Calibration of near infrared quantitative measurement device using optical measurement cross-products
First Claim
1. A method for calibrating a near infrared (NIR) measurement device to a subject, said method comprising the steps of:
- forming a data set comprising a plurality of optical measurement data terms for said NIR measurement device;
augmenting said data set by forming cross-products terms using said data terms;
forming a plurality of subsets having a first specified number of members randomly selected from said data set;
evaluating each of said plurality of subsets against a set of reliable measurement results for said subject;
selecting one of said subsets based on a preselected set of criteria related to said reliable measurement results; and
using said selected set to form an optimal calibration for said device to said subject.
1 Assignment
0 Petitions
Accused Products
Abstract
In a method to improve the calibration of a non-invasive, near infrared (NIR) measurement device, a plurality of data terms is formed for the NIR measurement device. Then the codependence of the data terms is evaluated by forming cross-products terms using the data terms. Next, sets of prespecified sizes are randomly formed from the data terms and the cross-product terms. Each of these sets of terms is evaluated by testing the ability of the set to predict a set of accurate measurements using regression analysis. The method then selects one of the sets based on preselected criteria and uses the selected set to calibrate the NIR measurement device.
353 Citations
17 Claims
-
1. A method for calibrating a near infrared (NIR) measurement device to a subject, said method comprising the steps of:
-
forming a data set comprising a plurality of optical measurement data terms for said NIR measurement device;
augmenting said data set by forming cross-products terms using said data terms;
forming a plurality of subsets having a first specified number of members randomly selected from said data set;
evaluating each of said plurality of subsets against a set of reliable measurement results for said subject;
selecting one of said subsets based on a preselected set of criteria related to said reliable measurement results; and
using said selected set to form an optimal calibration for said device to said subject. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. A method for calibrating a non-invasive NIR blood glucose measurement device to a subject, said method comprising the steps of:
-
forming a data set comprising a plurality of optical measurement data terms for said NIR measurement device;
augmenting said data set by forming cross-products terms using said data terms;
forming a plurality of subsets having at least one specified number of members randomly selected from said data set;
evaluating each of said plurality of subsets against a set of reliable blood glucose measurements for said subject;
selecting one of said sets based on a preselected set of criteria, including the ability of the subset to predict said set of reliable blood glucose measurements for said subject; and
using said selected set to calibrate said NIR blood glucose measurement device to said subject.
-
-
17. A non-invasive device for measuring blood glucose levels in a subject, said device comprising:
-
means for gathering a plurality of data terms using NIR energy;
means for storing said plurality of data terms in a data set;
means for augmenting said data set with at least one cross-product term formed from said data terms and;
means for forming a plurality of subsets from said data set;
means for evaluating each of said plurality of subsets against reliable blood glucose level measurements for said subject; and
means for selecting at least one of said subsets to calibrate said device to said subject.
-
Specification