| Food safety is a critical issue concerning the welfare of the nation. Thus, the responsibility to improve food safety risk assessment and regulation lies with the food safety supervision departments. Subject to the limitations of traditional mathematical statistical methods, food safety supervision departments, however, can only analyze the existing data obtained from sample inspection. Such analysis fails to present the links between data, which results in weak prediction and inadequate measures in prevention and early warning against food safety threats. In order to address the issue, this thesis discuses the feasibility of applying data mining techniques in food safety assessment with examples.Started with the research background, significance and the current situation at home and abroad, the thesis presents the basic knowledge of food safety and association rules of data mining. The followings are discussed in the thesis:(1) Due to the discrete feature of food safety inspection data, priority should be given to the connections between data and to apply Association Rule to initialize data mining.(2) Using the food sample inspection report as original data and to select the original data against the results and the nature of the inspection; starting initial process of the selected data, such as data cleaning, so as to get it ready for the mining tool.(3) Discussing the design of the database and function module in food safety risk assessment system and presenting the structure of the database and function module.(4) Due to the data volume of the food safety risk monitoring, applying SAS EM (Enterprise Miner) to analyze and mine the data against Association Rule. Filtering the mining results against appropriate descriptions for more accurate and concise data.(5) Conducting data mining against Association Rule in risk assessment for rice products. By analyzing the mining result, data mining technique is proved to be applied in food safety risk assessment. |
PDF Full Text Request