Skip to content

Read-only version: For security reasons, the Research Information System (FIS) has been temporarily replaced by this read-only version. All public content and full texts remain available; login, editing, submission and statistics are not available. Questions: fis@uni-bamberg.de

Home › Publications › Mirror, Mirror on the Wall, Who Is Leaving of Them All: Predictions for Employee Turnover with Gated Recurrent Neural Networks

Publication

Mirror, Mirror on the Wall, Who Is Leaving of Them All: Predictions for Employee Turnover with Gated Recurrent Neural Networks

de Oliveira, Joao M.; Zylka, Matthäus; Gloor, Peter A.; u. a. (2019): Mirror, Mirror on the Wall, Who Is Leaving of Them All: Predictions for Employee Turnover with Gated Recurrent Neural Networks, in: Yang Song, Francesca Grippa, Peter A. Gloor, u. a. (Hrsg.), Collaborative Innovation Networks. Studies on Entrepreneurship, Structural Change and Industrial Dynamics, Cham: Springer International, S. 43–59, doi: 10.1007/978-3-030-17238-1_2.

Details

Faculty/Chair
Author
de Oliveira, Joao M.
Zylka, Matthäus
Gloor, Peter A.
Joshi, Tushar
Title of the compilation
Collaborative Innovation Networks. Studies on Entrepreneurship, Structural Change and Industrial Dynamics
Editors
Song, Yang
Grippa, Francesca
Gloor, Peter A.
Leitao, Joao
Publisher Information
Springer International : Cham
Year of publication
2019
ISBN
978-3-030-17237-4
Language
English
DOI
Abstract

Employee turnover is a serious issue for organizations and disrupts the organizational behavior in several ways. Hence, predicting employee turnover might help organizations to react to these mostly negative events with, e.g., improved employee retention strategies. Current studies use a “standard analysis approach” (Steel, Academy of Management Review 27:346–360, 2002) to predict employee turnover; accuracy in predicting turnover by this approach is only low to moderate. To address this shortcoming, we conduct a deep learning experiment to predict employee turnover. Based on a unique dataset containing 12 months of time series of e-mail communication from 3952 managers, our model reached an accuracy of 80.0%, a precision of 74.5%, a recall of 84.4%, and a Matthews correlation coefficient value of 61.5%. This paper contributes to turnover literature by providing a novel analytical perspective on key elements of turnover models.

Keywords
turnover
Recurrent neural networks
Type
Contribution to an Articlecollection
URI
Activation date
2019-07-01, 2019-07-01

Permalink: https://fis.uni-bamberg.de/handle/uniba/45757 · Data as JSON · Cite: BibTeX · RIS

Data as of: 2026-11-16

Restricted read-only access: login, editing and submission are temporarily disabled.