The purpose of this study is to use an optimal hybrid ML model with oversampling techniques (SMOTE) and feature selection techniques (SA) to help predict which employees may churn. Is to investigate. It is integrated with the classification algorithm. KNN, Naive Bayes, MLP, LR, etc. The focus is on the true positive accuracy predicted by the model. The dataset was split in half, with 70% used to train the algorithm and 30% used to test it, resulting in an accuracy of ~93% percent. We compared these results between the features selected by the model in this study and those previously listed by domain experts to see which one yielded the more reliable results. The future for reliable results. This helps the HR system adopt the right scenarios in real time, correctly predict potential employees leaving the company, and know why they are doing so.
Authors: Sheetal S. Patil, S. H. Patil, Avinash M. Pawar, Piyush Pandey, Swastik Sharma, Mrunal S. Bewoor
DOI: https://doi.org/10.1109/asiancon55314.2022.9909382
Publish Year: 2022
