Cigarette internal quality index extimating method based on regression function estimating SVM

Abstract

The present invention provides a method for assessing the inner quality index of cigarettes based on regression function estimation SVM. Based on the sample data of cigarettes, an SVM model estimated by regression function is established, and then the inner quality of the cigarette to be tested is evaluated by using the above SVM model. The present invention utilizes the correlation analysis method to select characteristic parameters, reduce the number of input parameters, thereby further reducing the demand for sample data, and utilize the excellent characteristics of SVM learning in the case of limited samples, in the case of small samples It can also achieve relatively high evaluation and prediction accuracy, thereby reducing the amount of manual smoking evaluation and reducing the testing cost of enterprises, and is very suitable for popularization and application in the tobacco industry.

Description

technical field [0001] The invention relates to a support vector machine method for more accurately realizing the evaluation and prediction of the inner quality index of the cigarette under the condition of limited samples, in particular to an evaluation method of the inner quality index of the cigarette by using the SVM model estimated by the regression function. Background technique [0002] The internal quality of cigarettes, such as sensory quality, is closely related to certain physical and chemical components in cigarettes. These physical and chemical components determine the internal quality of cigarettes to a certain extent, such as nicotine in tobacco and sensory stimulation. Feature related. Domain experts often need to refer to these detected components when performing sensory evaluations. [0003] Tobacco leaves contain many chemical components; the interaction of various chemical components in the smoking process stimulates people's senses of taste, smell, and ...

AI Technical Summary

Problems solved by technology

[0007] In order to overcome the defects in the prior art, the present invention provides a method for evaluating the internal quality index of cigarettes based on regression function estimation SVM, which combines technologies and methods such as correlation analysis, regression function estimation SVM, and evaluation output value conversion rules. Combining and improving the algorithm and key parameters accordingly to make it more suitable for the data environment of the cigarette industry; its main purpose is to establish a machine learning model suitable for the small sample data environment of the cigarette industry, fully adapting to small samples, high dimensions, large noise, The four important data characteristics of non-linearity; thereby reducing the demand for sample data, reducing the amount of manual smoking evaluation, reducing the testing cost of enterprises, and achieving a relatively high degree of evaluation and prediction compliance under the condition of valid samples, meeting the needs of computer-aided tobacco industry The need for quality assessment; at the same time, it also provides a good prerequisite technical method for the intelligent design of cigarette leaf group formulations

Images