Research On The Discrimination Model And Quality Traceability System Of Rice Origin

As the old saying “Hunger breeds discontentment”,rice plays an important role in the national economy and people's livelihood because it is one of the main food sources.Heilongjiang Province,as one of the main producing areas of rice,is civilized because of the high quality of the rice.However,due to the frequent occurrence of rice food safety incidents such as “yellow grain rice”,“poisonous rice” and “adulterated rice”,not only the health of the consumer group but also the development of the rice industry has been affected.Therefore,people today are all concerning about food quality and safety,under this serious situation,it is an urgent priority to study on the quality traceability and origin discrimination system of rice with the characteristics of Heilongjiang Province to protect the legitimate rights and interests of consumers and the relevant interests of producers,and improve the brand effect for the development of rice industry in Heilongjiang Province.In this paper,it is the research goal to improve the discriminant classification systems based on the discrimination rate of the production area.Some algorithms,such as the principal component analysis,KNN discrimination,Fisher discriminant etc.,are used to achieve dimensionality and discriminant analysis of mineral elements in rice.After a full survey on consumer demand and rice enterprises in Heilongjiang Province.a Web-based system of rice quality traceability and origin discrimination was established under some development technologies such as,B/S 3-layer architecture,Mysql database,PHP to achieve an efficient and scientific traceability and a rapid identification.The mineral element data discrimination and analyzes were performed using MATLAB.It was shown that the positive discriminant rate of Fisher discriminant is higher than KNN discriminant method,linear regression and principal component regression analysis;discriminant analysis based on R-type clustering,an optimal scheme was selected from 864 element combination schemes,which can be discriminated by 23 mineral elements.The Fisher discriminant rate of 86.76% was slightly lower than that of the positive rate of 91.78% for the direct classification on original data.Moreover,the data analyses for the two years of 2015 and 2016 were carried out.Because the content of mineral elements is affected by natural factors,the discrimination results in 2015 are better than those in 2016.By analyzing 38 mineral elements from Fangzheng,Wuchang,Chahayang,Jiansanjiang and Xiangshui rice,and the discriminative models established by 8 kinds of mineral elements such as Mg,Al,K,V,Co,Rb,Cs and Er are gradually selected.The overall correct discriminant rate of the rice origin is 94.5%,It is indicated that these eight mineral elements carry sufficient identification information of the origin,and the traceability of the five rice producing areas of the other side has application value.At present,the system has been applied in Beidahuang Rice Industry Tailai Food Technology Corporation,Heilongjiang Province.By using this system,the production management level of the enterprise was improved and provide a convenient and effective platform for consumers to trace the original places of rice.