Enhancing revenue generation in Bangladesh’s FinTech sector: a comprehensive analysis of real-time predictive customer behavior modeling in AWS using a hybrid OptiBoost-EnsembleX model
Abstract Customer behavior holds pivotal significance within the fintech industry, both in offline and online domains, influencing revenue generation. The application of data analytics to scrutinize customer behavior is a critical factor in optimizing financial outcomes. The anticipation of future c...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2025-06-01
|
| Series: | Journal of Electrical Systems and Information Technology |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s43067-025-00209-w |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Customer behavior holds pivotal significance within the fintech industry, both in offline and online domains, influencing revenue generation. The application of data analytics to scrutinize customer behavior is a critical factor in optimizing financial outcomes. The anticipation of future customer conduct is a cornerstone for resource allocation in sales and marketing, enabling strategic decision-making in manufacturing operations, inventory planning, and point-of-sale scenarios. The intricate nature of customer behavior analysis necessitates innovative methodologies. This study introduces a real-time predictive model: OptiBoost-EnsembleX (Optuna-tuned CatBoost and LightGBM classifiers in a soft-voting ensemble framework) integrating data analytics and unsupervised machine learning techniques to discern and understand customer conduct. The investigation utilized a range of machine learning algorithms, such as random forest, support vector machine (SVM), XGBoost, CatBoost, and LightGBM, to develop models. This employs a unique dataset that consists of 10,000 examples of customer behavior. This investigation culminates in identifying CatBoost as the model that demonstrates the highest accuracy in predicting customer behavior. The selected model incorporated a real-time web application, representing a practical manifestation of the proposed solution. The seamless integration of the developed model into a machine-learning pipeline hosted on Amazon EC2 servers ensures its deployment in a production environment. This investigation makes a substantial contribution to the fintech industry by introducing a comprehensive and efficient method for analyzing customer behavior in real time, which has implications for improving decision-making and optimizing operations. |
|---|---|
| ISSN: | 2314-7172 |