GRU-based capsule network with an improved loss for personnel performance prediction

Personnel performance is a key factor to maintain core competitive advantages. Thus, predicting personnel future performance is a significant research domain in human resource management (HRM). In this paper, to improve the performance, we propose a novel method for personnel performance prediction...

Full description

Saved in:
Bibliographic Details
Published in:Applied intelligence (Dordrecht, Netherlands) Netherlands), 2021-07, Vol.51 (7), p.4730-4743
Main Authors: Xue, Xia, Gao, Yi, Liu, Meng, Sun, Xia, Zhang, Wenyu, Feng, Jun
Format: Article
Language:English
Subjects:
Artificial Intelligence
Classification
Clustering
Computer Science
Correlation analysis
Decision making
Deep learning
Design
Human performance
Human resource management
Machines
Manufacturing
Mechanical Engineering
Methods
Neural networks
Performance enhancement
Performance prediction
Personnel
Processes
Strategy
Success
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Personnel performance is a key factor to maintain core competitive advantages. Thus, predicting personnel future performance is a significant research domain in human resource management (HRM). In this paper, to improve the performance, we propose a novel method for personnel performance prediction which helps decision-makers select high-potential talents. Specifically, for modeling the personnel performance, we first devise a GRU model to learn sequential information from personnel performance data without any expertise. Then, to better cluster the features, we exploit capsule network. Finally, to precisely make predictions, we further design one strategy, i.e., an improved loss function, and embed it into the capsule network. In addition, by introducing this strategy, our proposed model can well deal with the imbalanced data problem. Extensive experiments on real-world data clearly demonstrate the effectiveness of the proposed approach.
ISSN:0924-669X
1573-7497
DOI:10.1007/s10489-020-02039-x