Using Machine Learning to Forecast User Satisfaction from Behavioural Data
Author(s): Khushbu Rahat & Hemant Sharma
Abstract - Traditional user satisfaction metrics, such as surveys and Net Promoter Scores, are inherently reactive and provide a limited, often delayed, view of user sentiment. This research explores a proactive paradigm by leveraging machine learning (ML) to forecast user satisfaction directly from behavioural data. We meticulously analysed diverse behavioural attributes, including clickstream patterns, session durations, error occurrences, and feature engagement. Utilising a suite of advanced classification and regression models, our study achieves high predictive accuracy, demonstrating the robust capability of ML to interpret subtle user signals. Crucially, it has identified and highlighted the most influential behavioural features, offering actionable insights for product teams. These findings underscore the transformative potential of ML-driven forecasting to enable timely, targeted interventions and significantly enhance the overall user experience.
Keywords - User Satisfaction, Machine Learning, Behavioural Data, Predictive Analytics, User Experience (UX), Customer Relationship Management (CRM).
DOI URL: https://doi.org/10.26761/ijrls.11.3.2025.1910
Cite This Article As: Rahat, K. & Sharma, H. (2025) Using Machine Learning to Forecast User Satisfaction from Behavioural Data. International Journal of Research in Library Science (IJRLS), 11(3) 1-9. www.ijrls.in
Copyright © 2025 Author(s) retains the copyright of this article. This article is published under the terms of the Creative Commons Attribution License 4.0.
Paper ID: IJRLS-1910 Page: 1-9 Publication Date: 12 July 2025