Back to search

Ionian University

Μοντέλα πρόβλεψης απόδοσης ανθρώπινου δυναμικού

Abstract

dc:description

The issue of human resource performance has been of great concern to the Greek public administration in recent years. In this research proposal, an attempt is made to predict performance by estimating the suitability of an employee for the position he holds, based on measurable criteria related to his/her qualifications and the specific job analysis. The key research question is whether the prediction of suitability of an employee for a job position and consequently the prediction of his/her performance could be based on artificial intelligence practices and machine learning models. The innovative objectives of this dissertation are: i) The evaluation of candidates, through a tool based on artificial intelligence for selection of candidates for recruitment, possible reallocation in the organization, placement of suitable employees in the appropriate job position or promotion to authority positions, ii) expression of the opinion of public sector employees regarding the factors that influence their evaluation, for selection and promotion, iii) The creation of an integrated framework with structural equations models for the evaluation parameters regarding the suitability of human resources for selection and promotion in authority positions and the appraisal evaluation parameters, iv) The creation of an evaluation model, which will be connected with the possibility of predicting the performance of employees, exploiting their qualifications and matching the appropriate person to the appropriate position. The interdisciplinarity of the dissertation, with the contribution of the sciences of administration, informatics and statistics is unquestionable and creates new perspectives for the future regarding the issues of human resources management. The selected parameters-criteria are related to the criteria used by the Supreme Council for Civil Personnel Selection (ASEP) for recruitment or promotion. Two questionnaires were conducted to civil servants to collect quantitative and qualitative data. Through a linear model, a total evaluation score was calculated based on the qualifications of the employees. Quantitative research through exploratory and confirmatory factor analyses and structural equations, highlighted the importance of formal qualifications for the evaluation of a candidate for recruitment or for promotion to authority positions as well as for performance evaluation. Many significant and positive correlations were found between formal qualifications and performance abilities, as well as between formal qualifications and experience-training, and finally, between formal qualifications and evaluation. In the next stage, from a large number of available classification algorithms, some of them were selected for the supervised machine learning experiments conducted, based on the criteria of forecasting performance, explainability of the produced model and academic recognition. The two machine learning approaches of the present dissertation, as part of a two-phase testing process, i.e. to predict the suitability of a candidate for a job and to predict the appropriate position for each candidate, according to the available qualifications and job descriptions, revealed as the most appropriate the tree models (REPtree and J48), as they showed the highest prediction accuracy compared to the other algorithms applied. The final proposed tree machine learning model is based on the J48 algorithm, which provides a reliable tool to support the human resource decision-making system for predicting human resource performance, by assessing candidates’ suitability for selection, placement/allocation (horizontal placement) and promotion to authority positions (vertical placement), so that the appropriate people are placed in the appropriate positions. From the attributes (qualifications) used in the machine learning experiments, the most important were the total score of each candidate, in combination to the requirements of the job analysis, as well as the previous experience in authority positions. The proposed model can be the basis for the development of a new automated system for the evaluation of human resources based on machine learning, supporting the relative procedures in the context of a holistic management of the human factor.

Degree

thesis:*
Grantor dc:publisher
Ionian University
Year dc:date
2022

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Pampouktsi, Panagiota
  • Παμπουκτσή, Παναγιώτα

Subjects

dc:subject × 20

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/51841

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Pampouktsi, Panagiota; Παμπουκτσή, Παναγιώτα. Μοντέλα πρόβλεψης απόδοσης ανθρώπινου δυναμικού. Ionian University, 2022. http://hdl.handle.net/10442/hedi/51841