Cuckoo search algorithm for clustering food offers

This paper presents a method for clustering food offers based on the cuckoo search algorithm. The proposed method clusters food offers based on the similarity between their nutritional features (e.g. calcium, vitamins etc.) and/or ingredients. The similarity is evaluated by using the Sorensen-Dice coefficient. To test the clustering method proposed here, we have developed in-house a set of 800 food offers. The food offers have been generated as starting from a set of food recipes (provided in an XML standard for sharing recipes) and a database containing information about nutritional features. This database stores the nutritional features of each food type, as provided by the Agricultural Research Service of the United States Department of Agriculture. We evaluated the performance of our clustering method by using the following metrics: the Dunn Index, the Davies-Bouldin index, and the Average Item-Cluster Similarity.