Predicting Employee Attrition Using Machine Learning
Main Article Content
Abstract
This paper delves into the necessity and significance of predicting employee attrition based on their individual circumstances. A key challenge addressed in this study pertains to the presence of noise within such datasets, as well as the inherent inaccuracies stemming from interdependencies among the data points. To tackle this challenge, the paper showcases the effectiveness of the LGBM algorithm in training models with noisy data while maintaining high accuracy levels. Comparative analyses between LGBM and alternative algorithms highlight the algorithm's superior performance, particularly in terms of accuracy in predicting termination.
In terms of future research directions, one potential avenue involves implementing automatic employee allocation. By leveraging the same algorithms and utilizing different datasets, it becomes possible to train models that can predict an employee's aptitude and specialization within the workplace. Consequently, this could facilitate automatic task allocation or work assignment processes.