Universidade Tecnológica Federal Do Paraná (UTFPR), 87301-899, Campo Mourão, Paraná, Brazil.
Universidade Tecnológica Federal Do Paraná (UTFPR), 87301-899, Campo Mourão, Paraná, Brazil.
Talanta. 2021 Jan 15;222:121526. doi: 10.1016/j.talanta.2020.121526. Epub 2020 Aug 15.
Professional cupping is a reliable methodology for the coffee industry and its professionals. However, it faces barriers for its implementation on an industrial scale. To date, no study has determined a coffee cup profile using a handheld near-infrared (NIR) spectrometer. Therefore, the aim of this study was to evaluate directly cup profiles in roasted and ground coffee blends via handheld NIR spectroscopy combined with partial least squares with discriminant analysis (PLS-DA), in an industrial case study. The sensitivity and specificity of the model obtained ranged from 91 to 100%, 84-100%, and 73-95% in the training, prediction, and internal cross-validation sets, respectively. These results are therefore promising for the industrial reality and the methodology could assist coffee professionals in their decisions during cup evaluation in further tests. The proposed method is viable for the direct determination of cup profile at industrial scale since it is portable, fast, simple, robust, and less expensive compared to the benchtops equipment.
专业的咖啡杯测是咖啡行业及其专业人士的可靠方法。然而，它在工业规模上的实施面临障碍。迄今为止，尚无研究使用手持式近红外（NIR）光谱仪确定咖啡杯型。因此，本研究的目的是通过手持式近红外光谱结合偏最小二乘判别分析（PLS-DA），在工业案例研究中直接评估烘焙和研磨咖啡混合物的杯型。该模型在训练集、预测集和内部交叉验证集中获得的灵敏度和特异性分别为 91%至 100%、84%至 100%和 73%至 95%。因此，这些结果对工业现实具有很大的希望，该方法可以帮助咖啡专业人士在进一步的测试中进行杯测评估时做出决策。与台式设备相比，该方法具有便携、快速、简单、稳健和更经济的特点，因此适用于工业规模的直接杯型确定。
