension. This approach surpasses traditional readability indices, offering substantial benefits for content creation and knowledge management in fields like education, business, technical support, and policy platforms. Strategic Insights into Blockchain Adoption in Automotive Supply Chains: A Comparative AHP-TOPSIS and TISM-MICMAC Analysis The authors explore blockchain adoption in the automotive industry using a multidisciplinary approach involving AHP, TOPSIS, TISM, and MICMAC analyses. This study identifies key enablers and their relationships, offering actionable insights and practical recommendations for automotive managers considering blockchain adoption. Avoid Maximum Cost Method for Solving Linear Fractional Transshipment Problem: The authors introduce a mathematical model for the linear fractional transshipment problem (LFTP) and suggest “Avoid Maximum Cost Method” to obtain an initial basic feasible solution for LFTP. This study conducts a comparative analysis with existing methods to demonstrate the efficiency of the proposed approach. Mathematical Study of Dispersion of Nano Biosensors in an Artery with Multiple Stenosis: This work examines nano-biosensors in a diseased artery with multiple stenoses, determining the temperature, velocity of nanofluid, and transport coefficients. The results lay the groundwork for developing nano-biosensors to diagnose, treat, and manage cardiovascular disease. The mathematical model had possible scope for target detection and drug delivery at stenosed sites. Integrating Generative AI in Business Intelligence: A Practical Framework for Enhancing Augmented Analytics: This study offers a practical framework for integrating generative AI (Gen AI) into Business Intelligence (BI). By adopting it, businesses can maximize GenAI and BI's potential, enhancing analytics, operations, and fostering a collaborative, data-driven culture. Data Monetization Through Cross Industry Collaboration in Retail Banking: This paper examines how data sharing between banks and e-commerce platforms, facilitated by data monetization, can improve banking customer experiences. This study proposes a framework using propensity models to identify promising customers and offer personalized products and promotions. Development of Dispatching Rule based Heuristic Algorithms for Real-Time Dynamic Scheduling of Non-identical Parallel Burn-in Ovens with Machine Eligibility Restriction: This study tackles a realistic problem in semiconductor manufacturing by scheduling non-identical parallel Burn-in ovens. The study proposes 25 heuristic algorithms for real-time dynamic scheduling with machine eligibility restrictions. Through empirical and statistical analysis, this study identifies top-performing algorithms. A Hybrid Framework for Real-Time Data Drift and Anomaly Identification Using Hierarchical Temporal Memory and Statistical Tests: This paper introduces a hybrid framework combining Hierarchical Temporal Memory and Sequential Probability Ratio Test for real-time data drift detection and anomaly identification. Retraining and false positives were minimized, outperforming traditional methods in experiments. Challenges and Future Directions 1. Enhanced Decision-Making: Addressing uncertainty and complexity in decision-making processes to improve outcomes. 2. Regulatory Frameworks: Conducting further research to establish flexible yet robust regulatory structures that can effectively adapt to the rapid evolution of blockchain technology. 3. Optimization Algorithms: Developing efficient algorithms to solve linear fractional transshipment problem with multi-objective linear fractional functions. 4. AI in Business Intelligence: Investigating long-term impacts of AI-enabled Business Intelligence tools on data-driven decision-making and organizational performance. 5. Ethical AI Considerations: Addressing ethical concerns such as data privacy, fairness, and biases in AI systems to ensure responsible use. 6. Digital Collaboration Frameworks: Developing frameworks that integrate digital footprints and cross-industry collaboration data to enhance strategic partnerships. 7. Advanced Scheduling Algorithms: Creating advanced meta-heuristic algorithms using efficient dispatching rule-based heuristics for dynamic scheduling of non-identical parallel batch processing machines with eligibility constraints. 8. Hybrid Anomaly Detection: Proposing a hybrid framework that combines multivariate extension of Hierarchical Temporal Memory with multivariate Sequential Probability Ratio Test for enhanced anomaly detection. 9. Advanced Text Generation: Utilizing advanced text generation techniques like prompt engineering and fine-tuning to produce more readable and engaging content. We extend our sincere appreciation to all contributing authors for their significant contributions and anonymous reviewers for their dedication and sincere evaluation of submissions. Their timely and excellent responses have been truly gratifying. Additionally, we would like to express our heartfelt thanks to Professor Mangey Ram, Editor-in-Chief of the International Journal of Mathematical, Engineering and Management Sciences, for his support in accepting this special issue and providing unwavering backing from its inception. Guest Editors

