Journal of IiME Volume 1 Issue 2 www.investinme.org Visible and near-infrared (Vis-NIR) spectroscopy: Introduction and Perspectives for Diagnosis of CFS Hb: deoxyhemoglobin Table 2. Representative clinical applications of Vis-NIR spectroscopy HbO2: oxyhemoglobin HIV: Human immunodeficiency virus PCA: principal component analysis PLS: partial least squares regression analysis gives extra weight to variables that show a high correlation with y variables. Therefore, PLS is usually more effective for predictions than PCR. Further detailed illustrations and mathematical formulas of algorithms are available in many reports about chemometrics [82, 83]. In a multivariate analysis, the number of PCs is important, because too few or too many PCs distort signals or diminish the signal-to-noise ratio, respectively. To choose the correct number of PCs, a validation step is usually included in the process of modeling [9]. For validation steps, internal validation or external validation is used. Most chemometrics software programs include internal cross validation. In internal cross validation, the sample set is repeatedly divided into two groups. One group is reserved for validation and the other, for calibration. This process is repeated until all groups have been used for validation once. In external validation, sample sets are first separated into calibration samples and test samples, which are subjected to validation and used for assessment of the calibration model. By finding the number of PCs when the model shows a minimum standard error of validation (SECV), the number of PCs can be used to describe the signal in the data. Invest in ME Charity Nr 1114035 These results suggest that combining Vis-NIR spectroscopy with chemometrics is a promising way to objectively diagnose CFS. They also suggest that an unknown factor or factors present in the serum of all CFS patients could provide important clues as to the agent causing this debilitating disease. Recently, commercially available chemometrics software programs such as Pirouette (Infometrics, Woodinville, Washington, USA) and Unscrambler (CAMO Inc., Woodbridge, New Jersey, USA) have been used for Vis-NIR analyse. The number of manufacturers of these software programs is increasing. The programs and their manufacturers are listed in Table 3. The software programs are designed to analyze spectral data, and because preprocessing such as standard normal variate (SNV) [84] and smoothing [85], which minimize differences between spectra caused by baseline shifts and noise, is carried out during the analysis, pre-processing handling, which is time(continued on page 12) Page 11/72 LDA: linear discriminant analysis MELAS: myopathy, encephalopathy, lactic acidosis, MERRF: myoclonic epilepsy with ragged red fibers and stroke-like episodes PCRA: principal component regression analysis
12 Publizr Home