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.
-
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