Method and system for differentiation of tea type

A tea classification system, comprising:

• a sampling module, configured to obtain tea mass spectrometric data corresponding to tea to be differentiated by using an LC-MS (Liquid Chromatography-Mass Spectrometry) technology;
• a classification module, configured to build a differentiation function by using ironic intensities of 20 characteristic compounds as evaluation indexes to perform classification processing on the obtained tea mass spectrometric data for obtaining a classification result of the tea to be differentiated, mass-to-charge ratios of the 20 compounds being as follows:
• 116.0648-116.0764, 267.1206-267.1474, 268.0906-268.1174, 280.1252-280.1532, 289.0561-289.085, 307.0657-307.0964, 308.0757-308.1065, 309.0814-309.1123, 364.0819-364.1183, 381.0604-381.0985, 425.0658-425.1083, 485.0833-485.1318, 518.2984-518.3503, 537.2765-537.3302, 554.1509-554.2063, 579.1207-579.1786, 607.2611-607.3218, 677.3378-677.4055, 744.2234-744.2978, and 869.1124-869.1993;
• a model building module configured to build tea classification models, wherein the model building module specifically comprises:
• a model building data obtaining submodule, configured to obtain sample tea mass spectrometric data corresponding to different types of sample tea and use a data set formed by the obtained sample tea mass spectrometric data as a sample tea mass spectrometric data set;
• a model building processing submodule, configured to randomly divide the obtained sample tea mass spectrometric data into a training set and a validation set, and perform model building processing on the training set by using a random forest method, a support vector machine method or a Fisher differentiation method to build and obtain tea type differentiation models;
• and
• a validation submodule, configured to validate a random forest model by using the validation set.

A tea type differentiation method, comprising:

• building a differentiation function by using ionic strengths of 20 compounds as evaluation indexes to differentiate tea types, wherein mass-to-charge ratios of the 20 compounds being as follows:
• 116.0648-116.0764, 267.1206-267.1474, 268.0906-268.1174, 280.1252-280.1532, 289.0561-289.085, 307.0657-307.0964, 308.0757-308.1065, 309.0814-309.1123, 364.0819-364.1183, 381.0604-381.0985, 425.0658-425.1083, 485.0833-485.1318, 518.2984-518.3503, 537.2765-537.3302, 554.1509-554.2063, 579.1207-579.1786, 607.2611-607.3218, 677.3378-677.4055, 744.2234-744.2978, and 869.1124-869.1993;
• wherein the building different tea type differentiation models further comprises:
• using data of the internal standard normalization ionic response intensity values of 20 compounds of different types of collected tea samples between the set mass-to-charge ratios as a data set of sample tea;
• and randomly dividing the data set of the sample tea into a training set and a validation set, wherein data of the training set is used for building the tea type differentiation models, data of the validation set is used for validating the built tea type differentiation models, and a ratio of the quantity of the sample tea samples of each tea type to the quantity of the tea samples to be differentiated is not smaller than 3:1.