Loading [MathJax]/extensions/TeX/extpfeil.js
Comparative Study of Predictive Models to Estimate Employee Attrition | IEEE Conference Publication | IEEE Xplore

Comparative Study of Predictive Models to Estimate Employee Attrition

Publisher: IEEE

Abstract:

Employees put in a lot of effort to make the company prosperous. As a result, this research work attempts to construct a model that, utilizing HR analytics data provided ...View more

Abstract:

Employees put in a lot of effort to make the company prosperous. As a result, this research work attempts to construct a model that, utilizing HR analytics data provided from the Kaggle platform, can forecast employee attrition rate in this research. Attrition proves to be a costly and time-consuming problem for the organization, and it also leads to loss of productivity. The relationship among attributes is shown using correlation matrix and heatmap. The histogram, which depicts the disparity between left and right employees, is constructed in the experimental portion in terms of income, department, level of pleasure, and so on. The proposed research employs four distinct machine learning techniques for prediction, including Logistic Regression, Random Forest, Naive Bayes classifier and K-Nearest Neighbor. This document proposes the factors that help an organization’s staff turnover rate to be as low as possible.
Date of Conference: 22-24 June 2022
Date Added to IEEE Xplore: 29 July 2022
ISBN Information:
Publisher: IEEE
Conference Location: Coimbatore, India

I. Introduction

Staff experiences a severe recession during the pandemic, which makes many organizations wary of completing projects successfully because it limits productive time, teamwork, and the cost of training new employees. This necessitates a significant financial investment as well as time to acclimate them to the workplace and its structure. However, if they abruptly leave their jobs, any firm may face significant financial losses. Because new hires not only cost money and time, but also take time to turn a profit for the company. As a result, the study’s purpose is to develop a model that HR analytics data may be used to anticipate staff turnover rates. Lockdowns and other restrictions imposed on companies and consumers at the start of the outbreak raised the jobless rate to 14.7 percent in March 2020, according to the Bureau of Labour Statistics with 21.8 million people unemployed. As a result, the idea of utilizing a machine to simulate the forecast of staff churn rate becomes increasingly practical for businesses. As a result, classification techniques are required to prepare the way for a typical real-time challenge. Machine learning techniques can be used to categorize an entity and assist in making appropriate decisions. The company’s recruitment and termination criteria are used to calculate the attrition rate. An employee may resign for a variety of reasons. During feature extraction we perform data pre-processing and finally extract the feature set. For prediction purposes, we employ four different machine algorithms. There are many algorithms out there for predicting the attrition of an employee but for measuring the efficiency we use accuracy which is not a good measure for an imbalanced data so in this paper we are using a good measure for finding efficiency of a machine learning model in such cases. So, in this paper we deal with the attributes which affect the employee attrition vote for the most using attribute relation techniques like correlation technique. By using this technique, we identify the features that best describe the output label i.e., the attrition of an employee. The paper is organized as follows: Section II presents the related work in employee attrition. The proposed work is explained in Section III. Model building is explained in Section IV. Section V gives a case study whereas we concluded the paper in Section VI.

References

References is not available for this document.