Food Recommendations to Support Unsold Food Marketplace Using Content-Based Filtering

Limanto, Susana and Liliana, Liliana and Soesanto, Daniel and Louk, Maya Hilda Lestari and Arwoko, Heru and Cahyawati, Fenny (2026) Food Recommendations to Support Unsold Food Marketplace Using Content-Based Filtering. Applied Information System and Management (AISM), 9 (1). pp. 71-80. ISSN 2621-2544

PDF Food Recommendation.pdf - Published Version Download (1MB)

Abstract

Unsold food is food that has a short shelf life, is intended to be sold, and is still edible. Based on survey results, business owners usually resell unsold food at a discounted price via WhatsApp or posters placed in front of the shop or distribute it to the surrounding community. Meanwhile, consumers buy unsold food by visiting the shop directly. Unsold food promotions via WhatsApp or posters do not reach the wider community. Meanwhile, direct purchases limit the opportunity to buy unsold food from several sellers simultaneously and increase the risk of stockouts, which can lead to wasted food and missed savings for consumers. This study aims to develop an unsold food marketplace integrated with two key features: bargaining and recommendations. Recommendations are generated using content-based filtering with cosine similarity to measure the similarity between the user's purchase history and each unsold food item. The recommendation feature's findings reveal that content-based filtering generates recommendations more in line with user preferences than popularity-based ones. Validation results confirm this finding, demonstrating 100% accuracy in matching the recommended food categories with the ones users have purchased. Meanwhile, during the marketplace validation stage, 15 respondents reported strong acceptance, with average scores of 4.67 out of 5 for usefulness and 4.71 out of 5 for usability. This study highlights how an unsold food marketplace supports consumers with limited budgets, reduces food waste, and increases seller revenue, while its bargaining and recommendation features enhance user satisfaction and engagement, thereby achieving mutual benefits.

Item Type:	Article
Subjects:	Q Science > Q Science (General) T Technology > T Technology (General)
Divisions:	Faculty of Engineering > Department of Informatic
Depositing User:	Susana 6169
Date Deposited:	04 May 2026 08:33
Last Modified:	04 May 2026 08:33
URI:	http://repository.ubaya.ac.id/id/eprint/50690

Actions (login required)

View Item