Recently, food safety and cold chain food have been closely related to the epidemic. The party and the state have intensified efforts to solve food safety problems and prevent possible epidemic risks. China has issued a series of policies and plans to strengthen food safety supervision to improve the food safety policy system. To our knowledge, little work has studied policy problems of food safety with in-depth quantitative analysis for an extended period. In accordance with the different national policies and regulations for food safety, this paper fills the gap by analyzing the policies and comparing the central and local policies issued in China from 2007–2022. In addition, the Latent Dirichlet Allocation (LDA) topic model and K-Means clustering model are constructed to analyze the content of food safety policies and identify hot topics. Finally, a quantitative analysis of China’s food safety policies is made from four aspects: the number of policy release years, the distribution area, the range of action, and the affiliated institutions. The results show that: (a) there is a partial surge in food safety policies issued in 2007, 2011, and 2017; (b) the local food safety policy has a high inheritance to the central policy content, and the trends of annual publication number are highly consistent; (c) the innovation of different policy contents in the region have their own characteristics; (d) the proportion of compulsory and capacity policies is much more significant than that of other types of policies. This paper provides some novel insights into food safety policies.

Over recent years, food safety has garnered widespread attention and concern from society. Concurrently, social media sites and online forums have become popular platforms to disseminate news, share opinions, and connect with one’s social network. In this research, we focus on the intersection of food safety and online social networking by utilizing natural language processing techniques and social network analysis to study public opinions related to food safety. Using real data collected from a popular Chinese question-and-answer platform, we first identify hot topics related to food safety, and then analyze the emotional state of users in each community (i.e., users communicating about the same topic) to understand the public’s sentiment related to different food safety topics. We proceed by forming semantic networks to analyze the characteristics of food safety opinion networks. Our results show that Internet users form modular communities, each with differences in topics of concern and emotional states of community users. Users focus on a wide range of topics, showing that overall, food safety awareness is increasing. This paper provides novel insights that can help interested stakeholders monitor the discussions and opinions related to food safety.

Pre-made dishes have drawn growing attention because of their convenience and rapid market expansion. Their food safety risks, however, are shaped not only by products themselves, but also by the gap between public perception, reported incidents, and inspection records. This study develops a three-stage analytical approach by combining Weibo public opinion data, news media reports, and food inspection records from Gansu Province. First, ERNIE and BERTopic are used to identify public sentiment and discussion topics. The results show that negative sentiment slightly exceeds positive sentiment, with school meals, additives, and food safety as the main concerns. Second, 11,110 pre-made dish-related food safety reports from Food Partner Network are clustered and assessed for incident severity. The results point to drug residues in aquatic products, microbial contamination in egg products, authenticity disputes over meat ingredients, and quality issues in frozen composite foods. Third, based on the 2024 official definition, 12,121 inspection records are screened, and 2783 definition-constrained pre-made dish-associated products are retained. Six imbalanced classification models are then constructed. The Weight + RF model performs relatively well for starch and starch products, with a Precision of 0.7857, an AUC-ROC of 0.7778, and an MCC of 0.4429. The study provides a reference for risk identification and inspection resource optimization under limited pre-made dish inspection data.