# US 2024/0203569 A1 · Intelligent triage method and device, storage medium and electronic device

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

- Number: US 2024/0203569 A1
- Document ID: US20240203569A1
- Status: Published application
- Kind: A1
- Application number: 17/907,775
- Filed: 2021-06-30
- Published: 2024-06-20
- Earliest claimed: 2021-06-30
- Assignees: Boe Technology Group Co., Ltd.
- Applicants: Boe Technology Group Co., Ltd.
- Inventors: Zhenzhong ZHANG
- Web page: https://patentgrep.com/patent/US20240203569A1

## CPC

- G16H 40/20: for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
- G06F 40/30: Semantic analysis
- G06F 40/40: Processing or translation of natural language (natural language analysis G06F40/20; semantic analysis G06F40/30)
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G06F 16/35: Clustering; Classification
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

The present disclosure provides an intelligent triage method and device, a storage medium and an electronic device, and is related to artificial intelligence technology. The method includes: obtaining disease information of a target outpatient, and obtaining a vector of the target outpatient according to the disease information; vectorizing each doctor in a doctor knowledge graph to obtain a plurality of doctor vectors; and calculating matching degrees between the vector of the target outpatient and the plurality of doctor vectors using a matching degree calculation model, and recommending a doctor for the target outpatient according to magnitudes of the matching degrees.

## Claims

### Claim 1

An intelligent triage method, comprising:
obtaining disease information of a target outpatient, and obtaining a vector of the target outpatient according to the disease information;
vectorizing each doctor in a doctor knowledge graph to obtain a plurality of doctor vectors; and
calculating matching degrees between the vector of the target outpatient and the plurality of doctor vectors using a matching degree calculation model, and recommending a doctor for the target outpatient according to magnitudes of the matching degrees.

### Claim 2 (depends on 1)

The intelligent triage method according to claim 1, wherein the disease information comprises text information;
wherein obtaining the vector of the target outpatient according to the disease information comprises:
encoding each text in the text information to obtain a plurality of phrase vectors; and
sequentially inputting the plurality of phrase vectors into a pre-trained neural network to obtain the vector of the target outpatient.

### Claim 3 (depends on 1)

The intelligent triage method according to claim 1, further comprising:
obtaining corpus information of a plurality of doctors;
performing semantic analysis on the corpus information through natural language processing, and extracting relationships among the plurality of doctors; and
with the plurality of doctors as entities, constructing the doctor knowledge graph according to the relationships among the plurality of doctors.

### Claim 4 (depends on 1)

The intelligent triage method according to claim 1, wherein vectorizing each doctor in the doctor knowledge graph to obtain the plurality of doctor vectors comprises:
vectorizing each doctor in the doctor knowledge graph through a graph neural network to obtain the plurality of doctor vectors.

### Claim 5 (depends on 4)

The intelligent triage method according to claim 4, wherein vectorizing each doctor in the doctor knowledge graph through the graph neural network to obtain the plurality of doctor vectors comprises:
initializing node vectors corresponding to individual doctors in the doctor knowledge graph; and
iteratively updating each of the node vectors by using a pre-trained graph neural network to obtain the plurality of doctor vectors.

### Claim 6 (depends on 5)

The intelligent triage method according to claim 5, wherein iteratively updating each of the node vectors by using the pre-trained graph neural network to obtain the plurality of doctor vectors comprises:
updating each of the node vectors according to the following equation to obtain the plurality of doctor vectors:
h
t
+
1
(
e
i
)
=
∑
e
k
∈
Np
(
e
i
)
σ
⁡
(
W
p
×
e
i
+
W
ph
×
h
t
(
e
k
)
)
+
∑
e
j
∈
Nc
⁡
(
e
i
)
σ
⁡
(
W
c
×
e
i
+
W
ch
×
h
t
(
e
j
)
)
wherein Wp, Wph, Wc, and Wch are parameters of the graph neural network, σ is an activation function in the graph neural network, t is number of network iterations, ei is a node corresponding to an i-th doctor in the doctor knowledge graph, Np(ei) is a parent node set for the node ei, ek is a k-th parent node of the node et, Nc(ei) is a child node set for the node ei, ej is a j-th child node of the node ei, and ht(ei) represents a vector of the node ei after t network iterations.

### Claim 7 (depends on 1)

The intelligent triage method according to claim 1, wherein calculating the matching degrees between the vector of the target outpatient and the plurality of doctor vectors using the matching degree calculation model comprises:
calculating a matching degree score between the vector of the target outpatient and each of the plurality of doctors according to:
score (doci, patj)=νTσ(W[doci, patj]+b);
wherein [doci, patj] is a vector obtained by concatenating a i-th doctor vector doci and the vector patj of the target outpatient, W, ν and b are matching parameters, and σ is an activation function in the matching degree calculation model.

### Claim 8 (depends on 1)

The intelligent triage method according to claim 1, further comprising:
training the matching degree calculation model, the graph neural network and the neural network to obtain corresponding matching parameters and neural network parameters; and
based on the matching parameters and the neural network parameters, calculating the matching degree score between the vector of the target outpatient and each of the plurality of doctor vectors.

