Optimisasi Menu Balita Berdasarkan Gizi Harian Menggunakan XGBoost, Genetic Algorithm, dan Particle Swarm Optimization

Safitri, Evyra Rizki (2026) Optimisasi Menu Balita Berdasarkan Gizi Harian Menggunakan XGBoost, Genetic Algorithm, dan Particle Swarm Optimization. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6026241001-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (10MB) | Request a copy

Abstract

Perencanaan menu balita perlu mempertimbangkan kondisi antropometri dan riwayat alergi makanan. Namun, perencanaan secara manual sering belum mampu menyesuaikan kebutuhan gizi yang berbeda. Penelitian ini mengembangkan sistem rekomendasi menu harian balita dengan mengintegrasikan XGBoost untuk klasifikasi kelas menu serta Genetic Algorithm (GA) dan Particle Swarm Optimization (PSO) untuk optimasi kombinasi bahan makanan. Data penelitian terdiri atas 236 data valid balita dari Posyandu Sandingrowo, Tuban, Tabel Komposisi Pangan Indonesia sebagai sumber kandidat bahan pangan, serta Angka Kecukupan Gizi sebagai acuan target gizi. XGBoost digunakan untuk memprediksi tiga kelas menu, yaitu Catch-Up Nutrition, Balanced Nutrition, dan Energy Control, yang kemudian dipetakan menjadi target gizi personal. GA dan PSO dibandingkan pada enam skenario berdasarkan RDI Score, error Z, konvergensi, deviasi nutrien, dan waktu komputasi. Hasil menunjukkan bahwa XGBoost mampu mengklasifikasikan kelas menu dengan baik, dengan zscore_bbtb sebagai fitur paling berpengaruh. GA lebih unggul pada beberapa skenario Catch-Up Nutrition dan Balanced Nutrition, sedangkan PSO lebih unggul pada Energy Control dan secara konsisten membutuhkan waktu komputasi lebih singkat. Uji pengguna memperoleh nilai 86,97% dan validasi ahli sebesar 83,33%, keduanya termasuk kategori sangat baik. Sistem berhasil menghasilkan meal plan personal yang mendekati target gizi dan bebas alergen.
===============================================================================================================================
Toddler meal planning should consider anthropometric conditions and food allergy history. However, manual planning often cannot accommodate different nutritional needs. This study developed a daily meal recommendation system integrating XGBoost for menu-class classification with Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) for food-combination optimization. The study used 236 valid toddler records from Posyandu Sandingrowo, Tuban, the Indonesian Food Composition Table as the food candidate source, and the Indonesian Recommended Dietary Allowance as the nutritional reference. XGBoost predicted three menu classes: Catch-Up Nutrition, Balanced Nutrition, and Energy Control, which were then mapped to personalized nutritional targets. GA and PSO were compared across six scenarios using RDI Score, Z error, convergence, nutrient deviation, and computation time. The results showed that XGBoost classified the menu classes effectively, with zscore_bbtb as the most influential feature. GA performed better in several Catch-Up Nutrition and Balanced Nutrition scenarios, while PSO performed better in Energy Control and consistently required less computation time. User testing scored 86.97%, and expert validation reached 83.33%, both categorized as very good. The system successfully generated personalized, allergen-free meal plans that closely matched nutritional targets.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Balita, Optimasi Menu, XGBoost, Genetic Algorithm, Particle Swarm Optimization Toddler, Menu Optimization, XGBoost, Genetic Algorithm, Particle Swarm Optimization
Subjects:	T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 59101-(S2) Master Thesis
Depositing User:	Evyra Rizki Safitri
Date Deposited:	28 Jul 2026 01:20
Last Modified:	28 Jul 2026 01:20
URI:	http://repository.its.ac.id/id/eprint/138184

Actions (login required)

View Item