Analyzing the variability of non-destructive boletus from "spatial and chemical" dual-latitude

文献类型： 外文期刊

第一作者： Tian, Jing

作者： Tian, Jing;Li, Jieqing;Tian, Jing;Wang, Yuanzhong;Liu, Honggao

作者机构：

关键词： Boletus; LSTM regression model; Content prediction; Data fusion; Species identification

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 148 卷

页码：

收录情况： SCI

摘要： The looking, smelling and tasting of edible mushrooms are criteria for species identification and quality assessment by consumers. Species identification is mostly carried out through human sensory analysis and traditional chemical methods, which are mostly subjective and cumbersome. In this study, species identification of boletus was carried out by fusing epigenetic features and internal chemical composition in a manner. The results of the study show that the accuracy of the training set and test set is 80.11 % and 87.69 %, respectively. When the partial least squares discriminant analysis(PLS-DA) model is built based on image data, and it can be used as an indicator for quality assessment. When spectroscopy was combined with chemometrics for chemical composition prediction, the Long and short-term memory neural network(LSTM) model predicted the best results, effectively predicting 29 chemical components with a test set R2 of up to 0.9856 and an relative percentage difference (RPD) value of 10.1963. The best model performance R2y and Q2 of 0.9864 and 0.8276 were obtained when fusing image-spectral data to build the PLS-DA model, and the ACC, SPE, SEN all of 100 %. Integrating image and spectral information creates a more comprehensive dataset that can enhance species identification in boletus. New methods and ideas for species information and quality assessment of boletus.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序