Patient diabetes monitoring system with clustering of unsupervised daily CGM profiles (or insulin profiles) and method thereof
First Claim
1. A patient diabetes monitoring system for a patient comprising:
- a physiological data input device which acquires a plurality of blood glucose concentration measurements of the patient within a time window to generate at least one time window dataset of collected unsupervised daily monitoring profiles;
a memory storing an unsupervised daily monitoring profile clustering algorithm; and
a processor in communication with said input device to receive said generated at least one time window dataset, and in communication with said memory in order to execute said unsupervised daily monitoring profile clustering algorithm,wherein said unsupervised daily monitoring profile clustering algorithm when executed by said processor causes said processor automatically to;
pre-process the dataset to generate a pre-processed dataset via a data transformation of the dataset that makes the pre-processed dataset symmetric for retrospective analysis,build a similarity matrix from the pre-processed dataset, andoutput an optimum number of similarity clusters found by the processor from the similarity matrix,wherein the data transformation for retrospective analysis results from processing the dataset with a hazard function defined by;
Gt=α
*ln(G−
β
)−
α
*ln(α
), where parameter α
=Tc−
β
, and parameter β
=Dr−
1, in which Tc is a center of a transformed space, Dr is a minimum defined glucose level, Gt is the transformed data of the blood glucose concentration measurements provided in the dataset, and G is original glucose level values of the blood glucose concentration measurements provided in the dataset and measured in millimoles per liter.
3 Assignments
0 Petitions
Accused Products
Abstract
A patient diabetes monitoring system with an efficient unsupervised daily monitoring profile clustering algorithm, a method, and a computer product thereof are disclosed. The system may include a physiological data input device or sensor which receives a plurality of physiological measurements to generate a dataset, a memory which stores a clustering algorithm, and a processor. The clustering algorithm when executed by the processor, causes the processor to automatically pre-process the dataset to control an amount of bias/aggressiveness from the collected unsupervised daily monitoring profiles, thereby generating a pre-processed dataset, build a similarity matrix from the pre-processed dataset, and output an optimum number of similarity clusters found by the processor from the similarity matrix.
45 Citations
18 Claims
-
1. A patient diabetes monitoring system for a patient comprising:
-
a physiological data input device which acquires a plurality of blood glucose concentration measurements of the patient within a time window to generate at least one time window dataset of collected unsupervised daily monitoring profiles; a memory storing an unsupervised daily monitoring profile clustering algorithm; and a processor in communication with said input device to receive said generated at least one time window dataset, and in communication with said memory in order to execute said unsupervised daily monitoring profile clustering algorithm, wherein said unsupervised daily monitoring profile clustering algorithm when executed by said processor causes said processor automatically to; pre-process the dataset to generate a pre-processed dataset via a data transformation of the dataset that makes the pre-processed dataset symmetric for retrospective analysis, build a similarity matrix from the pre-processed dataset, and output an optimum number of similarity clusters found by the processor from the similarity matrix, wherein the data transformation for retrospective analysis results from processing the dataset with a hazard function defined by;
Gt=α
*ln(G−
β
)−
α
*ln(α
), where parameter α
=Tc−
β
, and parameter β
=Dr−
1, in which Tc is a center of a transformed space, Dr is a minimum defined glucose level, Gt is the transformed data of the blood glucose concentration measurements provided in the dataset, and G is original glucose level values of the blood glucose concentration measurements provided in the dataset and measured in millimoles per liter. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. A non-transitory computer-readable medium that stores a program that, when executed by a processor, causes the processor to execute, via a patient diabetes monitoring system having a physiological data input device which acquires a plurality of blood glucose concentration measurements of the patient within a time window to generate at least one time window dataset of collected unsupervised daily monitoring profiles and which is in communication with said processor, such that said processor receives said generated at least one time window dataset, and in communication with said memory, an unsupervised daily monitoring profile clustering algorithm that causes said processor to automatically:
-
pre-process the dataset to generate a pre-processed dataset via a data transformation of the dataset that makes the pre-processed dataset symmetric for retrospective analysis, build a similarity matrix from the pre-processed dataset, and output an optimum number of similarity clusters, wherein the data transformation for retrospective analysis results from processing the dataset with a hazard function defined by;
Gt=α
*ln (G−
β
)−
α
*ln(α
), where parameter α
=Tc−
β
, and parameter β
=Dr−
1, in which Tc is a center of a transformed space, Dr is a minimum defined glucose level, Gt is the transformed data of the blood glucose concentration measurements provided in the dataset, and G is original glucose level values of the blood glucose concentration measurements provided in the dataset and measured in millimoles per liter. - View Dependent Claims (17)
-
-
18. A method for identifying day(s) where a diabetes control therapy was inadequate for a patient using a monitoring system comprising a display device, a physiological data input device and a processor, the method comprising:
-
receiving automatically from the physiological data input device a plurality of blood glucose concentration measurements of the patient within a time window to generate at least one time window dataset of collected unsupervised daily monitoring profiles; and executing from a memory a stored an unsupervised daily monitoring profile clustering algorithm and causing the processor automatically to; pre-process the dataset to generate a pre-processed dataset via a data transformation of the dataset that makes the pre-processed dataset symmetric for retrospective analysis, build a similarity matrix from the pre-processed dataset, and output on the display an optimum number of similarity clusters found by the processor from the similarity matrix, wherein the data transformation for retrospective analysis results from processing the dataset with a hazard function defined by;
Gt=α
*ln(G−
β
)−
α
*ln(α
), where parameter α
=Tc−
β
, and parameter β
=Dr−
1, in which Tc is a center of a transformed space, Dr is a minimum defined glucose level, Gt is the transformed data of the blood glucose concentration measurements provided in the dataset, and G is original glucose level values of the blood glucose concentration measurements provided in the dataset and measured in millimoles per liter.
-
Specification