• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.55.2022.tde-01122022-093014
Document
Author
Full name
Vinicius de Oliveira Boen
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Carlos, 2022
Supervisor
Committee
Nonato, Luis Gustavo (President)
Casaca, Wallace Correa de Oliveira
Pagliosa, Paulo Aristarco
Santos, Maristela Oliveira dos
Title in Portuguese
People Analytics: Aprendizado de máquina na gestão estratégica de pessoas, aplicando modelo preditivo de turnove
Keywords in Portuguese
Aprendizado computacional
Mineração de dados
Recursos humanos
Regressão logística
Turnover
Abstract in Portuguese
Este projeto engloba técnicas de aprendizado de máquina aplicadas no contexto de People Analytics com a aplicação de modelos preditivos supervisionados para classificação de turnover, com o objetivo de auxiliar o processo de tomada de decisão de uma empresa cuja o nome não será citado, impactando diretamente na performance organizacional. Os dados utilizados foram extraídos de uma plataforma de People Analytics, que centraliza as informações dos funcionários. A construção do conjunto de dados final ocorreu através da utilização de técnicas de mineração para manipulação e tratamento. Em seguida, foram aplicadas técnicas de pré-processamento para adequar os dados aos algoritmos de classificação Random Forest, Naive Bayes, Regressão Logística e Decision Tree. Os resultados foram compilados e analisados considerando as principais medidas de avaliação para classificação. Os classificadores Random Forest e Regressão Logística apresentaram os melhores desempenhos e por isso foram selecionados para a etapa de interpretação com a ferramenta SHAP. Dessa análise foi selecionado o classificador regressão logística para a última etapa de análise pós predição, onde foi observado a relação da probabilidade de turnover dos casos de falso positivo com o tempo até turnover futuro desses casos.
Title in English
People Analytics: Machine learning on the strategic management of human resources, applying a predicting turnover
Keywords in English
Computer learning
Data mining
Human resources
Logistic regression
Turnover
Abstract in English
This project encompasses machine learning techniques applied in the context of People Analytics with the application of supervised predictive models for turnover classification, with the objective of helping the decision-making process of a company (whose name will not be mentioned), directly impacting the organizational performance. The data used were extracted from a platform, which centralizes employee information. Data mining techniques were used to manipulate and clean the extracted data and to build a final dataset. Then, pre-processing techniques were applied to adapt the data to the Random Forest, Naive Bayes, Logistic Regression and Decision Tree classification algorithms. The results were compiled and analyzed considering the main evaluation measures for classification. The Random Forest and Logistic Regression classifiers presented the best performances and, therefore, were selected for the interpretation stage with the SHAP tool. From this analysis, the logistic regression classifier was selected for the last stage of post-prediction analysis, where the relationship between the probability of turnover of false positive cases and the time until future turnover of these cases was observed.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2022-12-01
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2023. All rights reserved.