Enhanced Customer Behaviour Analysis using Stacking Classifiers for Churn Prediction

Palathinkal Shibu, Nayana (2024) Enhanced Customer Behaviour Analysis using Stacking Classifiers for Churn Prediction. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (845kB) | Preview
Preview	PDF (Configuration Manual) Download (635kB) | Preview

Abstract

Customer behaviour is defined as how individuals or customers make decisions regarding the use and purchase of products or any other services. This study aims to predict customers’ behaviour and to use machine learning algorithms like Logistic Regression, AdaBoost Classifier, XGBoost Classifier and Stacking Classifier, to analyze their performance on the dataset used. Among the first three models, the XGBoost Classifier stood high in terms of accuracy and therefore required a more detailed analysis using hyperparameter tuning and GridSearchCV was applied to XGBoost Classifier. Among all the proposed algorithms, the Stacking Classifier developed using three base models, namely optimized XGBoost Classifier, Logistic Regression and AdaBoost Classifier, with XGBoost Classifier as the meta-classifier worked the best with an accuracy of 88% to predict customer behaviour. The performance of this system was further evaluated by comparing it with another methodology which used the dataset, “Ecommerce Customer Churn Analysis and Prediction”. When the models were retrained on that same set of data that was used in the referenced study, the Stacking Classifier achieved precisely 99% accuracy, surpassing the 98% of the referenced study while the XGBoost achieved an average accuracy of 93%, with an even better 97% after cross-validation and 98% after hyperparameter tuning compared to 91% in that study. These results confirm the effectiveness of the proposed system, showing how it can be generalized across datasets and how it outperforms other models.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Cosgrave, Noel UNSPECIFIED
Uncontrolled Keywords:	Customer Behaviour; Machine Learning; Predictive Modelling; Customer Churn; Hyperparameter Tuning; Stacking Ensemble Model
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science H Social Sciences > HF Commerce > Marketing > Consumer Behaviour Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	04 Sep 2025 08:28
Last Modified:	04 Sep 2025 08:28
URI:	https://norma.ncirl.ie/id/eprint/8764

Actions (login required)

View Item