Quantitatively determining the total bacterial count of raw goat milk using dielectric spectra

摘要

The objective of this study was to evaluate dielectricspectra as a means of quantitatively determining totalbacterial count (TBC) of raw goat milk. The dielectricspectra, including dielectric constant (ε′) spectra anddielectric loss factor (ε″) spectra, and TBC of 154 rawgoat milk samples were measured using network analyzerand plate count methods, respectively. Owing tothe poor linear relationship between TBC in logarithmand permittivities at a single frequency, chemometricswas used to reduce noise, identify outliers, select effectivevariables, and divide sample sets. Several linearmodels, such as multiple linear regression, ridgeregression, and least absolute shrinkage and selectionoperator, were established to determine TBC based onthe effective spectra of ε′, ε″, and their combination(ε′+ε″). The results indicated that the models builtusing the spectra of ε′+ε″ and ε′ had excellent TBCprediction performance. The best model was multiplelinear regression based on ε′+ε″ spectra with the residualpredictive deviation of 3.26. This study showsthat the dielectric spectra had great potential to quantitativelyand rapidly determine TBC of raw milk.