Near infrared spectroscopy coupled with sample set partitioning based on joint X-Y distances combined with partial least square regression was applied to the quantitative analysis of six routine chemicals, five physical indices and four macromolecular substances in reconstituted tobacco. The quantitative regression models of these indices were established by joint X-Y distances combined with partial least square regression. Results showed remarkable correlation between predicted and measured values of the 15 indices. The root mean square error of prediction of all the indices was low, and the correlation coefficients of these PLS models were all greater than 0.85. This was the first study in which NIR spectroscopy had been used to determine the macromolecular substances as well as certain physical indices in reconstituted tobacco. Results showed that this method could be feasibly applied for rapid detection of these properties of industrial products.

中文翻译：

---

近红外光谱结合样品组划分法测定再造烟草中的常规化学物质、物理指标和大分子物质

将近红外光谱结合基于联合XY距离的样本集划分结合偏最小二乘回归应用于再造烟草中6种常规化学物质、5种物理指标和4种大分子物质的定量分析。通过联合XY距离结合偏最小二乘回归建立这些指标的定量回归模型。结果表明，15 个指标的预测值与实测值之间存在显着相关性。所有指标的预测均方根误差较低，这些PLS模型的相关系数均大于0.85。这是第一次使用近红外光谱来确定再造烟草中的大分子物质以及某些物理指标的研究。