2026-04-17T17:46:53.897Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:10123 2025-02-06T16:46:32Z 3

Employee reviews sentiment classification using BERT encoding and AdaBoost classifier tuned by modified PSO algorithm 2024 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/10123 https://www.atlantis-press.com/proceedings/iciitb-24/126002410 V. Markovic A. Njegus D. Bulaja T. Zivkovic M. Zivkovic J. P. Mani N. Bacanin Sentiment analysis of the employee reviews is very important to understand the satisfaction in the company, predict the engagement of the employees, identify the risk of employee retention and improve general productivity of the company. Proper analysis of these reviews may provide valuable insight into the satisfaction and moral levels among employees, and identify the potential areas where improvement is possible. Moreover, employee analysis can help in detecting the risks of employee retention and drop in satisfaction within the company prior to their escalation. Companies can then intervene to mitigate identified problems, and boost morale among employees. This manuscript suggests application of the AdaBoost classification model to execute the classification of the employee reviews sentiment. To select the appropriate values of the AdaBoost hyperparameters, an enhanced version of the particle swarm optimization algorithm was developed and applied. The simulation results were put into comparisons to the outcomes achieved by several contenting potent optimizers. The overall findings suggest that the presented model obtained accuracy of 87.2%. was superior to other regarded methods, showing considerable potential for further applications in this domain. bookPart Atlantis Press 22 37 10.2991/978-94-6463-482-2_3 Proceedings of the 2nd International Conference on Innovation in Information Technology and Business (ICIITB 2024), Chapter in Advances in Computer Science Research