WO2024104005A1 - 医疗信息处理方法以及装置、设备及存储介质 - Google Patents

医疗信息处理方法以及装置、设备及存储介质 Download PDF

Info

Publication number
WO2024104005A1
WO2024104005A1 PCT/CN2023/123804 CN2023123804W WO2024104005A1 WO 2024104005 A1 WO2024104005 A1 WO 2024104005A1 CN 2023123804 W CN2023123804 W CN 2023123804W WO 2024104005 A1 WO2024104005 A1 WO 2024104005A1
Authority
WO
WIPO (PCT)
Prior art keywords
candidate
subgraph
resources
resource
graph
Prior art date
Application number
PCT/CN2023/123804
Other languages
English (en)
French (fr)
Inventor
南静文
许利群
乔丰
王青松
卜昌郁
Original Assignee
中移(成都)信息通信科技有限公司
中国移动通信集团有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中移(成都)信息通信科技有限公司, 中国移动通信集团有限公司 filed Critical 中移(成都)信息通信科技有限公司
Publication of WO2024104005A1 publication Critical patent/WO2024104005A1/zh

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/901—Indexing; Data structures therefor; Storage structures
    • G06F16/9024—Graphs; Linked lists
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
    • G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records

Definitions

  • the present disclosure relates to, but is not limited to, the field of medical application technology or the field of big data technology, and in particular to a medical information processing method and apparatus, device and storage medium.
  • Medical records are records of medical activities such as examination, diagnosis, and/or treatment conducted by medical personnel on the occurrence, development and/or outcome of patients' diseases; such medical records play an important role in medical treatment, prevention, teaching, scientific research and/or hospital management.
  • the embodiments of the present disclosure provide a medical information processing method, apparatus, device and storage medium to at least solve some of the above-mentioned technical problems.
  • an embodiment of the present disclosure provides a medical information processing method, the method comprising:
  • a target structure graph is determined based on the second candidate structure subgraph.
  • the method comprises:
  • the selection matrix is determined; wherein the diagonal elements in the selection matrix are used to describe the first-order similarity of each of the resources; and the elements other than the diagonal elements in the selection matrix are used to describe the second-order similarity between the resources.
  • the resource in the first candidate structure subgraph is the first resource;
  • the resources in the selection matrix include the first resource and the second resource;
  • the determining of the second candidate structure subgraph based on the first candidate structure subgraph and the selection matrix includes:
  • the bias is the sum of the second-order similarities between the second resource and the first resource respectively;
  • the second candidate structure subgraph is determined.
  • the method comprises:
  • the acquiring, from the first original structure graph, a first candidate structure subgraph including at least two of the resources comprises:
  • the method of determining the selection matrix based on the first original structure graph and the first candidate structure subgraph Array including:
  • the selection matrix is determined based on the second original structure graph and the first candidate structure subgraph.
  • the resource related to the disease information and/or treatment information is selected from the resources of the first original structure diagram to form a second original structure diagram, including:
  • the second original structure graph is determined based on the connection relationship between each of the source nodes and each of the destination nodes, and the weight values between each of the source nodes and each of the destination nodes.
  • the method comprises:
  • determining a target structure graph based on the second candidate structure subgraph includes:
  • the target structure graph is determined based on the second candidate structure subgraph.
  • determining the support of the second candidate structure subgraph includes:
  • the first original structure graph contains all the resources contained in the second candidate structure subgraph, determine that the first original structure graph is a first similar structure graph; determine the support of the second candidate structure subgraph according to the ratio of the number of the first similar structure graphs to the number of the first original structure graphs;
  • the second original structure graph contains all the resources contained in the second candidate structure subgraph
  • the second original structure graph is determined to be a second similar structure graph
  • the supporting number of the second candidate structure subgraph is determined according to the ratio of the number of the second similar structure graphs to the number of the second original structure graphs.
  • an embodiment of the present disclosure provides a medical information processing device, the device comprising:
  • An acquisition module is configured to acquire at least one first original structure diagram, wherein the first original structure diagram is used to describe a topological relationship between resources; wherein one of the resources is used to describe a type of medical information;
  • a processing module configured to obtain a first candidate structure subgraph including at least two of the resources from the first original structure graph
  • the processing module is configured to determine a second candidate structure subgraph based on the first candidate structure subgraph and a selection matrix, wherein the selection matrix is used to determine the candidate resources added by the first candidate structure subgraph to obtain the second candidate structure subgraph;
  • the determination module is configured to determine a target structure graph based on the second candidate structure subgraph if the second candidate structure subgraph meets a predetermined condition.
  • an embodiment of the present disclosure provides a device, comprising a processor and a memory for storing a computer program that can be run on the processor; wherein, when the processor is used to run the computer program, the medical information processing method described in any embodiment of the present disclosure is implemented.
