The Research On The System Of Raw Milk Quality Evaluation And Information Management System

Quality problems of dairy products occur frequently in recent years, which havearoused people’s high attention. To ensure the safety of dairy products and improvetheir quality, it becomes especially important to appraise the quality of raw milkscientifically. However, nowadays there are few researches on appraisals of raw milkquality and basic information management system. In this paper,41129raw milksamples of year2009from134pastures of Shanghai as well as its surrounding areaswere analysised. Raw milk standard systems both in China and abroad werecompared. Raw milk quality appraisal index system was established using analytichierarchy process method. A survey on the quality variation rules and the influencingfactors of the raw milk was conducted applying the correlation analysis, the pathanalysis and the multiple regression analysis. The classification of the raw milkquality and its judgment model were constructed by the cluster analysis, factoranalysis and the discriminant analysis. Meanwhile, the reliability of classificationmodel was proved by decision tree method. Raw milk quality database andinformation management system was developed using SQL Server andC/S(Client/Server). The main results are as following:1.The raw milk standard system at home and abroad were compared.The resultsindicated that there were noticeable changes on technical indexes of both the newnational standard and the old.The differences between standard of China and that ofsome organizations related dairy products like CAC、IDF and of developed countrieswere obvious in the aspects of standard structure, standard quantity, qualitygrading,price determined.2.The technical route and principles of raw milk quality appraisal index wereconfirmed. Raw milk quality appraisal index system structure was established andthe weight of every index was determined by analytic hierarchy process.5itemshave greater weight. They were protein0.192, fat0.185, total bacterial count0.183,somatic cell0.159and ice point0.151. Combined with the present situation of thedomestic and foreign relevant research,protein,fat,ice point,the aerobic plate countand somatic cell count were finally determined as the quality evaluation index.3. The quality variation rules and the influencing factors of the raw milk wereanalysised. The result showed that somatic cell count and the aerobic plate count hadan obvious change. There was an ostensible positive relation between the fat contentand the protein content, the protein content and the freezing points,the aerobic platecount and the average temperature,while the relation between somatic cell count andthe average temperature was negative. Meanwhile, a model was establishedconcerning the aerobic plate, the temperature, the fat,the protein,the humidity in theraw milk which was displayed as Y=7.5920.465X1-4.029X2-3.494X3-0.182X4. Itscoefficients of determination was R2=0.716.4.The grade of the raw milk quality was studied by the cluster analysis, factoranalysis and discriminant analysis. The result indicated that the raw milk could be classified as below four grades: excellent, good, medium and poor. A mathematicmodel of raw milk quality classification with93.7percent accuracy was constructedby discriminant analysis based on the classified results, and the decision tree methodwith98.6percent accuracy was constructed between quality level and five indexes.The two models together proved the rationality of classification in two ways.5.The database which contains the relevant standards and regulations of dairyproducts,the regular testing programs,the index of the raw milk quality,the cowfarmlands,forage grass,the index of the additives,the quality of the cows and otherbasic information was constructed by SQL Server and C/S(Client/Server). The rawmilk quality information management system was constructed by using the C/S.Some of the concrete functions could be achieved such as entering the data,therevisions,the deletion,the search,the analysis,the printing. The classification processof raw milk quality was simplified with the classification model. This system hasgreat utility value and can meet different needs of users for research as well asmanagement.