Indian Journal of Science and Technology
DOI: 10.17485/ijst/2015/v8i23/79216
Year: 2015, Volume: 8, Issue: 23, Pages: 1-9
Original Article
Su-Do Kim1 , Yun-Jung Lee2 , Sang Hoon Kang3 , Hwan-Gue Cho4 and Seong-Min Yoon5*
1 Research Institute for Computer and Information Communication, Pusan National University, Republic of Korea; [email protected]
2 Research Institute for Social Criticality, Pusan National University, Republic of Korea; [email protected]
3 Department of Business Administration, Pusan National University, Republic of Korea; [email protected]
4 Department of Computer Science, Pusan National University, Republic of Korea; [email protected]
5 Department of Economics, Pusan National University, Republic of Korea; [email protected]
Food recipes, from traditional recipes to fusion recipes, are easily uploaded and shared online. Recipes consist of a set of ingredients, the cooking procedure, cooking time, etc. It is not easy to classify recipes in terms of the taste of cooked foods, the cuisine styles, or the characteristics of foods. In this paper, we construct the recipe similarity network by adding edges if two different recipes share common ingredients. For this, we newly define the similarity measure among recipes using the probabilistic entropy measures over ingredients. And we construct the ingredient relation network that shows the correlations of ingredients in the recipes. We show these networks can be applied to show the hierarchical structure of 683 recipes and 375 ingredients and the similar recipes are well clustered according to the entropy measure.
Keywords: Complex Network, Ingredient, Ingredient Entropy, Recipe Entropy, Recipe, Recipe Similarity
Subscribe now for latest articles and news.
