A Cross-language Approach for Semantic Recipe Annotation and Nutrition Annotation Augmentation Based on Extraction Ontologies and Culture-Relevant Policy Reasoning

Abstract

This paper summarizes an approach for cross language annotation and querying of semi-structured data with the use of cross-language extraction ontologies. The paper describes the capabilities of the approach in the context of annotation of nutritional components and their respective values for ingredients found in English, Chinese and Japanese recipes. To this end, a cross-language, data-extraction ontology is designed to annotate the recipes, and then augments the annotations with nutrition information on a per-ingredient basis using nutritional information obtained from nutritional databases found in China, Japan and the USA. The relevant nutritional information is used to augment ingredient annotations, which can be used to respond to specific queries related to nutritional goals or to meet specific dietary regimens. The approach meets cross-cultural constraints by incorporating a cultural policy service subsystem to enable responses based on cultural policy ontologies. The approach allows cross-language queries by annotating recipes and parsing queries with a cross-language extraction ontology. For example, responses to an English or Chinese query might be obtained from a Japanese recipe and vice versa.