Text-based Research of Early Warning Platform from Food Complaint Texts

Abstract

This study proposes a food complaint text early warning method based on the guidance of ontology and establishes a scientific and reasonable system of early warning, builds and improves the food security early warning platform. All of those make this study play a supplementary role in the research content of food safety regulators. Based on traditional early warning system, this study constructs food safety complaints warning platform model and builds the food domain ontology and expands food complaint document semantics to highlight the implicit semantics and improve the document's semantic accuracy. Through the calculation of similarity of theme characteristic vector and text categorization constructing classifier, make the automatic classification of food complaint documents based on the theme come true for those which are not correctly classified documents for unsupervised clustering, which can be the purpose of food safety alarm. Then, it is possible to use complaints about food safety for rapid and accurate text data processing, make the food safety regulators understand the food safety hidden trouble in time to protect consumers' rights and interests.