Method and apparatus for enhanced estimation of an analyte property through multiple region transformation
View Patent ↗The invention comprises transformation of a section of a data block independently of the transformation of separate or overlapping data blocks to determine a property related to the original matrix, where each of the separate or overlapping data blocks are derived from an original data matrix. The transformation enhances parameters of a first data block over a given region of an axis of the data matrix, such as signal-to-noise, without affecting analysis of a second data block derived from the data matrix. This allows for enhancement of analysis of an analyte property, such as concentration, represented within the original data matrix. A separate decomposition and factor selection for each selected data matrix is performed with subsequent score matrix concatenization. The combined score matrix is used to generate a model that is subsequently used to estimate a property, such as concentration represented in the original data matrix.
1. A method for developing a calibration score matrix, comprising the steps of:
providing a matrix of calibration spectra;
generating N spectral matrices corresponding with N spectral regions of said calibration spectra, wherein N is an integer greater than or equal to two and wherein at least two of said N spectral regions are non-identical spectral regions;
independently decomposing each of said N spectral matrices to yield N score matrices;
independently selecting a plurality of factors for each of said N score matrices;
concatenating said selected number of factors for each of said N score matrices, to generate a calibration score matrix, and
estimating, with a spectrometer, a target property from a prediction matrix using said calibration score matrix, wherein said spectrometer comprises a base module, a communication bundle, and a sample module, wherein said communication bundle integrates said sampling module and said base module.
2. The method of claim 1 , wherein N consists of any of the integers 2, 3, 4, and 5.
3. The method of claim 1 , wherein said step of independently decomposing employs multivariate techniques.
4. The method of claim 3 , wherein said multivariate techniques comprise any of:
principal component analysis;
partial least squares;
wavelet regression;
Fourier series;
Taylor series; and
factor analysis.
5. The method of claim 1 , wherein at least two of said N spectral regions overlap.
6. The method of claim 1 , wherein at least two of said N spectral regions are discontinuous regions.
7. The method of claim 1 , wherein at least two of said N spectral regions abut.
8. The method of claim 1 , further comprising the step of developing a calibration with said calibration score matrix.
9. The method of claim 8 , further comprising the step of estimating said target property from an estimation spectrum.
10. The method of claim 1 , further comprising the step of independently preprocessing each of said generated N spectral matrices prior to said step of decomposing.
11. The method of claim 10 , further comprising the step of developing a calibration with said calibration score matrix.
12. The method of claim 11 , wherein said calibration spectra comprise noninvasively obtained near-infrared spectra.
13. The method of claim 12 , wherein at least two of said N spectral regions comprise regions within the first and second overtone spectral region.
14. The method of claim 13 , further comprising the step of estimating a glucose concentration from a noninvasive spectrum.
15. A method for estimating a property from a noninvasive spectrum, comprising the steps of:
providing a matrix of calibration spectra;
generating N spectral matrices corresponding to N spectral regions of said calibration spectra, wherein N is an integer greater than or equal to two and wherein at least two of said N spectral regions are non-identical spectral regions;
independently decomposing each of said N spectral matrices yielding N score matrices;
independently selecting a plurality of factors to represent each of said N score matrices;
concatenating said selected number of factors for each of said N score matrices to generate a calibration score matrix;
generating a calibration model using said calibration score matrix; and
estimating, with a spectrometer, a target property represented by said noninvasive spectrum, wherein said spectrometer comprises a base module, a communication bundle, and a sample module, wherein said communication bundle integrates said sampling module and said base module.
16. A method for developing a calibration for estimating a target property, comprising the steps of:
providing a first calibration data block;
generating N data blocks from said calibration data block, wherein N is an integer greater than or equal to two;
transforming each of said N selected data blocks;
combining said N transformed data blocks to form a second calibration matrix; and
developing a calibration with said second calibration matrix; and
estimating, with a spectrometer, a target property using said calibration, wherein said spectrometer comprises a base module, a communication bundle, and a sample module, wherein said communication bundle integrates said sampling module and said base module.
17. The method of claim 16 , wherein said step of transforming comprises at least one of:
independently preprocessing each of said N data blocks; and
independently decomposing and selecting scores for each of said N data blocks.
18. The method of claim 1 , wherein said target property comprises at least one of:
an analyte concentration;
an analyte composition; and
an analyte constituent identification.