  • an embodiment of the present disclosure further provides a computer storage medium, wherein the computer storage medium contains computer executable instructions, and the computer executable instructions are executed by a processor to implement the medical information processing method described in any embodiment of the present disclosure.
  • the terminal obtains at least one first original structure graph, obtains a first candidate structure subgraph including at least two resources from the first original structure graph; and based on the first candidate structure subgraph and the selection matrix, determines the second candidate structure subgraph, and if the second candidate structure subgraph meets the predetermined conditions, determines the target structure graph based on the second candidate structure subgraph.
  • the selection matrix can be used to determine the candidate resources added to the second candidate structure subgraph by obtaining the first candidate structure subgraph, the node addition direction can be restricted by the selection matrix, so that a second candidate structure subgraph with strong directionality can be obtained; in this way, a target structure graph with a relatively tight structure of the second candidate structure subgraph can be obtained, which can provide a more directional data foundation in the field of medical applications and improve the focus of the mining results of medical information.
  • FIG1 is a flow chart of a medical information processing method provided in an embodiment of the present disclosure.
  • FIG. 2 is a schematic diagram of a first original structure diagram provided in an embodiment of the present disclosure.
  • FIG3 is a schematic diagram of a second first original structure diagram provided in an embodiment of the present disclosure.
  • FIG. 4 is a schematic diagram of a third first original structure diagram provided in an embodiment of the present disclosure.
  • FIG. 5 is a schematic diagram of a fourth first original structure diagram provided in an embodiment of the present disclosure.
  • FIG. 6 is a schematic diagram of a second original structure diagram provided in an embodiment of the present disclosure.
  • FIG. 7 is a flow chart of another medical information processing method provided in an embodiment of the present disclosure.
  • FIG8 is a schematic diagram of the structure of a medical information processing device provided by an embodiment of the present disclosure.
  • FIG. 9 is a schematic diagram of the hardware structure of a device provided in an embodiment of the present disclosure.
  • the present disclosure provides a method for processing medical information, including the following steps:
  • Step S11 Acquire at least one first original structure diagram, wherein the first original structure diagram is used to describe the topological relationship between resources; wherein one of the resources is used to describe a type of medical information;
  • Step S13 Acquire a first candidate including at least two of the resources from the first original structure diagram. Select the structural subgraph;
  • Step S15 determining a second candidate structure subgraph based on the first candidate structure subgraph and the selection matrix, wherein the selection matrix is used to determine the candidate resources added by the first candidate structure subgraph to obtain the second candidate structure subgraph;
  • Step S17 If the second candidate structure subgraph meets a predetermined condition, a target structure graph is determined based on the second candidate structure subgraph.
  • the medical information processing method provided in the embodiments of the present disclosure may be executed by a terminal; the terminal may be any mobile terminal or fixed terminal.
  • the terminal may be, but is not limited to, at least one of the following: a mobile phone, a computer, a server, a medical device, an industrial device, and/or a wearable device, etc.
  • the resource can be any kind of medical information or a data structure related to any kind of medical information; for example, the resource can be but is not limited to the following medical information or a data structure related to medical information: disease information, treatment information, management information, payment information, and workflow information.
  • the resource can be a minimum data unit.
  • the resources included in the first original structure graph, the first candidate structure subgraph, the second candidate structure subgraph, and the second original structure graph involved below may be two or more types.
  • At least some resources of any two first original structure diagrams are different, and/or the medical information described by at least some resources of any two first original structure diagrams is different.
  • FHIR Fast Healthcare Interoperability Resources
  • FHIR Release 4 has released a total of 145 resources, one resource corresponds to a data structure of medical information.
  • the Observation resource defines the data structure that describes symptom information
  • the CarePlan resource defines the data structure that describes the treatment plan, and so on.
  • each type of medical information may be associated with one or more resources; for example, as shown in Table 1 below, disease information may be associated with but not limited to at least one of the following resources: Allergy Intolerance resource, Condition resource, Observation resource, Diagnostic Report resource, Questionnaire Response resource, Family History resource, The resource includes FamilyMemberHistory, Immunization, and RiskAssessment.
  • the treatment information may be associated with, but not limited to, at least one of the following resources: MedicationRequest, ServiceRequest, NutritionOrder, Procedure, DeviceRequest, and CarePlan.
  • the "definition" in Table 1 is used to indicate the meaning of each resource; for example, the family medical history resource is used to describe the family medical history; and the service request resource is used to describe the test and/or examination prescription.
  • first original structure graph, the first candidate structure subgraph, the second candidate structure subgraph, the target structure graph and the second original structure graph described below can all be any structure graphs that describe the connection relationship or topological relationship between resources.
