Digital Repository

SentiChurn - Integrating Sentiment Analysis and Ensemble Based Churn Prediction to create a Retention System.

Show simple item record

dc.contributor.author Nawaz, Ammar
dc.date.accessioned 2025-06-18T03:57:52Z
dc.date.available 2025-06-18T03:57:52Z
dc.date.issued 2024
dc.identifier.citation Nawaz, Ammar (2024) SentiChurn - Integrating Sentiment Analysis and Ensemble Based Churn Prediction to create a Retention System. BSc. Dissertation, Informatics Institute of Technology en_US
dc.identifier.issn 2019675
dc.identifier.uri http://dlib.iit.ac.lk/xmlui/handle/123456789/2631
dc.description.abstract "In the ever-changing landscape of customer relationship management, successfully anticipating and mitigating customer turnover is crucial. This work aims to improve the performance and qualitative insights of customer churn prediction by combining sentiment analysis modeling via the Keras functional API and a Churn Prediction system based on an ensemble approach. The holistic research technique begins with the systematic collection of consumer feedback, then sentiment analysis is used to extract nuanced insights from textual data. Initiating the research journey, the first phase introduces deep learning-based sentiment analysis as a potent tool for sentiment modeling. This technique leverages neural networks to capture intricate patterns and nuances within the collected feedback, offering a nuanced portrayal of customer sentiments. By employing deep learning models, the research aims to extract rich semantic representations from the text data, enabling a detailed understanding of the emotional tone within customer comments. This approach facilitates the detection of subtle nuances and complexities in sentiment expression, which may not be adequately captured by conventional methods. Subsequently, the research advances to a novel phase, leveraging ensemble learning techniques to enhance the customer churn prediction model. This approach integrates multiple diverse predictors in a synergistic manner, harnessing the collective intelligence of multiple algorithms. By incorporating ensemble learning, the model aims to achieve superior predictive performance by aggregating the strengths of individual classifiers and mitigating their weaknesses. Beyond the quest for heightened prediction accuracy, this research aspires to furnish organizations with a profound understanding of the multifaceted factors influencing customer decisions. The proposed integrated approach empowers businesses to proactively identify potential churners collectively, facilitating the implementation of targeted retention strategies. Through this comprehensive strategy, organizations can cultivate customer product loyalty, optimize resource allocation, and fortify their competitive edge in the market, aligning with the evolving demands of customer-centric industries." en_US
dc.language.iso en en_US
dc.subject Sentiment Analysis en_US
dc.subject Churn Prediction en_US
dc.subject Ensemble Learning en_US
dc.title SentiChurn - Integrating Sentiment Analysis and Ensemble Based Churn Prediction to create a Retention System. en_US
dc.type Thesis en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search


Advanced Search

Browse

My Account