Companies are investing more in analytics to obtain a competitive edge in the market and decision makers are required better identification among their data to be able to interpret complex patterns more easily. Alluring thousands of new customers is worthless if an equal number is leaving. Business Intelligence (BI) systems are unable to find hidden churn patterns for the huge customer base. In this paper, a decision support system has been proposed, which can predict the churning behaviour of a customer efficiently. We have proposed a procedure to develop an analytical system using data mining as well as machine learning techniques C5, CHAID, QUEST, and ANN for the churn analysis and prediction for the telecommunication industry. Prediction performance can be significantly improved by using a large volume and several features from both Business Support Systems (BSS) and Operations Support Systems (OSS). Extensive experiments are performed; marginal increases in predictive performance can be seen by using a larger volume and multiple attributes from both Telco BSS and OSS data. From the results, it is observed that using a combination of techniques can help to figure out a better and precise churn prediction model.

Customer turnover is a crucial issue in banking since maintained profitability depends on keeping clients. This work aims to categorize consumer turnover in banks by using a new ensemble approach combining many machine learning methods, hence enhancing churn prediction models. Using a comprehensive dataset including demographic, financial, and behavioral data—such as credit score, account balance, tenure, and activity levels—the study employs the goal variable revealing if a customer has left the bank. The study starts with univariate, bivariate, and multivariate feature exploration and subsequently uses the Interquartile Range (IQR) approach to identify outliers thereby improving the data quality. Five models—K-Nearest Neighbors, Support Vector Classifier, Random Forest, Decision Tree, and XGBoost—a Voting Classifier ensemble—are used to estimate project churn. Building upon all the strengths of each model, this approach improves the prediction of classification and provides a balanced and highly robust classification system. The applied approaches are K-Nearest Neighbors (KNN), Support Vector Classifier (SVC), Random Forest, Decision Tree, and XGBoost within a Voting Classifier configuration. The performance of the Voting Classifier without SMOTE yields the following results: Accuracy: 0.87, precision: 0.87, recall: 0.80, and F1-Score: 0.87. The proposed model that extend the base model using SMOTE (Synthetic Minority Over-sampling Technique), yields a higher prediction accuracy of 0.90, precision of 0.90, recall of 0.90 and F1-Score of 0.90. This enhancement is proving the efficiency of SMOTE to handle the class imbalance problem in order to render the churn prediction more balanced and reliable system. The proposed approach assures a reliable solution to the strategies to retain the customers in the banking organisations.

Owing to saturated markets, fierce competition, dynamic criteria, along with introduction of new attractive offers, the considerable issue of customer churn was faced by the telecommunication industry. Thus, an efficient Churn Prediction (CP) model is required for monitoring customer churn. Therefore, this work proposes a novel framework to predict customer churn through a deep learning model namely Swish Recurrent Neural Network (S-RNN). Finally, SRNN is adapted to classify the Churn Customer (CC) and a normal customer. If the result is a churn customer, network utilisation history is analysed for retention process. Whereas, the number of churn customers based on the area network usage is not recognised in this frameworkOwing to saturated markets, fierce competition, dynamic criteria, along with introduction of new attractive offers, the considerable issue of customer churn was faced by the telecommunication industry. Thus, an efficient Churn Prediction (CP) model is required for monitoring customer churn. Therefore, this work proposes a novel framework to predict customer churn through a deep learning model namely Swish Recurrent Neural Network (S-RNN). Finally, S-RNN is adapted to classify the Churn Customer (CC) and a normal customer. If the result is a churn customer, network utilisation history is analysed for retention process.

CRM at a pay-TV company: Using analytical models to reduce customer attrition by targeted marketing for subscription services

School dropout in higher education remains a significant global challenge with profound socioeconomic consequences. To address this complex issue, educational institutions increasingly rely on business intelligence (BI) and related predictive analytics, such as machine learning and data mining techniques. This systematic review critically examines the application of BI and predictive analytics for analyzing and preventing student dropout, synthesizing evidence from 230 studies published globally between 1996 and 2025. We collected literature from the Google Scholar and Scopus databases using a comprehensive search strategy, incorporating keywords such as “business intelligence”, “machine learning”, and “big data”. The results highlight a wide range of predictive tools and methodologies, notably data visualization platforms (e.g., Power BI) and algorithms like decision trees, Random Forest, and logistic regression, demonstrating effectiveness in identifying dropout patterns and at-risk students. Common predictive variables included personal, socioeconomic, academic, institutional, and engagement-related factors, reflecting dropout’s multifaceted nature. Critical challenges identified include data privacy regulations (e.g., GDPR and FERPA), limited data integration capabilities, interpretability of advanced models, ethical considerations, and educators’ capacity to leverage BI effectively. Despite these challenges, BI applications significantly enhance institutions’ ability to predict dropout accurately and implement timely, targeted interventions. This review emphasizes the need for ongoing research on integrating ethical AI-driven analytics and scaling BI solutions across diverse educational contexts to reduce dropout rates effectively and sustainably.

