2024_Deep Ensemble Learning Averaging Method (DELAM): An Enhancement Model For Student’s Performance Classification
| Format: | General Document |
|---|
| _version_ | 1860798223189278720 |
|---|---|
| building | INTELEK Repository |
| collection | Online Access |
| collectionurl | https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 |
| copyright | Copyright©PWB2025 |
| country | Malaysia |
| date | 2024-02-15 14:40 |
| format | General Document |
| id | 16488 |
| institution | UniSZA |
| originalfilename | 16488_7e68a8cb36fee9a.pdf |
| person | PDFsam Basic v5.2.5 Hasnah Binti Nawang |
| recordtype | oai_dc |
| resourceurl | https://intelek.unisza.edu.my/intelek/pages/view.php?ref=16488 |
| sourcemedia | Server storage Scanned document |
| spelling | 16488 https://intelek.unisza.edu.my/intelek/pages/view.php?ref=16488 https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 General Document Malaysia Library Staff (Top Management) Library Staff (Management) Library Staff (Support) Terengganu Faculty of Islamic Contemporary Studies English application/pdf 1.5 Server storage Scanned document Universiti Sultan Zainal Abidin UniSZA Private Access Universiti Sultan Zainal Abidin PDFsam Basic v5.2.5 SAMBox 3.0.10; modified using iTextSharp™ 5.5.10 ©2000-2016 iText Group NV (AGPL-version) Copyright©PWB2025 278 Hasnah Binti Nawang Machine learning Educational data mining Dissertations, Academic 2024-02-15 14:40 16488_7e68a8cb36fee9a.pdf Deep Ensemble Learning Averaging Method DELAM Student’s Performance Classification Deep learning (Machine learning) Ensemble learning (Machine learning) Prediction of scholastic success Education—Data processing Academic achievement—Evaluation Students—Rating of 2024_Deep Ensemble Learning Averaging Method (DELAM): An Enhancement Model For Student’s Performance Classification Classifying educational data can be challenging due to variations and inconsistencies caused by students' characteristics and teaching methods. Although machine learning techniques using single classifiers are widely utilized for this purpose, there is an increasing need to explore approaches that can further enhance prediction accuracy. Studies show that ensemble methods can significantly improve accuracy compared to single classifiers by leveraging multiple classifiers' diversity and collective decisionmaking to achieve more accurate and reliable predictions. Therefore, this thesis aims to improve the accuracy of students’ performance classification using the ensemble method and deep learning approaches. To address this objective, the thesis proposed an ensemble method based on deep learning approaches for students’ certificate classification. The dataset used in this study comprises actual records from seven different high schools in Malaysia, with a total of 3,664 students' records. Six machine learning techniques were applied to determine the most accurate model for students' certificate classification: Support Vector Machine (SVM), Random Forest (RF), k- Nearest Neighbor (KNN), Decision Tree (DT), Logistic Regression (LR), and Naïve Bayes (NB). In order to enhance the performance of the single predictive model, a combination of filter feature selection (FFS) approaches was employed to select relevant features. FFS calculates the optimal score of features using a combination of Information Gain and Pearson Correlation, which are relevant to the target variable. This thesis proposed a Multiple Classifier System (MCS) using the Weighted Majority Voting (WeMV) approach as the combiner rule to improve the accuracy of a single classification. Then, the study applied an additional ensemble approach using Multilayer Perceptrons (MLPs) algorithm to compare the performance of different models using deep ensemble learning. Deep ensemble learning utilizes the output of six classifiers as metadata and implements multiple runs of deep learning models. The predictions generated by these models are then combined using the Deep Ensemble Learning Averaging Method (DELAM) to determine the most accurate classification outcome. The highest accuracy achieved by the ensemble approach was 91.47 percent using the DELAM method, which combined RF, SVM, NB, and KNN classifiers. This outperformed the accuracy achieved by MCS with WeMV, which combined RF, SVM, and DT classifiers with an accuracy of 88.15 percent. In conclusion, the results show that the most effective fusion of classifiers used together with the DELAM in building the ensemble can significantly enhance the performance of the ensemble in terms of accuracy. The results of this study have practical applications for enhancing classification performance in educational environments and could provide insights for further research in this field. Thesis |
| spellingShingle | 2024_Deep Ensemble Learning Averaging Method (DELAM): An Enhancement Model For Student’s Performance Classification |
| state | Terengganu |
| subject | Machine learning Educational data mining Dissertations, Academic Deep learning (Machine learning) Ensemble learning (Machine learning) Prediction of scholastic success Education—Data processing Academic achievement—Evaluation Students—Rating of |
| summary | Classifying educational data can be challenging due to variations and inconsistencies caused by students' characteristics and teaching methods. Although machine learning techniques using single classifiers are widely utilized for this purpose, there is an increasing need to explore approaches that can further enhance prediction accuracy. Studies show that ensemble methods can significantly improve accuracy compared to single classifiers by leveraging multiple classifiers' diversity and collective decisionmaking to achieve more accurate and reliable predictions. Therefore, this thesis aims to improve the accuracy of students’ performance classification using the ensemble method and deep learning approaches. To address this objective, the thesis proposed an ensemble method based on deep learning approaches for students’ certificate classification. The dataset used in this study comprises actual records from seven different high schools in Malaysia, with a total of 3,664 students' records. Six machine learning techniques were applied to determine the most accurate model for students' certificate classification: Support Vector Machine (SVM), Random Forest (RF), k- Nearest Neighbor (KNN), Decision Tree (DT), Logistic Regression (LR), and Naïve Bayes (NB). In order to enhance the performance of the single predictive model, a combination of filter feature selection (FFS) approaches was employed to select relevant features. FFS calculates the optimal score of features using a combination of Information Gain and Pearson Correlation, which are relevant to the target variable. This thesis proposed a Multiple Classifier System (MCS) using the Weighted Majority Voting (WeMV) approach as the combiner rule to improve the accuracy of a single classification. Then, the study applied an additional ensemble approach using Multilayer Perceptrons (MLPs) algorithm to compare the performance of different models using deep ensemble learning. Deep ensemble learning utilizes the output of six classifiers as metadata and implements multiple runs of deep learning models. The predictions generated by these models are then combined using the Deep Ensemble Learning Averaging Method (DELAM) to determine the most accurate classification outcome. The highest accuracy achieved by the ensemble approach was 91.47 percent using the DELAM method, which combined RF, SVM, NB, and KNN classifiers. This outperformed the accuracy achieved by MCS with WeMV, which combined RF, SVM, and DT classifiers with an accuracy of 88.15 percent. In conclusion, the results show that the most effective fusion of classifiers used together with the DELAM in building the ensemble can significantly enhance the performance of the ensemble in terms of accuracy. The results of this study have practical applications for enhancing classification performance in educational environments and could provide insights for further research in this field. |
| title | 2024_Deep Ensemble Learning Averaging Method (DELAM): An Enhancement Model For Student’s Performance Classification |
| title_full | 2024_Deep Ensemble Learning Averaging Method (DELAM): An Enhancement Model For Student’s Performance Classification |
| title_fullStr | 2024_Deep Ensemble Learning Averaging Method (DELAM): An Enhancement Model For Student’s Performance Classification |
| title_full_unstemmed | 2024_Deep Ensemble Learning Averaging Method (DELAM): An Enhancement Model For Student’s Performance Classification |
| title_short | 2024_Deep Ensemble Learning Averaging Method (DELAM): An Enhancement Model For Student’s Performance Classification |
| title_sort | 2024_deep ensemble learning averaging method (delam): an enhancement model for student’s performance classification |