Ensemble Methods with Bidirectional Feature Elimination for Prediction and Analysis of Employee Attrition Rate During COVID-19 Pandemic

Abstract

Due to this pandemic, millions of people have been laid off from their jobs, and in developed countries like the USA, the national unemployment rate was at an astonishing 14.7% in the month of April 2020. Alongside the unfavorable situation and increasing loss in a business, various factors account for employee attrition. The factors affecting employee status are extensively studied, and the observations found are further explained in detail.Attributes such as department, job role, and education have to be primarily considered for analyzing the trend. But considering only these factors is not sufficient to successfully comprehend this issue. Many other factors that might even appear to be trivial at a first glance have to be included to significantly improve the quality of this research. These are stated and explained further. BackgroundThe workforce of a nation largely determines its economic progression [1,2]. The dissatisfaction of employees in an organization could be a potential warning that an organization needs to change its policies [3,4]. The study done by Silpa et al. looks at statistical measures like the coefficient of correlation, Chi-square test, and mean of employee's data to understand the reasons behind them leaving the organization.The researcher in his proposed study [5] has explained unique factors like turnovers in BPO sectors. But the individual employee factors mentioned are hardly two, and the data collection was restricted to only 100 employees.

Predicting Employee Attrition Using Machine Learning Approaches

“…The employee attrition prediction using a machine learning pipeline [21] was proposed in this study [22]. The study findings analyze the factors such as the number of years of work experience, educational qualifications, gender, and department were that caused employee attrition.…”