  • the connection relationship or topological relationship can be used to indicate that resources are associated with each other.
  • the first original structure diagram may be a structure diagram with connection relationships between resources.
  • the first original structure diagram may be as shown in FIG2, in which the resources include a physical sign record 1 resource, a patient resource, a service request resource, a medical event resource, a diagnosis record resource, and a physical sign record 2 resource; the first original structure diagram also includes a connection relationship between an observation resource and a patient resource, and a connection relationship between a service application resource and a patient resource, etc.
  • the second original structure diagram may be a structure diagram with topological relationships between resources.
  • the first original structure diagram may be as shown in FIG3, and the first original structure diagram may be a topological structure diagram; in the topological structure diagram, "1", “2", “3", “4", "5" and “6” may respectively indicate a resource, for example, they may be used to indicate the vital sign record 1 resource, patient resource, diagnosis record resource, medical event resource, service application resource and vital sign record 2 resource as shown in FIG2.
  • the arrows in the first candidate structure sub-graph may be used to describe the connection relationship between resources; for example, the arrow between "1" and "2” is used to describe that the resources indicated by “1" and "2" are associated.
  • the first original structure graph, the first candidate structure subgraph, the second candidate structure subgraph, the target structure graph, and the second original structure graph can all be directed structure graphs.
  • the arrows connecting resources are directional; the arrow between "1" and “2" in FIG3 is used to describe that the resource indicated by “1" is reachable to the resource indicated by "2".
  • step S13 includes: acquiring any at least two resources from the first original structure graph to form a first candidate structure subgraph.
  • the first original structure diagram can be selected from resource “1” and resource “2” constitutes the first candidate structural subgraph.
  • the method includes: determining a selection matrix.
  • the terminal may determine the selection matrix based on the correlation between each resource, or the terminal may determine the selection matrix based on the historical structure diagram; or the terminal may determine the selection matrix based on the first original structure diagram.
  • the diagonal elements in the selection matrix are used to describe the correlation between each resource itself, and the elements other than the diagonal elements in the selection matrix are used to describe the correlation between resources.
  • the correlation can be represented by a weight value.
  • the first original structure diagram includes resource "1", resource “2”, resource “3”, resource “4", resource “5" and resource “6”;
  • the selection matrix can be as shown in the following formula:
  • the diagonal elements are all “ ⁇ ”, which are used to indicate the association between resource “1”, resource “2”, resource “3”, resource “4", resource “5" and resource “6” respectively.
  • Elements other than the diagonal elements are used to refer to the association between resources; for example, the elements in the 1st row and the 2nd column are used to indicate the association between resource "1" and resource "2", and resource "1" is directly connected to resource "2"; then the weight value of the association between resource "1” and resource “2” can be determined to be 1.
  • the elements in the 3rd row and the 5th column are used to indicate the association between resource “3” and resource "5", and resource "3" to resource "5" needs to pass through one resource; then the weight value of the association between resource “1” and resource "5" can be determined to be
  • the element in the 5th row and the 1st column is used to indicate the association between resource “6" and resource "1". Resource “6" cannot be connected to resource "1", so the weight value of the association between resource “6” and resource “1” can be determined to be 0.
  • the resource included in the first candidate structure subgraph is the first resource; the candidate matrix The resources include a first resource and a second resource; the candidate matrix is used to select a second resource with the strongest correlation with the first candidate structure subgraph from the second resources as a candidate resource.
  • the selection matrix is used to determine candidate resources among resource “3", resource "4", resource "5" and resource "6".
  • step S15 may be repeatedly executed to continuously obtain new first candidate structure subgraphs and second candidate structure subgraphs until the obtained second candidate structure subgraph meets a predetermined condition, and then the target structure graph is determined based on the second candidate structure subgraph.
  • step S15 when step S15 is repeatedly executed, it may be that: after the first execution of step S15, the first second candidate structure subgraph is obtained; the first second candidate structure subgraph is used as the first candidate structure subgraph for the second execution of step S15; and so on, the i-th second candidate structure subgraph obtained after the i-1th execution of step S15 is used as the first candidate structure subgraph for the i-th execution of step S15; wherein i is an integer greater than 1.
  • the second candidate structure subgraph in step S17 satisfies the predetermined condition, which may be but is not limited to at least one of:
  • the support degree of the second candidate structural subgraph is less than or equal to the first predetermined value
  • the number of resources included in the second candidate structure subgraph is greater than or equal to a second predetermined value.
  • the method includes: determining the support of the second candidate structural subgraph
  • the step S17 includes: if the support degree of the second candidate structure subgraph is less than or equal to a first predetermined value, determining the target structure graph based on the second candidate structure subgraph.
