Optimizing Culinary Association Rules with Genetic Algorithms Using Lift and Novelty
Journal of Applied Informatics and Computing · 22.04.2026 · 10 (2) · 1760-1773
DOI: 10.30871/jaic.v10i2.12284 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Applied Informatics and Computing, online erschienen am 22.04.2026. Verfasst von Larasati Romadhani Yunita Putri und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).
Abstract (Originalsprache)
The culinary industry generates large volumes of transaction data, yet conventional Association Rule Mining often produces excessive rules due to rule explosion. This issue can be addressed using Genetic Algorithms to optimize and select the most relevant rules. While the integration of ARM and Genetic Algorithms (GA) has been widely explored, most existing studies rely on the Apriori algorithm, which is computationally expensive and memory-intensive due to repeated database scans and candidate generation. Furthermore, many GA-based models prioritize confidence and novelty, often overlooking the lift metric, which may lead to rules lacking genuine positive correlation This study proposes an optimized framework that integrates FP-Growth with a Genetic Algorithm-driven evaluation based on lift and novelty to address these efficiency and quality gaps. The use of FP-Growth significantly enhances scalability by reducing memory consumption compared to traditional Apriori-based approaches. The main contribution of this research is a unified integration of FP-Growth and Genetic Algorithm with a multi-layered filtering mechanism enforcing minimum confidence and lift before optimization. This method effectively eliminates rule redundancy, reducing the redundancy rate from 6.06% in standard FP-Growth to 0%. Experiments on 36,164 transaction culinary records (Jan 2024–Sept 2025) demonstrate that the proposed FP-Growth+GA method achieves a 65.1% reduction in peak memory usage (50.68 MB) compared to Apriori+GA (145.54 MB). The method successfully prunes redundant patterns into 11 high-quality rules with lift values > 1 and novelty scores between 0.5 and 1.0, providing concise insights for menu bundling strategies. The implementation is available at https://bit.ly/github_ARM_with_GeneticAlgorithm.
Zitieren
APA
Putri, L. R. Y., Muhaimin, A., & Trimono, T. (2026). Optimizing Culinary Association Rules with Genetic Algorithms Using Lift and Novelty. Journal of Applied Informatics and Computing, 10(2), 1760-1773. https://doi.org/10.30871/jaic.v10i2.12284
Vancouver
Putri LRY, Muhaimin A, Trimono T. Optimizing Culinary Association Rules with Genetic Algorithms Using Lift and Novelty. Journal of Applied Informatics and Computing. 2026;10(2):1760-1773. doi:10.30871/jaic.v10i2.12284
BibTeX
@article{Putri2026,
author = {Putri, Larasati Romadhani Yunita and Muhaimin, Amri and Trimono, Trimono},
title = {Optimizing Culinary Association Rules with Genetic Algorithms Using Lift and Novelty},
journal = {Journal of Applied Informatics and Computing},
year = {2026},
volume = {10},
number = {2},
pages = {1760--1773},
doi = {10.30871/jaic.v10i2.12284}
}
EUVIMED