### Claim 9 (depends on 8)

The intelligent triage method according to claim 8, wherein training the matching degree calculation model to obtain the matching parameters of the matching degree calculation model comprises:
obtaining a training data set, wherein the training data set comprises a positive training data set and a negative training data set;
inputting the training data set into the matching degree calculation model, and constructing an objective function; and
determining the matching parameters of the matching degree calculation model according to the objective function.

### Claim 10 (depends on 9)

The intelligent triage method according to claim 9, wherein determining the matching parameters of the matching degree calculation model according to the objective function comprises:
updating the matching parameters of the matching degree calculation model using a stochastic gradient descent algorithm, and when the objective function converges, completing the training of the matching parameters.

### Claim 11 (depends on 10)

The intelligent triage method according to claim 10, wherein the objective function is:
L
=
∑
(
doc
i
,
pat
i
)
∑
(
doc
j
,
pat
i
)
[
score
(
doc
j
,
pat
i
)
+
γ
-
score
(
doc
i
,
pat
i
)
]
+
wherein (doci, pati) is positive training data and represents that an i-th doctor is suitable for dealing with a disease of an i-th patient, (docj, pati) is negative training data and represents that a j-th doctor is not suitable for dealing with a disease of the i-th patient, γ represents a preset difference threshold between a matching degree score score(doci, pati) of the positive training data and a matching degree score score(docj, pati) of the negative training data.

### Claim 12 (depends on 8)

The intelligent triage method according to claim 8, wherein training the graph neural network to obtain neural network parameters of the graph neural network comprises:
iteratively updating the neural network parameters of the graph neural network by using a backpropagation algorithm, and when an objective function converges, completing the training of the neural network parameters.

### Claim 13 (depends on 1)

The intelligent triage method according to claim 1, wherein recommending the doctor for the target outpatient according to magnitudes of the matching degrees comprises:
determining a set of doctors to be recommended according to the magnitudes of the matching degrees; and
recommending the doctor for the target outpatient based on the set of doctors to be recommended.

### Claim 14 (depends on 13)

The intelligent triage method according to claim 13, wherein recommending the doctor for the target outpatient based on the set of doctors to be recommended comprises:
recommending the set of doctors to be recommended to the target outpatient; and
according to a filtering condition input by the target outpatient, performing filtering on each doctor in the set of doctors to be recommended, so as to recommend a target doctor for the target outpatient.

### Claim 15

(canceled)

### Claim 16

(canceled)

### Claim 17

An electronic device, comprising:
a processor; and
a memory storing instructions executable by the processor;
wherein the processor is configured to:
obtain disease information of a target outpatient, and obtaining a vector of the target outpatient according to the disease information;
vectorize each doctor in a doctor knowledge graph to obtain a plurality of doctor vectors; and
calculate matching degrees between the vector of the target outpatient and the plurality of doctor vectors using a matching degree calculation model, and recommending a doctor for the target outpatient according to magnitudes of the matching degrees.

### Claim 18 (depends on 17)

The electronic device according to claim 17, wherein the disease information comprises text information;
wherein the processor is configured to:
encode each text in the text information to obtain a plurality of phrase vectors; and
sequentially input the plurality of phrase vectors into a pre-trained neural network to obtain the vector of the target outpatient.

### Claim 19 (depends on 17)

The electronic device according to claim 17, wherein the processor is further configured to:
obtain corpus information of a plurality of doctors;
perform semantic analysis on the corpus information through natural language processing, and extracting relationships among the plurality of doctors; and
with the plurality of doctors as entities, construct the doctor knowledge graph according to the relationships among the plurality of doctors.

### Claim 20 (depends on 17)

The electronic device according to claim 17, wherein the processor is configured to:
vectorize each doctor in the doctor knowledge graph through a graph neural network to obtain the plurality of doctor vectors.

### Claim 21 (depends on 20)

The electronic device according to claim 20, wherein the processor is configured to:
initialize node vectors corresponding to individual doctors in the doctor knowledge graph; and
iteratively update each of the node vectors by using a pre-trained graph neural network to obtain the plurality of doctor vectors.

### Claim 22 (depends on 21)

The electronic device according to claim 21, wherein the processor is configured to:
update each of the node vectors according to the following equation to obtain the plurality of doctor vectors:
h
t
+
1
(
e
i
)
=
∑
e
k
∈
Np
(
e
i
)
σ
⁡
(
W
p
×
e
i
+
W
ph
×
h
t
(
e
k
)
)
+
∑
e
j
∈
Nc
⁡
(
e
i
)
σ
⁡
(
W
c
×
e
i
+
W
ch
×
h
t
(
e
j
)
)
wherein Wp, Wph, Wc, and Wch are parameters of the graph neural network, σ is an activation function in the graph neural network, t is number of network iterations, ei is a node corresponding to an i-th doctor in the doctor knowledge graph, Np(ei) is a parent node set for the node ei, ek is a k-th parent node of the node ei, Nc(ei) is a child node set for the node ei, ej is a j-th child node of the node ei, and ht (ei) represents a vector of the node ei after t network iterations.
