1. RIUdeG
  2. Tesis
  3. Doctorado
  4. CUCEA
Por favor, use este identificador para citar o enlazar este ítem: https://hdl.handle.net/20.500.12104/109923
Registro completo de metadatos
Campo DCValorLengua/Idioma
dc.contributor.advisorChavoya Peña, Arturo
dc.contributor.advisorLarios Rosillo, Víctor Manuel
dc.contributor.advisorMorales Gamboa, Rafael
dc.contributor.authorUlloa Cazarez, Rosa Leonor
dc.date.accessioned2025-09-01T22:42:31Z-
dc.date.available2025-09-01T22:42:31Z-
dc.date.issued2016-11-28
dc.identifier.urihttps://wdg.biblio.udg.mx
dc.identifier.urihttps://hdl.handle.net/20.500.12104/109923-
dc.description.abstractIn this research I consider the following three issues as the context for online higher education (OHE): (1) Higher education (HE) has a steady enrolment growth [1, 2] and budget issues are becoming structural and survival matters for universities and HE institutions [3]. Moreover, HE institutions (HEI) are required to pay attention of the society needs regarding to the demands of the workplace according to international organizations, which are considered as stakeholders, assessors and certifying bodies of HEIs [4, 5, 6, 7]: HEIs should meet international organizations quality criteria; and OHE as a subsystem within HE system is embedded in this environment. (2) Most of the OHE offers are dealing with self-sustainability issues and are experiencing an enrolment growth [8, 9]. Thus, the increasing of economic resources is a matter of concern for the OHEs. (3) Investment is dependent on evaluation [10], and based on indicators such as failure, success, and terminal efficiency; these three indicators are quality issues for OHE [11, 12], they become important in order to sustain and increase the inflow of resources. Furthermore, economic indicators are related to knowledge production, long term education, and training of human resources of a country: these indicators are affected by student success, failure, retention, dropout, and graduation rates among other educational indicators. OHE as an extension of HE [7] has a similar behavior to HE: the enrolment growth, failure, success and dropout rates have comparable patterns. However, OHE has its own challenges and characteristics, which I describe in the following sections. Failure rates are high in OHE [13]: to deal with this major issue I will link the field of information technologies with the educational field; my contribution is to propose a solution to handle one of the indicators that are important for the development of the OHE institutions and programs: the student failure rate.
dc.description.tableofcontentsINDEX Abbreviations and Acronyms V List of Tables VII List of Figures VIII Introduction 1 Problem approach 2 Research questions 2 Objectives 2 Hypotheses 3 Chapter I 5 Online Higher Education Framework 5 1.1 General characteristics of Online Higher Education 5 1.2 The learning setting 7 1.3 Context and growth of Online Higher Education 7 1.4 Student performance 8 1.4.1 Approaches to student performance from psychological, cognitive and personality traits 121 1.4.2 Approaches to student performance from demographic data, academic data, and student background 164 1.4.3 Approaches to student performance from data captured in the LMS 17 Chapter II 19 Related works on Student Performance Prediction 19 2.1 Studies predicting student performance 22 2.1.1 Student performance predicted in a binary, ordinal and categorical way 23 2.1.2 Student Performance predicted in a numerical way 26 2.1.2.1 Prediction of student performance by statistical techniques 29 2.1.2.2 Prediction of student performance by machine learning techniques 31 2.2 Conclusion of the related work review 33 Chapter III 353 II Theoretical framework 353 3.1 Genetic Programming 364 3.1.1 Genetic programming general process 396 3.1.1.1 The fitness 408 3.1.1.2 Operations to modify structures: reproduction and crossover 40 3.1.2 Simple Symbolic Regression of Koza 475 3.1.2.1 Definition of terminals and functions 486 3.1.2.2 Definition of fitness measures 497 3.1.2.3 Definition of several parameters 519 3.1.2.4 The termination criterion 50 3.2 Regression analysis 51 3.2.1 Model specification 541 3.2.1.1 Scatter plots 552 3.2.1.1.1 Outliers identification 563 3.2.1.1.2 Functional form 596 