Relationship Between Total Plate Count And Dielectric Property Of Raw Milk

Total plate count(TPC)is an important index to evaluate the quality of raw milk,and it could directly reflect the food hygienic quality.Excessive amounts of bacteria would influence the nutritive value and properties of dairy products.At present,much of the bulk raw milk without detection is on sale.The hidden security problems can not be ignored.National standard method for detecting TPC is the plate count method,which has high accuracy but low efficency,and is unable to realize on-line inspection or site inspection.Therefore,it's significant to explore rapid detection method of TPC in raw milk,which is helpful to supervise the safety of the dairy industry.To explore the feasibility of detecting TPC in raw milk based on dielectric spectra,the advanced open-ended coaxial probe technique and plate count method were used to measure dielectric properties(dielectric constant ?' and dielectric loss factor ?'')and TPC of cow's milk and goat's milk,respectively.The variation trend of TPC and the influence of TPC on dielectric properties during storage were analyzed.The relationships between dielectric properties and TPC of large number of raw milk samples were studied by building qualitative and quantitative models.The results are as follows:(1)Relationship between TPC and dielectric property of raw milk was studied.With the increase of storage time,the TPC of cow's milk increased but the pH values decreased,and ?' decreased at over 100 MHz,and ?'' increased over the frequency range of 20~800 MHz.There were good linear correlations between dielectric parameters(?' and ?'')and pH values and TPC at certain frequencies during storage at 4 ?.(2)Quantitative models were established for predicting TPC in cow's milk based on dielectric spectra.The sample sets were partitioned by Sample Set Partitioning based on Joint X-Y Distances(SPXY).Partial Least Squares(PLS),Extreme Learning Machine(ELM)and Least Square Support Vector Machine(LSSVM)were used for modeling based on full frequencies(FF)and the selected optimal characteristic dielectric parameters by Successive Projections Algorithm(SPA).The results showed that all the correlation coefficients(Rc and Rp)were less than 0.8,and all the root mean square errors of calibration set were above 1.0 lgCFU/mL,and all the residual prediction deviation(RPD)were less than 1.50 except FF-LSSVM model.(3)Quantitative models were set up for predicting TPC in goat's milk based on dielectric spectra.The sample sets were partitioned by SPXY.PLS,ELM and LSSVM were used for modeling based on FF and the selected optimal characteristic dielectric parameters by SPA.The results showed that FF-PLS and FF-LSSVM models had the higher correlation coefficients and lower root mean square error.The RPD value of FF-PLS was greater than 2.0,and it indicated that TPC of goat's milk could be predicted roughly.(4)Qualitative models were set up for identifying the cow's milk samples with TPC exceeding standard based on dielectric spectra.The sample set was partitioned by Kennard-Stone(KS).PLS,Back Propagation(BP)network and Support Vector Machine(SVM)were used for modeling based on FF and the selected optimal characteristic dielectric parameters by SPA.The results showed that the recognition rate of FF-BP model was better than others,and the recognition rates of calibration set and prediction set were 86.0% and 94.4%.(5)Qualitative models were set up for identifying the goat's milk samples with TPC exceeding standard based on dielectric spectra.The sample sets were partitioned by KS.PLS,BP and SVM were used for modeling based on FF and the selected optimal characteristic dielectric parameters by SPA.The results showed that the recognition rate of FF-BP model was better than others,and the recognition rates of calibration set and prediction set were 96.1% and 90.4%.The study results indicated that dielectric properties and TPC of raw milk had correlation to some extent,and it's feasible to carry out qualitative analysis while quantitative analysis remained to be further research based on large number of raw milk samples.This study provided experiment and data foundation for exploring detection technology of TPC in raw milk based on dielectric spectra.