Determining Factors Resulting to Employee Attrition Using Data Mining Techniques

What it’s about

The Business Process Outsourcing (BPO) industry is plagued by high employee attrition rates, leading to significant costs in recruitment, training, and lost productivity. This study addresses this critical challenge by demonstrating how data mining techniques can be used to build a highly accurate predictive model for employee turnover. By analyzing a real-world dataset from a BPO company, the authors use a Fast Correlation-Based Filtering Algorithm to isolate the 17 most significant predictors of attrition and then apply a Naïve Bayes Algorithm to classify employees as likely 'stayers' or 'hoppers'. The resulting model achieves 84.69% accuracy and reveals that Total Working Years, Marital Status, and Age are the most powerful predictors. This provides HR managers with a practical, data-driven tool to anticipate turnover, develop targeted retention strategies, and make smarter hiring decisions, ultimately saving costs and building a more stable workforce.

The through-line

Who it’s for

An HR manager or business leader in the Business Process Outsourcing (BPO) industry who is struggling with high employee turnover and its associated costs.

The problem

The constant cycle of employees leaving the company (high attrition rate) leads to increased recruitment costs, loss of productivity, and operational disruption. Feeling frustrated, overwhelmed, and uncertain about how to identify at-risk employees and effectively retain talent.

The plan

1. Gather relevant employee data from HR records.
2. Apply a filtering algorithm to identify the most significant predictor variables for attrition.
3. Use the Naïve Bayes classification algorithm to build a predictive model.
4. Evaluate the model's accuracy using a confusion matrix cross-validation.
5. Deploy the model to identify at-risk employees and inform retention strategies.

The payoff

Reduced employee turnover and lower recruitment and training costs. · The ability to create targeted and effective employee retention strategies. · Improved strategic workforce planning and budgeting.

See our guide

• The Four-S Discipline — One System, Four Capabilities →
• Build a People Analytics Practice →

Related profiles we’ve built

• The basic principle of people analytics learn how to use hr data to drive better outcomes for your business and employees →
• Handbook of Regression Modeling in People Analytics →
• An Everyone Culture: Becoming a Deliberately Developmental Organization →
• Apa Handbook Io V2 →
• Armstrongs Handbook of Strategic Human Resource Management Improve Business Performance Through Strategic People Management →
• Beyond Hr Boudreau Ramstad →
• The Differentiated Workforce →
• Drive Pink →

Additional reading

• Using Data Mining Techniques to Build a Classification Model for Predicting Employees Performance · Q. a. Al-Radaideh and E. Al Naqi

  This cited paper shows a similar application of data mining techniques within HR, but focuses on predicting performance instead of attrition, providing a complementary perspective.

• How The Naive Bayes Classifier Works In Machine Learning · R. Saxena

  Provides a deeper, foundational understanding of the core algorithm used in this study, which is useful for readers who want to understand the mechanics of the model.

• Data Mining: A prediction for performance improvement using classification · B. K. Bhardwaj and S. Pal

  This reference demonstrates the versatility of classification models by showing their use in an educational context to predict student performance, paralleling this study's corporate application.

Featured in these guides

• Build a People Analytics Practice →Cited source