Use of Raman spectroscopy to screen diabetes mellitus with machine learning tools

• DOI: 10.1364/boe.9.004998
• Publication Date: 2018-09-26T20:33:49Z
• Authors: Edgar Guevara; Juan Carlos Torre...; Miguel G. Ramírez...; Claudia Luevano-C...; Francisco Javier ...
• Keywords: 6. Clean water; 3. Good health

Type 2 diabetes mellitus (DM2) is one of the most widely prevalent diseases worldwide and currently screened by invasive techniques based on enzymatic assays that measure plasma glucose concentration in a laboratory setting. A promising plan action for screening DM2 to identify molecular signatures non-invasive fashion. This work describes application portable Raman spectroscopy coupled with several supervised machine-learning techniques, discern between diabetic patients healthy controls (Ctrl), high degree accuracy. Using artificial neural networks (ANN), we accurately discriminated Ctrl groups 88.9-90.9% accuracy, depending sampling site. In order compare ANN performance more traditional methods used spectroscopy, principal component analysis (PCA) was carried out. subset features from PCA generate support vector machine (SVM) model, albeit decreased accuracy (76.0-82.5%). The 10-fold cross-validation model performed validate both classifiers. technique relatively low-cost, harmless, simple comfortable patient, yielding rapid diagnosis. Furthermore, ANN-based method better than typical measurement capillary blood glucose. These characteristics make our tool identifying automated