Aroma evaluation for chili pepper using an E-nose combined with a novel feature fusion technology based on machine learning

文献类型： 外文期刊

第一作者： Yan, Jia

作者： Yan, Jia;Sun, Ruihong;Zhang, Changfu;Li, Zhe;Zhang, Aimin;Wang, Xueya;Chen, Ju;Yin, Yong;Liu, Tao

作者机构：

关键词： Electronic nose; Aroma evaluation; Chili pepper; Variety identification; Feature fusion

期刊名称：JOURNAL OF FOOD COMPOSITION AND ANALYSIS （ 影响因子：4.6； 五年影响因子：4.6 ）

ISSN： 0889-1575

年卷期： 2025 年 145 卷

页码：

收录情况： SCI

摘要： Different varieties of vegetables and fruits have unique flavors, and accurately identifying these varieties can assist companies in facilitating product packaging and transportation and make the distribution and supply of vegetables and fruits more normative, which increases market competitiveness and satisfies diverse consumer demands. In this study, we applied an electronic nose (E-nose) coupled with a novel feature fusion technology, known as maximum separability discriminant correlation analysis (MSDCA), to evaluate the flavors of different pepper varieties effectively. Principal component analysis, linear discriminant analysis, hierarchical clustering analysis and support vector machines were used to evaluate 5 varieties of chili peppers. The proposed MSDCA method introduces discriminative information into canonical correlation analysis while considering both interclass and intraclass dispersion to generate a more representative and distinguishable feature set. This approach achieved 100 % accuracy for pepper variety recognition. The capsaicin content in 5 chili pepper varieties was predicted with high accuracy via principal component regression. With this method, it is possible to reliably identify 5 chili pepper varieties on the basis of flavor assessment and to develop sorting machines according to flavor characteristics, providing technical support for product optimization and increased market competitiveness in the pepper industry.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序