化学计量学
电子鼻
食品科学
化学
发酵
感官分析
砖
红外光谱学
计算机科学
色谱法
材料科学
人工智能
有机化学
复合材料
作者
Yan Hu,Wei Chen,Mostafa Gouda,Hailong Yao,Xinxin Zuo,Haixiang Yu,Yuying Zhang,Lejia Ding,Fengle Zhu,Yuefei Wang,Yuanyuan Liu,Jihong Zhou,Yong He
标识
DOI:10.1016/j.foodres.2024.114401
摘要
Fuzhuan brick tea (FBT) fungal fermentation is a key factor in achieving its unique dark color, aroma, and taste. Therefore, it is essential to develop a rapid and reliable method that could assess its quality during FBT fermentation process. This study focused on using electronic nose (e-nose) and spectroscopy combination with sensory evaluations and physicochemical measurements for building machine learning (ML) models of FBT. The results showed that the fused data achieved 100 % accuracy in classifying the FBT fermentation process. The SPA-MLR method was the best prediction model for FBT quality (R2 = 0.95, RMSEP = 0.07, RPD = 4.23), and the fermentation process was visualized. Where, it was effectively detecting the degree of fermentation relationship with the quality characteristics. In conclusion, the current study's novelty comes from the established real-time method that could sensitively detect the unique post-fermentation quality components based on the integration of spectral, and e-nose and ML approaches.
科研通智能强力驱动
Strongly Powered by AbleSci AI