Granted Patent B2

US 12,249,422 · App. 17/755,124 · Granted Mar 11, 2025

Inventors: Balasubramaniam Krishnan (Bangalore, IN); Ramesh Balaji (Chennai, IN); Srinivasa Raghavan Venkatachari (Chennai, IN); Arun Vijayakumar (Kochi, IN); Harish Kumar Dhanasekaran (Chennai, IN)

Assignee: Tata Consultancy Services Limited

G16H50/20A61B5/202G16H40/67

Abstract

Claims (40)

1. A processor implemented method for detection and validation of Nocturia in a person, the method comprising:

receiving readings from a plurality of sensors sensing location of the person over a predefined number of days, wherein the plurality of sensors are present in one or more rooms and one or more bathrooms;

extracting, via one or more hardware processors, a plurality of features from the plurality of sensors data;

creating, via one or more hardware processors, a featured engineered sensor dataset using the plurality of features;

applying, via one or more hardware processors, a set of statistical techniques on the featured engineered sensor dataset to detect anomalies in the bathroom visit patterns wherein the anomalies are used to detect the presence of Nocturia in the person based on a first predefined criterion;

applying, via one or more hardware processors, a set of rules on the featured engineered sensor dataset to identify bathroom visit patterns, wherein the bathroom visit patterns are used to determine the presence of Nocturia in the person based on a second predefined criterion;

applying, via one or more hardware processors, a one support vector machine (SVM) classifier on the plurality of features to detect anomalies in the bathroom visit patterns, wherein the anomalies are used to detect the presence of Nocturia in the person;

ensembling, via one or more hardware processors, outputs obtained from the set of statistical techniques, the set of rules and the SVM classifier to identify whether the person is showing a positive Nocturia or a negative Nocturia depending on a third predefined criterion;

identifying, via one or more hardware processors, anomalies in a sleeping pattern of the person using the featured engineered sensor dataset by applying a machine learning technique, wherein the sleeping pattern is either a positive or negative depending on a fourth predefined criterion, wherein the anomalies in the bathroom visit patterns are identified based on data points and data densities in the featured engineered sensor dataset; and

detecting and validating, via one or more hardware processors, the presence of Nocturia in the person if the person is showing positive Nocturia and the negative sleeping pattern.

2. The method of claim 1 further comprising detecting and validating the presence of Nocturia in the person using a Pearson correlation function.

3. The method of claim 1 , wherein each reading of the plurality of sensor readings comprises a sensor id from which reading is captured, location of the sensor and the date of the sensor reading.

4. The method of claim 1 , where in the anomalies in the number of visits in the one or more bathrooms is represented by coefficient of variance.

5. The method of claim 1 , wherein the second predefined criterion is the person visited bathroom more than 3 times in a night on last 2 weeks for more than 3 days.

6. The method of claim 1 , wherein the third predefined criterion comprising: identifying the Nocturia as positive if at least two of the set of statistical techniques, the set of rules and the one SVM classifier are showing the presence of Nocturia.

7. The method of claim 1 , wherein the fourth predefined criterion is a specific number of hours spent continuously in a bedroom during the night.

8. The method of claim 1 , wherein the plurality of features includes “SensorID”, date, hour, minute, second, day, week, month, year, weekday/weekend, start location, end location, time spent, room change indicator, daytime indicator.

9. A system for detection and validation of Nocturia in a person, the system comprises:

a plurality of sensors present in one or more rooms and one or more bathrooms, wherein the plurality of sensors are configured to sense location of the person over a predefined number of days;

one or more hardware processors; and

a memory in communication with the one or more hardware processors, the memory configured to perform:

receive a plurality of readings from the plurality of sensors;

extract a plurality of features from the plurality of readings;

create a featured engineered sensor dataset using the plurality of features;

apply a set of statistical techniques on the featured engineered sensor dataset to detect anomalies in the bathroom visit patterns wherein the anomalies are used to detect the presence of Nocturia in the person based on a first predefined criterion;

apply a set of rules on the featured engineered sensor dataset to identify bathroom visit patterns, wherein the bathroom visit patterns are used to determine the presence of Nocturia in the person based on a second predefined criterion;

apply a one support vector machine (SVM) classifier on the plurality of features to detect anomalies in the bathroom visit patterns, wherein the anomalies are used to detect the presence of Nocturia in the person;

ensemble outputs obtained from the set of statistical techniques, the set of rules and the SVM classifier to identify whether the person is showing a positive Nocturia or a negative Nocturia depending on a third predefined criterion;

identify anomalies in a sleeping pattern of the person using the featured engineered sensor dataset by applying a machine learning technique, wherein the sleeping pattern is either a positive or negative depending on a fourth predefined criterion, wherein the anomalies in the bathroom visit patterns are identified based on data points and data densities in the featured engineered sensor dataset; and

detect and validate the presence of Nocturia in the person if the person is showing positive Nocturia and the negative sleeping pattern.

10. A computer program product comprising a non-transitory computer readable medium having a computer readable program embodied therein, wherein the computer readable program, when executed on a computing device, causes the computing device to:

receiving readings from a plurality of sensors sensing location of the person over a predefined number of days, wherein the plurality of sensors are present in one or more rooms and one or more bathrooms;

extracting a plurality of features from the plurality of sensors data;

creating a featured engineered sensor dataset using the plurality of features;

applying a set of statistical techniques on the featured engineered sensor dataset to detect anomalies in the bathroom visit patterns wherein the anomalies are used to detect the presence of Nocturia in the person based on a first predefined criterion;

applying a set of rules on the featured engineered sensor dataset to identify bathroom visit patterns, wherein the bathroom visit patterns are used to determine the presence of Nocturia in the person based on a second predefined criterion;

applying a one support vector machine (SVM) classifier on the plurality of features to detect anomalies in the bathroom visit patterns, wherein the anomalies are used to detect the presence of Nocturia in the person;

ensembling outputs obtained from the set of statistical techniques, the set of rules and the SVM classifier to identify whether the person is showing a positive Nocturia or a negative Nocturia depending on a third predefined criterion;

identifying anomalies in a sleeping pattern of the person using the featured engineered sensor dataset by applying a machine learning technique, wherein the sleeping pattern is either a positive or negative depending on a fourth predefined criterion, wherein the anomalies in the bathroom visit patterns are identified based on data points and data densities in the featured engineered sensor dataset; and

detecting and validating, the presence of Nocturia in the person if the person is showing positive Nocturia and the negative sleeping pattern.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (13)