3.2.1.2 Correlation coefficient 607 3.2.1.3 Determination Coefficient 618 3.2.1.4 Adjusted R2 629 3.2.1.5 Multiple Linear Regression analysis 629 3.2.2 Parameters identification 60 3.2.2.1 Regression coefficients of the SLR 61 3.2.2.2 The standard deviation ( 2 ) 663 3.2.2.3 Regression coefficients of the MLR 674 3.2.3 Model validation 706 3.2.3.1 Coefficient analysis and predicted values by the model 707 3.2.3.2 Gathering of new data 717 3.2.3.3 Data splitting 718 3.2.4 Model Adequacy 728 3.2.4.1 Residuals and outliers analysis 728 3.2.4.2 Determination coefficient 729 3.2.5 ANOVA analysis 739 III 3.2.6 Prediction accuracy criterion 71 3.2.7 Statistical test for samples comparison 751 Chapter IV 773 Empirical analysis 773 4.1 Data sample description (raw data) 794 4.2 Method for splitting the data sample into training and testing data sets 80 4.3 Statistical analysis 861 4.3.1 Correlational analysis between FG and eleven independent variables 861 4.3.2 Outliers analysis for FG and 11 independent variables 872 4.3.3 Normality test of FG and eleven independent variables 893 4.3.4 Multiple linear regression analysis 905 4.4 Simple linear regression model 926 4.4.1 Calculation of ARs of the LRLS model 938 4.5 Generation of Genetic Programming model based on Symbolic Regression 948 4.5.1 SRGP model generation 959 4.5.2 Calculation of the ARs of the SRGP model 90 4.5.3 Testing the LRLS and the SRGP models 960 4.6 Results 960 4.6.1 Prediction accuracy: LRLS vs. SRGP models for the training data set 971 4.6.2 Comparison of accuracies of the training data sets 971 4.6.3 Prediction accuracy: LRLS vs. SRGP models for the testing data set 983 Chapter V 1004 Empirical analysis 2 1004 5.1 Generation of a Statistical Linear Regression model using transformed data 1004 5.2 Generation of a Genetic Programming model using transformed data 1015 5.2.1 lnGP model 1015 5.2.2 Prediction accuracy: lnLR vs. lnGP models (training data set) 1015 5.2.4 Testing the lnLR and lnGP models 1026 IV 5.3 Results 1026 5.3.1 Comparison of accuracies: lnLR vs. lnGP (training data set) 1026 5.3.2 Comparison of accuracies: lnLR_LN vs. lnGP (testing dataset) 103 Chapter VI 105 Discussion and future work 105 6.1 Contributions 105 6.1.1 Two literature reviews performed 105 6.1.1.1 Literature review about SP 105 6.1.1.2 Literature review about SP prediction 107 6.1.2 The simple linear regression model - LRLS 108 6.1.3 The Genetic Programming model based on Symbolic Regression proposed by Koza 109 6.1.4 Two experimental analysis performed 111 6.2 Discussion 105 6.2.1 Analysis of linear vs. quadratic functions 1135 6.2.2 Patterns observed from the empirical analysis 105 6.3 Limitations and future work 113 References 115 Appendices 129 Appendix 1: Data samples 123 Appendix 2: Calculation of Pearson correlation coefficients 135 Appendix 3: Outlier plots 139 Appendix 4: Normality tests 153 Appendix 5: MLR Analyses 148 Appendix 6: Calculation of simple linear regression parameters 152 Appendix 7: Calculation of ARs for training and testing data sets 154 Appendix 8: Genetic Programming, winner runs for raw and transformed data 160
dc.formatapplication/PDF
dc.language.isoeng
dc.publisherBiblioteca Digital wdg.biblio
dc.publisherUniversidad de Guadalajara
dc.rights.urihttps://www.riudg.udg.mx/info/politicas.jsp
dc.subjectGenetic Programming
dc.subjectOnline Higher Education
dc.subjectStudents
dc.subjectPredict
dc.subjectCourse
dc.titleApplying Genetic Programming to Predict the Final Grade of Students from an Online Higher Education Course
dc.typeTesis de Doctorado
dc.rights.holderUniversidad de Guadalajara
dc.rights.holderUlloa Cazarez, Rosa Leonor
dc.coverageZAPOPAN, JALISCO
dc.type.conacytdoctoralThesis
dc.degree.nameDOCTORADO EN TECNOLOGIAS DE INFORMACION
dc.degree.departmentCUCEA
dc.degree.grantorUniversidad de Guadalajara
dc.rights.accessopenAccess
dc.degree.creatorDOCTOR EN TECNOLOGIAS DE INFORMACION
dc.contributor.directorLópez Martín, Cuauhtémoc
dc.contributor.codirectorAbran, Alain
Aparece en las colecciones:CUCEA

Ficheros en este ítem:
Fichero TamañoFormato 
DCUCEA10158FT.pdf3.23 MBAdobe PDFVisualizar/Abrir
Mostrar el registro sencillo del ítem


Los ítems de RIUdeG están protegidos por copyright, con todos los derechos reservados, a menos que se indique lo contrario.