Research On Pork Source-Tracing And Its Price Early Warning Model Based On The Internet Of Things

Food safety is not only critical to public health and safety, but also closely related to economic development and social stability. Our government has always been dedicating herself to the protection of people's health and life safety, makes it a basic national policy, and attaches great importance to the supervision and management of the food safety. Comprehensively improve the quality and safety of agricultural products has become a global strategic mission. This thesis considers the pork products safety problem which comes from the livelihood business that is very closely related and extremely important to our people's daily life, we study on the backward source-tracing model and the price precaution model of the pork products from the perspective of the Internet of Things. The major research topics of this thesis include: food source-tracing process control scheme and management system; food source-tracing data collection, transport and storage; high-performance computation and large-scale data backup; precaution of the pork price risks, etc. We developed two core software systems, one for pork products safety backward source-tracing, another for multi-authorities cooperating supervision of the food safety problems. Based on these two systems, we finally set up a demonstrating application system for pork products backward source-tracing, which can be adapted to many practical scenarios of the Internet of Things. In summary, the major achievements of this thesis are outlined as follows:Firstly, we proposed a novel SVM-based pork price precaution model. This model was built upon the support vector machine theory, by making use of the pig feed price ratio as predictors, it can provide analysis and precaution support on risk levels of the pork prices under small sample conditions. Empirical studies show that the proposed model is adapted to small sample analysis, and can perform risk level precaution on pork prices with high accuracy, so as to provide necessary information for government decision-making. These research results offer us a new way to predict price risks of the pork products under limited sample conditions--which was typical in practice.Secondly, by using of the Panel Data Model, we proposed a novel approach for analyzing on the impact factors and regional characteristics of the sale prices of the pork products. We then perform empirical study on the public dataset issued by the Department of Commerce of our country, it was collected from thirty two regions of our country through a statistical monitoring system for livestock and poultry slaughter, which consists of many characteristics including the purchase price, the slaughter number, the production and sale numbers, and the ex-factory price of carcass meat, etc. According to our analysis, the sales price of the pork products of our country has significant positive correlation with the purchase price of live pigs. Except for Gui-Zhou province and construction corps areas, the sales price changes are not sensitive to the changes of the slaughter numbers, the production and the sale numbers in other areas.Thirdly, we proposed a practical backward food source-tracing model based on the Internet of Things application scenario, a finite state automaton model for food backward source-tracing, and a clustering analysis approach for time series analysis of the food trace data. The RFID based food source-tracing application model was integrated with a variety of information technology, through conversion, fusion and data mining techniques on a large number heterogeneous information sources, make it practical to manage the whole food trace system. This model also provides a new scheme for our government to efficiently supervise the food safety. The finite state automaton model for food traceability was responsible for analyzing and simulation of the state transition process, so as to make the process more rational and efficient. The data generated by the supply chain and backward source trace procedure were typically a huge amount of time series data, our clustering analysis approach treated it as a complete functional object with respect to time, and then perform data analysis on it to provide the basis for decision making.Finally, we build up a demonstrating application system for pork products backward source-tracing based on the Internet of Things, and this is an innovation for food safety supervision and management. According to the characteristics of the food safety supervision of our government, by using of the internet of things techniques to combine the information from pork products supply chains, we set up a joint supervision scheme between government bodies and a pork products supply chain management system, thus formed a demonstration system and application mode of the internet of things techniques for the food safety industry. This approach makes the food products traceable, brings us an innovative approach for food safety supervision and regulation, changes the way monitoring supply chain of the pork products from local independent monitoring to collaborative monitoring during the whole process; shares the supervision responsibility with communal participation; changes the supervision work mode from single way to duplex way; changes the work mode from human supervision to technical supervision; changes the manner of the government management from extensive to fine-grit supervision. Detailed, accurate, and real-time source-tracing information makes sure that once occurring a food safety event, it can be traced back to the source in the first moment, and this will greatly promote the supervisory capacity and level of our government in public health and safety area.