# US 12,008,458 B2 · Method and apparatus for visualizing health status information by using health space model

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,008,458 B2
- Document ID: US12008458B2
- Status: Granted
- Kind: B2
- Application number: 17/880,162
- Filed: 2022-08-03
- Published: 2023-02-02
- Granted: 2024-06-11
- Earliest claimed: 2021-08-02
- Assignees: Seoul National University R&Db Foundation
- Applicants: Seoul National University R&Db Foundation
- Inventors: Tae Sung Park; Cheol Gyun Park; Chan Hee Lee; O Ran Kwon; Yun Soo Kim; Eun Ok Lee
- Web page: https://patentgrep.com/patent/US12008458B2

## CPC

- G06N 3/045: Combinations of networks
- G06N 3/09: Supervised learning
- A61B 5/00: Measuring for diagnostic purposes (radiation diagnosis A61B6/00; diagnosis by ultrasonic, sonic or infrasonic waves A61B8/00); Identification of persons
- A61B 5/145: Measuring characteristics of blood in vivo, e.g. gas concentration or pH-value {; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid or cerebral tissue} (non-radiation detecting or locating of foreign bodies in blood A61B5/06)
- G06N 3/04: Architecture, e.g. interconnection topology
- G06N 3/047: Probabilistic or stochastic networks
- G06N 3/08: Learning methods
- G16H 10/60: for patient-specific data, e.g. for electronic patient records
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G06N 3/048: Activation functions
- G06N 5/02: Knowledge representation; Symbolic representation
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

An apparatus for visualizing a health status information of each individual by using a health space model includes a memory storing a health status information visualization program, and a processor configured to execute the visualization program. The visualization program inputs multidimensional data on the health status of each individual to the health space model to visually display a position of each individual in a two-dimensional health space, the health space model includes a first ordinal regression deep neural network model for outputting a first health status value based on multidimensional data of a first group and a second ordinal regression deep neural network model for outputting a second health status value based on multidimensional data of a second group, and the visualization program displays the health status information in a two-dimensional health space.

## Claims

### Claim 1

An apparatus for visualizing a health status information of each individual by using a health space model, the apparatus comprising:
a processor; and
a memory coupled to the processor and storing a health status information visualization program;
wherein, when executed by the processor, the health status information visualization program inputs multidimensional data on a health status of a target person to the health space model to visually display a position of each individual in a two-dimensional health space,
wherein the health space model includes a first ordinal regression deep neural network model for outputting a first health status value based on multidimensional data of a first group and a second ordinal regression deep neural network model for outputting a second health status value based on multidimensional data of a second group, and
the health status information visualization program displays the health status information of each individual in a two-dimensional health space by causing the first health status value to correspond to a first axis and the second health status value to correspond to a second axis,
wherein the first health status value indicates oxidative stress of the target person and the second health status value indicates metabolic stress of the target person.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein
the health space model further includes a third ordinal regression deep neural network model for outputting a third health status value based on multidimensional data of a third group, and
the health space model visually displays the position of each individual in a three-dimensional space based on the first to third health status values.

### Claim 3 (depends on 1)

The apparatus of claim 1, wherein
the multidimensional data of the first group includes age, gender smoking status, white blood cell count, and glutamic pyruvic transaminase (GPT) data of the target person, which are used to measure the oxidative stress of at least the target person,
the first ordinal regression deep neural network model outputs the first health status value indicating the oxidative stress of at least the target person,
the multidimensional data of the second group includes gender, body mass index (BMI), triglyceride level, high-density lipoprotein cholesterol index, and blood sugar level data of the person, which are used to measure the metabolic stress of at least the target person, and
the second ordinal regression deep neural network model outputs the second health status value indicating metabolic stress of at least the target person.

### Claim 4 (depends on 1)

The apparatus of claim 1, wherein
the first ordinal regression deep neural network model includes a deep neural network trained based on the multidimensional data of the first group for each individual and label values indicating the health status of each individual, and a classifier for dividing health status of each individual into k pieces (k is a plural natural number) based on the first health status value converted into a scalar value by multiplying an output of the deep neural network by a vector indicating a sharing coefficient,
the classifier classifies health statuses according to an ordinal regression analysis technique,
the second ordinal regression deep neural network model includes a deep neural network trained based on the multidimensional data of the second group for each individual and the label values indicating the health status of each individual, and a classifier for dividing the health status of each individual into k pieces (k is a plural natural number) based on the second health status value converted into a scalar value by multiplying an output of the deep neural network by a vector indicating a sharing coefficient, and
the classifier classifies health statuses according to the ordinal regression analysis technique.

### Claim 5 (depends on 4)

The apparatus of claim 4, wherein the classifier divides the health status into the k pieces based on k−1 values obtained by adding k−1 different intercept values to the values converted into the scalar value.

### Claim 6

A method of visualizing health status information by using an apparatus for visualizing health status information, the method comprising:
receiving multi-dimensional data on a health status of a target person;
inputting the received multidimensional data on the health status to a health space model;
outputting a first health status value as an output for multidimensional data of a first group by a first ordinal regression deep neural network model included in the health space model, and outputting a second health status value as an output for multidimensional data of a second group by a second ordinal regression deep neural network model; and
displaying the first health status value corresponding to a first axis and the second health status value corresponding to a second axis,
wherein the first health status value indicates oxidative stress of the target person and the second health status value indicates metabolic stress of the target person.

### Claim 7 (depends on 6)

The method of claim 6, wherein
the outputting further includes outputting a third health status value as an output for multidimensional data of a third group by a third ordinal regression deep neural network model included in the health space model, and
the displaying includes displaying the third health status value corresponding to a third axis.

### Claim 8 (depends on 6)

The method of claim 6, wherein
the multidimensional data of the first group includes age, gender, smoking status, white blood cell count, and glutamic pyruvictransaminase (GPT) data of the target person, which are used to measure the oxidative stress of at least the target person,
the first ordinal regression deep neural network model outputs the first health status value indicating the oxidative stress of at least the target person,
the multidimensional data of the second group includes gender, body mass index (BMI), triglyceride level, high-density lipoprotein cholesterol index, and blood sugar level data of the person, which are used to measure the metabolic stress of at least the target person, and
the second ordinal regression deep neural network model outputs the second health status value indicating the metabolic stress of at least the target person.

### Claim 9 (depends on 6)

The method of claim 6, wherein
the first ordinal regression deep neural network model includes a deep neural network trained based on the multidimensional data of the first group for each individual and label values indicating the health status of each individual, and a classifier for dividing the health status of each individual into k pieces (k is a plural natural number) based on the first health status value converted into a scalar value by multiplying an output of the deep neural network by a vector indicating a sharing coefficient,
the classifier classifies health statuses according to an ordinal regression analysis technique,
the second ordinal regression deep neural network model includes a deep neural network trained based on the multidimensional data of the second group for each individual and the label values indicating the health status of each individual, and a classifier for dividing the health status of each individual into k pieces (k is a plural natural number) based on the second health status value converted into a scalar value by multiplying an output of the deep neural network by a vector indicating a sharing coefficient, and
the classifier classifies health statuses according to the ordinal regression analysis technique.

### Claim 10 (depends on 9)

The method of claim 9, wherein the classifier divides the health status into the k pieces based on k−1 values obtained by adding k−1 different intercept values to the values converted into the scalar value.