Fourier transform near‐infrared spectroscopy (FT‐NIRS) was applied to quantitatively determine the macro components (dry matter, protein, carbohydrate, and usable energy) of different cultivated mushroom species. Prediction models were built by partial least‐square regression (PLSR) using 192 mushroom samples. The validation of the models was based on fivefold cross‐validation and test set validation. The root mean square errors of cross‐validation and prediction (RMSECV/RMSEP) were as follows: dry matter 0.36/0.37% w/w, carbohydrate 1.68/1.76% w/w total carbohydrate 1.50/1.50% w/w, and protein 0.47/0.81% w/w. The goodness of cross‐validation (Q²) was over 0.90 in every case. Interval PLSR was applied for model building in case of energy. The RMSECV was 44 kJ/100 g and the Q² was 0.867. Moreover, six species of mushrooms (Agaricus bisporus, Lentinula edodes, Hericium erinaceus, Pleurotus x hybrid, Hypsizygus tessellatus, Lentinula edodes, Hericium erinaceus) were classified correctly using linear discriminant analysis.

中文翻译：

---

近红外光谱法定量测定某些栽培蘑菇的宏成分和分类

傅里叶变换近红外光谱（FT-NIRS）用于定量确定不同栽培蘑菇种类的宏观成分（干物质，蛋白质，碳水化合物和可用能量）。通过使用192个蘑菇样本的偏最小二乘回归（PLSR）建立了预测模型。模型的验证基于五重交叉验证和测试集验证。交叉验证和预测的均方根误差（RMSECV / RMSEP）如下：干物质0.36 / 0.37％w / w，碳水化合物1.68 / 1.76％w / w总碳水化合物1.50 / 1.50％w / w和蛋白质0.47 / 0.81％w / w。在每种情况下，交叉验证的优势（Q ²）均超过0.90。在能源情况下，将区间PLSR应用于模型构建。RMSECV为44 kJ / 100 g，Q²为0.867。此外，六种蘑菇（双孢蘑菇，香菇，猴头，侧耳X混合，真姬tessellatus，香菇，猴头）使用线性判别分析正确分类。