Quantitative analysis of blood glucose by FT‐Raman spectroscopy and multivariate statistical analysis

Diabetes is one of the common chronic diseases in worldwide and increasing among young children. Testing glycated hemoglobin levels as well as blood glucose levels is a method for detecting diabetes. In this article, with multivariate statistical analysis was used to monitor the blood glucose levels by the Fourier Transform Raman spectroscopy. Principal component regression with direct orthogonal signal correction is developed and applied to analyze the glucose concentration with the blood Raman spectroscopy. The root means square error of calibration is 3.0824 mg/dL (0.1712 mmol/L), the root means square error of prediction is 3.2688 mg/dL (0.1816 mmol/L), the correlation coefficients (R ²) is.9946 and ratio of performance to deviation is 6.4789 and, respectively. The results show that this method can be used as a strong potential diagnostic tool for diabetes mellitus.