KLASIFIKASI PENDERITA DIABETES DENGAN PEMERIKSAAN GULA DARAH SEWAKTU (GDS) DENGAN TEKNIK SMOTE MENGGUNAKAN MODEL MULTILAYER PERCEPTRON (MLP) DAN XGBOOST

Putra, Ferry Trilaksana (2025) KLASIFIKASI PENDERITA DIABETES DENGAN PEMERIKSAAN GULA DARAH SEWAKTU (GDS) DENGAN TEKNIK SMOTE MENGGUNAKAN MODEL MULTILAYER PERCEPTRON (MLP) DAN XGBOOST. Undergraduate thesis, UPN "Veteran" Jawa Timur.

	Text (Cover) 21081010280.-cover.pdf Download (1MB)
	Text (Bab 1) 21081010280.-bab1.pdf Download (226kB)
	Text (Bab 2) 21081010280.-bab2.pdf Restricted to Repository staff only until 5 December 2028. Download (637kB)
	Text (Bab 3) 21081010280.-bab3.pdf Restricted to Repository staff only until 5 December 2028. Download (638kB)
	Text (Bab 4) 21081010280.-bab4.pdf Restricted to Repository staff only until 5 December 2028. Download (2MB)
	Text (Bab 5) 21081010280.-bab5.pdf Download (212kB)
	Text (Daftar Pustaka) 21081010280.-daftarpustaka.pdf Download (224kB)
	Text (Lampiran) 21081010280.-lampiran.pdf Restricted to Repository staff only Download (523kB)

Abstract

This study develops a classification model to determine the health status of diabetes patients based on Random Blood Sugar data. The approach applies a serial hybrid method combining the Multilayer Perceptron and XGBoost algorithms. The MLP model functions as a feature extractor for numerical data, while XGBoost serves as the final classifier. The dataset undergoes Encoding, normalization, and class balancing using SMOTE to address the imbalance in Normal, Prediabetes, and Diabetes categories. SMOTE increases the representation of minority classes, enabling the model to learn more evenly across all categories. Several experimental scenarios were conducted, including variations in data splitting, learning rate, Max depth, and the number of epochs. Model performance was evaluated using Accuracy , precision, recall, F1-score, and Confusion Matrix. The results indicate that the MLP model performs adequately but struggles to distinguish Prediabetes samples, achieving an Accuracy of 0.84. XGBoost demonstrates stronger performance on tabular data, with improved stability when combined with MLP. The hybrid MLP and XGBoost model yields a significant performance improvement. With the integration of SMOTE, the model achieves an Accuracy of 0.99, a macro average of 0.98, and a weighted average of 0.99. The Confusion Matrix shows perfect identification of Normal and Diabetes classes. The Prediabetes class, which was previously prone to misclassification in the standalone MLP model, shows more stable performance in the hybrid approach. These findings confirm that integrating MLP and XGBoost enhances the model's ability to identify patterns in Random Blood Sugar data, supporting more accurate medical decision-making.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Puspaningrum, Eva Yulia	NIDN0005078908	evapuspaningrum.if@upnjatim.ac.id
Thesis advisor	Haromainy, Muhammad Muharrom Al	NIDN0701069503	muhammad.muharrom.if@upnjatim.ac.id

Subjects:

Q Science > QA Mathematics > QA76.6 Computer Programming
T Technology > T Technology (General)

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Putra Ferry Trilaksana

Date Deposited:

05 Dec 2025 08:49

Last Modified:

05 Dec 2025 08:49

URI:

https://repository.upnjatim.ac.id/id/eprint/48093