Effect of moisture variation on the robustness of NIR spectroscopy-based calibration models
NIR spectroscopy is well-known for its sensitivity to water, which can be useful for detecting water variation in the sample matrix. However, for many other applications, water intake of the samples may be an issue and must be closely monitored to gauge the predictive performance of calibration models. This article looks at how moisture fluctuations in drug products affect the performance of classification models. To achieve this, we built hit quality index (HQI) and principal component analysis (PCA) identification models on pharmaceutical finished products to see how humidity variations affect the robustness of these models.
While near-infrared spectroscopy (NIRS) is a highly valued tool with many applications in industry due to its fast and noninvasive nature, a significant drawback is its extreme susceptibility to humidity changes in the environment. In this study, we evaluate the influence of humidity variation on the predictive performance of NIR‑based multivariate calibration models.