Automated Prediction of Employee Attrition Using Ensemble Model Based on Machine Learning Algorithms.

Area of Science:

• Business Analytics
• Human Resources Management
• Machine Learning

Background:

• Employee attrition poses a significant financial and operational challenge for organizations due to the high cost of replacing experienced personnel.
• Retaining competent and experienced employees is crucial for company success and sustainability.
• Existing methods for predicting employee attrition may not be universally optimal across different business contexts.

Purpose of the Study:

• To develop and evaluate an automated model for predicting employee attrition.
• To identify the most effective predictive analytical techniques and pipeline architectures for employee attrition prediction.
• To implement an autotuning approach for optimizing model hyperparameters and create an ensemble model for superior performance.

Main Methods:

• Application of various predictive analytical techniques with diverse pipeline architectures.
• Implementation of an autotuning approach to determine optimal hyperparameter combinations.
• Development of an ensemble model to select the most efficient predictive model based on assessment measures.

Main Results:

• The study demonstrated that no single model is universally ideal for all business contexts.
• The proposed ensemble model achieved optimal performance according to the defined requirements.
• The automated model effectively predicted employee attrition, offering a valuable tool for businesses.

Conclusions:

• Automated predictive models, particularly ensemble approaches, can significantly aid in addressing employee attrition.
• Hyperparameter autotuning is essential for maximizing the efficiency of predictive models in HR analytics.
• While no model is perfect for every scenario, the developed approach provides a robust and adaptable solution for employee retention challenges.