基于Rosetta的粗糙集理论在糕点食品安全风险监测中的应用

目的以沈阳市2012~2015年糕点食品标准指标监测数据为研究对象,应用Rosetta粗糙集理论进行指标约简,挖掘各指标属性重要性,从而提出更加科学的指标风险监测方案。方法对数据合并分类整理,建立指数分级标准并离散数据生成决策表,使用遗传算法对数据进行约简。结果获得2组数据约简结果,确定核心指标为硫酸铝钾、糖精钠、山梨酸、柠檬黄、胭脂红、菌落总数、大肠菌群和霉菌计数。通过筛选得到18组指标决策规则,经分析硫酸铝钾和菌落总数为关键指标。在2组约简的不合格规则中,硫酸铝钾的覆盖度之和均为0.23,菌落总数分别为0.38和0.36,并且指数等级值均达到6级和7级,提示存在高风险。结论影响糕点食品质量安全的重要指标为硫酸铝钾、糖精钠、山梨酸、柠檬黄、胭脂红、菌落总数、大肠菌群和霉菌计数,建议至少1年监测2次,其中硫酸铝钾和菌落总数为关键监测指标,应给予高度重视,至少1年监测3次。其余指标影响较小,建议1年监测1次。

Application of rough set theory on pastry food safety risk monitoring based on Rosetta

Objective To reduce pastry food standard testing data in Shenyang city from 2012 to 2015 by rough set theory, in order to mine the more important indexes and obtain scientific risk monitoring program. Methods The testing data were obtained and collated, the index classification standard was established, the discrete data decision table was generated, and the data were reduced by genetic algorithm. Results The core indicators that were obtained from the 2 sets of data reduction results were aluminium potassium sulfate, sodium saccharin, sorbic acid, lemon yellow, carmine cochineal, aerobic plate count, coliforms and moulds. Aluminium potassium sulfate and aerobic plate count were the key indexes by analyzing the rules of 18 sets. The coverage values of aluminium potassium sulfate were 0.23, and aerobic plate count were 0.38 and 0.36, respectively, in the substandard rule of the 2 reduction results. They indicated high risk because of the index grades all reached level 6 and 7. Conclusion The important indexes affecting the quality and safety of pastry food are aluminium potassium sulfate, sodium saccharin, sorbic acid, lemon yellow, carmine cochineal, aerobic plate count, coliforms and moulds, and they are recommended to be monitored 2 times every year at least. Aluminium potassium sulfate and aerobic plate count are the key of the important indexes, which must be monitored 3 times every year at least. The others which are not important should be monitored annually.