Pengembangan Sistem Rekomendasi Makanan Untuk Menurunkan Berat Badan Bagi Penderita Obesitas Dengan Metode Content-based Filtering

Suhaymi, Parisya Naylah (2026) Pengembangan Sistem Rekomendasi Makanan Untuk Menurunkan Berat Badan Bagi Penderita Obesitas Dengan Metode Content-based Filtering. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026221138-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Obesitas merupakan kondisi medis akibat penumpukan lemak tubuh berlebih yang menjadi faktor risiko berbagai penyakit kronis seperti diabetes tipe 2, hipertensi, dan penyakit jantung. Penelitian ini mengembangkan sistem rekomendasi menu makanan berbasis content-based filtering untuk mendukung program penurunan berat badan pada penderita obesitas. Sistem menggunakan algoritma K-Nearest Neighbor (KNN) untuk memilih kandidat makanan berdasarkan kedekatan profil nutrisi menggunakan euclidean distance, dan Multiple-Choice Knapsack Problem yang diselesaikan dengan Dynamic Programming untuk mengoptimalkan kombinasi makanan lintas kelompok dalam batas kapasitas kalori setiap slot waktu makan. Kebutuhan kalori dihitung menggunakan rumus Mifflin-St Jeor dan disesuaikan dengan defisit kalori harian untuk mencapai target penurunan berat badan yang aman. Dataset yang digunakan adalah data nutrisi USDA yang telah melalui tahap preprocessing meliputi pembersihan data, pengelompokan makanan, dan normalisasi. Evaluasi menggunakan metrik Mean Absolute Percentage Error (MAPE) menunjukkan bahwa sistem mampu memenuhi target kalori harian secara konsisten dalam rentang toleransi 90-110%. Evaluasi oleh ahli gizi menyatakan sistem sudah sesuai dari sisi perhitungan kebutuhan energi dan distribusi kalori per waktu makan, namun memerlukan pengembangan konten makanan agar lebih sesuai dengan pedoman Isi Piringku Kementerian Kesehatan RI dan pola makan lokal Indonesia.
===================================================================================================================================
Obesity is a medical condition caused by excessive body fat accumulation and is a major risk factor for chronic diseases such as type 2 diabetes, hypertension, and heart disease. This study develops a food recommendation system based on content-based filtering to support weight loss programs for obese patients. The system employs the K-Nearest Neighbor algorithm to select food candidates based on nutritional profile similarity using Euclidean distance, and the Multiple-Choice Knapsack Problem solved with Dynamic Programming to optimize food combinations across food groups within the caloric budget of each meal slot. Caloric requirements are calculated using the Mifflin-St Jeor formula and adjusted to a daily caloric deficit to achieve a safe weight loss target. The dataset used is USDA nutritional data that underwent preprocessing including data cleaning, food grouping, and normalization. Evaluation using the Mean Absolute Percentage Error (MAPE) metric shows that the system consistently meets daily caloric targets within a 90-110% tolerance range. Clinical evaluation by a nutritionist confirms the system is appropriate in terms of energy requirement calculation and calorie distribution per meal time, but requires further development on food content to better align with Indonesia's Ministry of Health Isi Piringku dietary guidelines and local eating habits.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Obesitas, Sistem rekomendasi, Content-Based Filtering, K-Nearest Neighbor, Penurunan Berat Badan, Obesity, Recommendation System, Content-Based Filtering, Calorie Needs, Weight Loss, K-Nearest Neighbor
Subjects:	Q Science > QA Mathematics > QA76.9.I58 Recommender systems (Information filtering)
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Parisya Naylah Suhaymi
Date Deposited:	20 Jul 2026 06:55
Last Modified:	20 Jul 2026 07:08
URI:	http://repository.its.ac.id/id/eprint/135690

Actions (login required)

View Item