Quantitative prediction method of clear flavor of cigarettes

Abstract

The invention discloses a quantitative prediction method of a clear flavor of cigarettes based on detected contents of 23 volatile components in the cigarettes. The quantitative prediction method comprises the following steps: firstly, measuring the contents of the 23 volatile components in cut tobacco according to a sample treatment method and a GC-MS / MS (Gas Chromatography-Tandem Mass Spectrometry) analytical method which are independently developed and established; secondly, substituting the measured contents into a regression equation to calculate a value of Y clear flavor, thus obtaining a predictive value of a score of the quantitative prediction method of the clear flavor of the cigarettes, wherein the range of the predictive value is 0 to 10 scores. The fit goodness between the quantitative prediction method of the clear flavor of the cigarettes, disclosed by the invention, and a sensory actual smoking result is better, and a predicted mean absolute error is 0.1007 score. The quantitative prediction method provides favorable technical support for maintaining the stability of sensory quality of the cigarettes.

Description

technical field [0001] The invention belongs to the field of quantitative prediction and control of the fragrance type of cigarettes, and in particular relates to a method for quantitatively predicting the fragrance type of cigarettes based on the detection content of 23 volatile components in shredded tobacco of cigarettes and the method of constructing a regression model in combination with regression coefficients. Background technique [0002] The key to the stability and improvement of cigarette product quality lies in the stability and improvement of sensory quality. How to establish an evaluation method for cigarette sensory quality has always been a research hotspot in the industry. At present, the evaluation of the sensory quality of cigarettes is mainly carried out through the sensory evaluation of experts, and relevant national and industry standards have been issued and implemented. For example, GB 5606.4-2005 "Cigarette Part 4: Sensory Technical Requirements" and...

AI Technical Summary

Problems solved by technology

For example, Mo Hongwei from Harbin Engineering University provided a method for predicting cigarette sensory evaluation and smoke index through immune neural network algorithm (CN 101419209 A), but the ingredients included in the calculation modeling are only total sugar, More than 10 conventional chemical composition indicators such as total nicotine and reducing sugar, which cannot represent the overall sensory characteristics of cigarettes

Hu Shengguo of Yizhong Tobacco (Group) Co., Ltd. invented a BP neural network model combined with fuzzy set method, and used the model to predict the sensory evaluation and smoke index of cigarettes (CN 1525394 A), but it is used for There are more than 110 chemical components to be modeled and calculated, and the qualitative and quantitative index detection is cumbersome and difficult to calculate

This leads to disadvantages such as low or unreliable prediction results.

Images