PERBANDINGAN KINERJA ALGORITMA XGBOOST DAN LIGHTGBM DALAM KLASIFIKASI STATUS GIZI PADA BALITA DENGAN OPTIMASI HYPERPARAMETER GRIDSEARCHCV

Hegar, Muhammad Yustaf Lana Badriul (2025) PERBANDINGAN KINERJA ALGORITMA XGBOOST DAN LIGHTGBM DALAM KLASIFIKASI STATUS GIZI PADA BALITA DENGAN OPTIMASI HYPERPARAMETER GRIDSEARCHCV. Undergraduate thesis, Universitas Pembangunan Nasional "Veteran" Jawa Timur.

	Text (Cover) Cover.pdf Download (1MB)
	Text (Bab 1) Bab 1.pdf Download (280kB)
	Text (Bab 2) Bab 2.pdf Restricted to Repository staff only until 5 December 2027. Download (483kB)
	Text (Bab 3) Bab 3.pdf Restricted to Repository staff only until 5 December 2027. Download (580kB)
	Text (Bab 4) Bab 4.pdf Restricted to Repository staff only until 5 December 2027. Download (1MB)
	Text (Bab 5) Bab 5.pdf Download (263kB)
	Text (Daftar Pustaka) Daftar Pustaka.pdf Download (207kB)
	Text (Lampiran) Lampiran.pdf Restricted to Repository staff only Download (414kB)

Abstract

This study aims to compare the performance of the XGBoost and LightGBM algorithms in classifying nutritional status in toddlers into four classes: well nourished, undernourished, overnourished, and severely malnourished. The initial stages of the study included data preprocessing, which included removing irrelevant columns, removing duplicate data, handling missing values, and data transformation. To address class imbalance, a Smote process was performed on the training data. Experiments were conducted by testing various data split schemes of 80:20, 70:30, and 60:40, as well as several key parameters of both models, such as learning rates of 0.01, 0.05, 0.1, and 0.15, max depths of 3, 4, 5, and 6, and n_estimators of 50, 100, 150, and 200. The test results showed that the best parameter combination for both models was a data split of 70:30, a learning rate of 0.15, a max depth of 6, and n_estimators of 200. The models were then evaluated using three approaches, namely GridSearchCV optimization, evaluation of performance metrics of accuracy, precision, recall, and F1-Score, and 1–10-fold cross-validation. The results showed that the LightGBM algorithm provided the best performance at a data split of 70:30 with accuracy, precision, recall, and F1 Score values of 92% each. Meanwhile, the XGBoost algorithm, with the same proportions, achieved 88% accuracy, 89% precision, 88% recall, and an F1-score of 88%. In terms of computational time, LightGBM required 304 seconds, while XGBoost was more efficient with a time of 156 seconds. Overall, the results of this study indicate that LightGBM excels in classification performance, while XGBoost is more efficient in computational time.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Rahajoe, Ani Dijah	NIDN0012057301	anidijah.if@upnjatim.ac.id
Thesis advisor	Haromainy, Muhammad Muharrom Al	NIDN0701069503	muhammad.muharrom.if@upnjatim.ac.id

Subjects:

T Technology > T Technology (General)

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Muhammad Yustaf Lana Badriul Hegar

Date Deposited:

05 Dec 2025 08:01

Last Modified:

05 Dec 2025 08:39

URI:

https://repository.upnjatim.ac.id/id/eprint/48041