APPLICATION OF ADAPTIVE CROSS VALIDATION AND PRINCIPAL COMPONENT ANALYSIS OPTIMIZATION FOR EMPLOYEE TURNOVER PREDICTION USING ENHANCED GRAPH EMBEDDING

Authors

  • ABDULLAHI JIBRIL ABDULLAHI Department of Computer Science, Kano state polytechnic, Kano, P.M.B.3401, Kano, Nigeria
  • NASIMA IBRAHIM Department of Computer Science, Kano state polytechnic, Kano, P.M.B.3401, Kano, Nigeria
  • AISHA FARIDA AHMAD Department of Computer Science, Kano state polytechnic, Kano, P.M.B.3401, Kano, Nigeria

DOI:

https://doi.org/10.56892/bima.v7i01.411

Keywords:

Cross validation, Graph embedding, Bipartite graph, Horary random walk.

Abstract

A rapid disclosure of an experience employee from organization (known as employee turnover), had necessitate organizations to make use of the recent development of information technology, to model their data collected over a period of time to enhance decision making, considering that, employee turnover continue to degrade performance of their organizations., Sometime, it requires much time and money for organization to get the equivalent replacements and get them train., consequently, predicting the likely hood of an employee to resign will allow the organization to take proactive measures to control the losses and costs., To address this problem, previous researches focus on exermining some impact factors., In this study, we consider modelling employee’s job historical data to form a dynamic bipartite graph between employees and organizations and learnt a vector representation of this graph, we achieved this by developing a model that generates a sequence for each vertex in the graph using a horary random walk (HRW) method and input the sequence to a skip-gram-negative-sampling (SGNS) to obtain the vector representation for each vertex., and add this vectors to the employee’s basic information and use it as input to machine learning classifiers to predict the employee’s turnover., specifically , we proposed a graph embedding and prediction model that investigate the role of cross validation in predicting employee turnover called Enhanced Graph Embedding and Prediction (EGEP)., Moreover, an experimental result indicated that our method had significantly enhanced the prediction performance of the employee turnover with 11.09%,  11.06% and 13.93% in precision, recall and F1 respectively.

Downloads

Published

2023-03-30

How to Cite

JIBRIL ABDULLAHI, A. ., IBRAHIM , N. ., & FARIDA AHMAD , A. . (2023). APPLICATION OF ADAPTIVE CROSS VALIDATION AND PRINCIPAL COMPONENT ANALYSIS OPTIMIZATION FOR EMPLOYEE TURNOVER PREDICTION USING ENHANCED GRAPH EMBEDDING. BIMA JOURNAL OF SCIENCE AND TECHNOLOGY (2536-6041), 7(01), 308-323. https://doi.org/10.56892/bima.v7i01.411