Tools
Tools
Nawanjana Basnayake, Jayani and Homayounvala, Elaheh (2026) Integrating sentiment analysis of employee reviews and organisational factors for employee retention risk analysis. In: 14th International Conference on Frontiers of Intelligent Computing: Theory and Applications (FICTA-2026), 8-9 June 2026, London (UK) / Online. (In Press)
Organisations focus on improving employee retention strategies by using employee feedback and its emotional context. However, the large volume of unstructured employee feedback makes this a major challenge for organisations. Many studies have examined sentiment analysis of textual data to compare models and sentiment classification performance, with limited attention given to how to use the results to improve employee retention policies and support evidence-based decision-making. To address this gap, this study uses sentiment analysis to identify negative employee feedback and maps it to predefined organisational dimensions to derive both retention risk levels and risk themes. The RoBERTa model achieved approximately 75% of accuracy. The risk categorisation revealed that work-life balance is the most critical risk factor, and the warehouse department has the highest turnover risk. These insights were delivered as actionable results via an interactive dashboard to support improved decision-making around the organisation's employee retention. Overall, the study shows that integrating sentiment analysis results with risk categorisation helps organisations identify key retention issues and take more targeted action.
![[thumbnail of 260_CRC.pdf]](https://repository.londonmet.ac.uk/style/images/fileicons/text.png "260_CRC.pdf")
Restricted to Repository staff only until 13 May 2027.
Download (820kB) | Request a copy
 |
View Item |