Method for classifying characteristics of base wine fingerprint

Abstract

The invention provides a method for classifying characteristics of base wine fingerprint, which relates to the related fields of food field, fermentation field, information processing, pattern recognition and the like. The liquor base liquor fingerprint classification method of the invention combines the principal component analysis and the neural network method to classify the liquor base liquorfingerprint. The invention firstly screens the important components of the base liquor through the principal component analysis, then the screened important features are used for the neural network weight training, and finally the base liquor is tested with the real base liquor spectrum data to realize the classification of the base liquor. The invention can better overcome the influence of unimportant components of redundant data in the spectrum of liquor base liquor on the classification result, greatly reduces the input amount of the neural network and the calculation amount of the classification, and the classification result can better reflect the difference in the characteristics of the liquor base.

Description

technical field [0001] The invention relates to related fields such as information and data processing, pattern recognition and classification, and in particular to a classification method based on fingerprint features of liquor base wine. Background technique [0002] The fingerprints of liquor-based liquor are as unique as human fingerprints, which can show the distinct specificity and typicality of different substances. It is a relatively conventional comprehensive evaluation method to analyze the base wine by its fingerprint. It has the characteristics of high sampling accuracy, obvious features, strong specificity, good reproducibility, fast speed and strong recognition ability. Some unique characteristics of liquor (such as: aroma type, origin, etc.) determine the overall similarity of liquor with the same aroma type and the same origin, which determines the applicability and importance of fingerprints in liquor classification research. Extensive. Liquor is a mixtur...

AI Technical Summary

Problems solved by technology

However, with the continuous improvement of monitoring technology, many trace components of liquor have been determined. If all the trace components are used for neural network weight training, more input units and training time are required.

At the same time, for wine samples of the same flavor type and grade, the content of some trace components is also quite different, which will lead to inaccurate training weights, resulting in inaccurate final recognition and classification

Images