The invention relates to the technical field of
food safety detection, in particular to an
analysis method and
system for identifying the heat treatment degree and
doping of
liquid milk, and the method comprises three steps of sample pretreatment,
mass spectrometry and
data analysis. According to the invention, a
machine learning technology and
mass spectrometric detection are combined to establish a milk identification method and
system with different heating degrees. In this way, a model is established by utilizing a
machine learning
algorithm according to the difference of
polypeptide composition and content change of milk of different heated types in
mass spectrum information, and information of a heat-sensitive
peptide fragment is obtained. Then the model is continuously trained and optimized, a powerful prediction model is screened out, and therefore efficient, stable and accurate identification is achieved. The identification method and
system are scientific and effective. The method established by the invention has the advantages of simplicity and convenience in operation, low
organic solvent consumption, high
throughput, high prediction accuracy and the like.