Standardizing the fish freshness class during ice storage using clustering approach

• Published in: Ecological informatics 2024-05, Vol.80, p.102533, Article 102533
• Main Authors: Prasetyo, Eko, Suciati, Nanik, Fatichah, Chastine, Aminin, Pardede, Eric
• Format: Article
• Language: English
• Subjects: Clustering; Elbow method; Fish; Freshness class standardization; Organoleptic inspection
• ISSN: 1574-9541
• DOI: 10.1016/j.ecoinf.2024.102533

The freshness of fish before consumption affects the taste of the food and human health, so many parties should consider it carefully. Fish is usually stored on ice to maintain its freshness. However, there has yet to be a standard for the fish freshness class during ice storage. On the other hand, each fish species has a different vigor in retaining freshness. In this research, we propose (1) a dataset of fish freshness inspection generated from daily inspection during ice storage for 11 days, (2) a standard of the freshness class of seven fish species using clustering, and (3) a framework for automatically determining the freshness class of fish with internal validation. The dataset is generated from daily inspection during ice storage for 11 days. Experts conduct the organoleptic inspection using six parameters according to the SNI 2729–2013 standards: eyes, gills, body surface mucus, meat, smell, and body textures. We experimented with six clustering methods on the dataset to obtain freshness clusters automatically. Our experimental results indicate that most species fall into two freshness categories, but Upeneus Moluccensis stands out by being classified into three distinct classes. Each species achieves a different number of days for each freshness class. K-Means and K-Means++ obtained the best freshness group with mean scores of 0.538 and 21.92 for Silhouette and Calinski-Harabasz, respectively. Therefore, the standardization of fish freshness class during ice storage for 11 days achieved satisfactory results and was acceptable as guidance and standard in freshness examination. •The freshness of fish before consumption affects the taste of the food and human health.•A dataset of fish freshness inspection generated from daily inspection.•A standard of the freshness class of seven fish species using clustering.•A framework for automatically determining the freshness class of fish with internal validation.•Organoleptic inspection using six parameters according to SNI 2729–2013 standards.