Pengembangan Sistem Rekomendasi Berbasis Collaborative Filtering Dengan Algoritma K-Nearest Neighbor Untuk Pembelian Frozen Food

Dewantara, Ananda Surya Dahana (2025) Pengembangan Sistem Rekomendasi Berbasis Collaborative Filtering Dengan Algoritma K-Nearest Neighbor Untuk Pembelian Frozen Food. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026211142-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Frozen food merupakan produk pangan olahan yang semakin populer di masyarakat modern karena kemudahan dalam penyimpanan dan konsumsi. Penelitian ini bertujuan untuk mengembangkan sistem rekomendasi produk frozen food menggunakan algoritma K-Nearest Neighbors (KNN) dengan pendekatan collaborative filtering yang mencakup tiga jenis preferensi, yaitu berbasis pengguna (user-based), produk (item-based), dan merek (brand-based). Sistem dikembangkan melalui tahapan pengumpulan dan preprocessing data transaksi, pembentukan matriks interaksi, pelatihan model KNN, serta evaluasi menggunakan metrik precision, recall, dan confusion matrix. Hasil evaluasi menunjukkan bahwa sistem mampu memberikan rekomendasi yang akurat, khususnya pada nilai precision yang tinggi meskipun nilai recall masih rendah pada Top-5. Sistem kemudian diimplementasikan ke dalam bentuk website interaktif menggunakan React.js pada frontend dan FastAPI pada backend, serta berhasil dideploy secara daring menggunakan layanan Vercel dan Railway. Website ini memungkinkan pengguna untuk memilih produk, menentukan preferensi rekomendasi, dan memperoleh hasil rekomendasi secara real-time melalui antarmuka yang interaktif dan responsif.
====================================================================================================================================
Frozen food is a processed food product that has become increasingly popular among modern consumers due to its convenience in storage and preparation. This study aims to develop a recommendation system for frozen food products using the K-Nearest Neighbors (KNN) algorithm with a collaborative filtering approach, covering three main preference types: user-based, item-based, and brand-based. The system is developed through stages of data collection, preprocessing, interaction matrix construction, KNN model training, and evaluation using information retrieval metrics such as precision, recall, and confusion matrix. Evaluation results indicate that the system is capable of generating accurate recommendations, particularly with high precision values, although recall remains low at Top-5. The system is then implemented into an interactive website using React.js for the frontend and FastAPI for the backend, and successfully deployed online using Vercel and Railway services. This web-based platform enables users to select products, determine recommendation preferences, and receive real-time suggestions through an intuitive and responsive interface.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Collaborative Filtering, Frozen Food, K-Nearest Neighbors, Sistem Rekomendasi, Website Interaktif, Interactive Website, Recommendation System
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:	Ananda Surya Dahana Dewantara
Date Deposited:	28 Jul 2025 08:00
Last Modified:	28 Jul 2025 08:00
URI:	http://repository.its.ac.id/id/eprint/122385

Actions (login required)

View Item