Machine learning has become useful in daily life thanks to improvements in machine learning techniques. Text classification as an important part in machine learning. There are already many methods used for text classification such as Artificial Neural Network (ANN), Naïve Bayes, SVM, Decision Tree etc. ANN is a branch in machine learning which approximate the function of natural neural network. ANN have been used extensively for classification. In this research a simple architecture of ANN is used. But it needs to be pointed out that the architecture used in this research is relatively simple compared to the cutting edge in ANN development and research to show the potential that ANN have compared to other classification method. ANN, Naïve Bayes and SVM performance are measured using f1-macro. Performance of classification model is measured of multiple single-label and multi-label dataset. This research found that in single-label classification ANN have a comparable f1-macro with 0.79 compared to 0.82 for SVM. In multi-label classification ANN have the best f1-macro with 0.48 compared to 0.44 in SVM.

This paper proposes universal coarse-grained reconfigurable computing architecture for hardware implementation of decision trees (DTs), artificial neural networks (ANNs), and support vector machines (SVMs), suitable for both field programmable gate arrays (FPGA) and application specific integrated circuits (ASICs) implementation. Using this universal architecture, two versions of DTs (functional DT and axis-parallel DT), two versions of SVMs (with polynomial and radial kernel) and two versions of ANNs (multi layer perceptron ANN and radial basis ANN) machine learning classifiers, have been implemented in FPGA. Experimental results, based on 18 benchmark datasets of standard UCI machine learning repository database, show that FPGA implementation provides significant improvement (1–2 orders of magnitude) in the average instance classification time, in comparison with software implementations based on R project.

Text mining is a special case of data mining which explore unstructured or semi-structured text documents, to establish valuable patterns and rules that indicate trends and significant features about specific topics. Text mining has been in pattern recognition, predictive studies, sentiment analysis and statistical theories in many areas of research, medicine, financial analysis, social life analysis, and business intelligence. Text mining uses concept of natural language processing and machine learning. Machine learning algorithms have been used and reported to give great results, but their performance of machine learning algorithms is affected by factors such as dataset domain, number of classes, length of the corpus, and feature selection techniques used. Redundant attribute affects the performance of the classification algorithm, but this can be reduced by using different feature selection techniques and dimensionality reduction techniques. Feature selection is a data preprocessing step that chooses a subset of input variable while eliminating features with little or no predictive information. Feature selection techniques are Information gain, Term Frequency, Term Frequency-Inverse document frequency, Mutual Information, and Chi-Square, which can use a filters, wrappers, or embedded approaches. To get the most value from machine learning, pairing the best algorithms with the right tools and processes is necessary. Little research has been done on the effect of feature selection techniques on classification accuracy for pairing of these algorithms with the best feature selection techniques for optimal results. In this research, a text classification experiment was conducted using incident management dataset, where incidents were classified into their resolver groups. Support vector machine (SVM), K-Nearest Neighbors (KNN), Naïve Bayes (NB) and Decision tree (DT) machine learning algorithms were examined. Filtering approach was used on the feature selection techniques, with different ranking indices applied for optimal feature set and classification accuracy results analyzed. The classification accuracy results obtained using TF were, 88% for SVM, 70% for NB, 79% for Decision tree, and KNN had 55%, while Boolean registered 90%, 83%, 82% and 75%, for SVM, NB, DT, and KNN respectively. TF-IDF, had 91%, 83%, 76%, and 56% for SVM, NB, DT, and KNN respectively. The results showed that algorithm performance is affected by feature selection technique applied. SVM performed best, followed by DT, KNN and finally NB. In conclusion, presence of noisy data leads to poor learning performance and increases the computational time. The classifiers performed differently depending on the feature selection technique applied. For optimal results, the classifier that performed best together with the feature selection technique with the best feature subset should be applied for all types of data for accurate classification performance. Keywords: Text Classification, Supervised Machine Learning, Feature Selection DOI : 10.7176/JIEA/9-3-06 Publication date :May 31 st 2019