  • determining the support of the second candidate structure subgraph includes:
  • the first original structure graph contains all the resources contained in the second candidate structure subgraph, determining that the first original structure graph is a first similar structure graph
  • the support degree of the second candidate structure subgraph is determined according to the ratio of the number of the first similar structure graphs to the number of the first original structure graphs.
  • the first original structure graph includes resources “1”, “2”, “3” and “4"
  • the second candidate structure subgraph includes resources “1” and “2”
  • the first original structure graph includes all resources included in the second candidate structure subgraph
  • the first original structure graph is the first similar structure graph.
  • the first original structure graph includes resources “1”, “2”, “3” and "4"
  • the second candidate structure subgraph includes resources “1", "4" and "5"
  • the first original structure graph does not include all resources included in the second candidate structure subgraph, and the first original structure graph is not the first similar structure graph.
  • the support degree may be the ratio of the number of first similar structure graphs to the number of first original structure graphs; or the support degree may be a value that differs from the ratio of the number of first similar structure graphs to the number of first original structure graphs by a predetermined value.
  • the first predetermined value and the second predetermined value can be determined by the terminal based on the user input operation, or determined based on historical experience information or preset.
  • the first predetermined value can be 90% or 80% or 70%, etc.
  • the second predetermined value can be 100, 200, 1000 or 10000, etc.
  • determining a target structure graph based on the second candidate structure subgraph in step S17 includes:
  • the first candidate structure subgraph corresponding to the second candidate structure subgraph that meets the predetermined condition is determined as the target structure graph.
  • G-SUP( gi ) is the support of the second candidate structure subgraph
  • is the number of the first similar structure graphs
  • is the number of the first original structure graphs.
  • the terminal obtains at least one first original structure diagram, and
  • the first candidate structural subgraph including at least two resources is obtained in the graph; and based on the first candidate structural subgraph and the selection matrix, the second candidate structural subgraph is determined. If the second candidate structural subgraph meets the predetermined conditions, the target structural graph is determined based on the second candidate structural subgraph.
  • the selection matrix can be used to determine the candidate resources added to the second candidate structural subgraph obtained by the first candidate structural subgraph, the node addition direction can be restricted by the selection matrix, so that a second candidate structural subgraph with strong directionality can be obtained; in this way, the target structural graph with a relatively tight structure of the second candidate structural subgraph can provide a more directional data foundation in the field of medical applications, and improve the focus of the mining results of medical information.
  • the target structure graph is determined based on the second candidate structure subgraph; in this way, a relatively large or largest frequent subgraph can be mined as the target structure graph, and the relatively large or largest frequent subgraph can fit the disease course mining target in terms of content, and can further ensure the tightness and focus of the structure; the target structure graph obtained in this way can provide important data support and thinking guidance for the medical application field.
  • the method includes: determining the selection matrix based on the first original structure graph and the first candidate structure subgraph; wherein the diagonal elements in the selection matrix are used to describe the first-order similarity of each of the resources; and the elements other than the diagonal elements in the selection matrix are used to describe the second-order similarity between the resources.
  • the first-order similarity can be determined based on the sum of the in-degree and out-degree of the resource; the second-order similarity refers to the first-order similarity between the resource and the adjacent resource.
  • the adjacent resource of a resource refers to the resource directly connected to the resource in the structure diagram; for example, as shown in Figure 3, the resource indicated by "1" can be the adjacent resource of the resource indicated by "2".
  • the first-order similarity can be determined based on the sum of the weight values of the resource and the first adjacent resource and the weight values of the resource and the second adjacent resource; the second-order similarity can be determined based on the weight value of the resource and the first adjacent resource or based on the weight value of the resource and the second adjacent resource; wherein the first adjacent resource is an adjacent resource entering the resource, and the second adjacent resource is an adjacent resource leaving the resource.
  • the diagonal elements in the selection matrix as the first-order similarity of each resource
  • the elements outside the diagonal elements are the second-order similarities between resources, and a tightly structured selection matrix can be determined, so that the candidate resources that can be grown are the resources that make the structure diagram more tightly structured.
  • the resource in the first candidate structure subgraph is a first resource; the resources in the selection matrix include the first resource and the second resource;
  • the step S15 determines the second candidate structure subgraph based on the first candidate structure subgraph and the selection matrix, including:
  • the bias is the sum of the second-order similarities between the second resource and the first resource respectively;
  • the second candidate structure subgraph is determined.
  • the second-order similarity may be the sum of weight values of the associations between resources.
  • the second resource with the largest deviation can be selected as the candidate resource, so that the increased candidate resources can be the resources that make the structure graph relatively the most compact, which is conducive to determining the second candidate structure subgraph with a more compact structure.
  • the second resource with the second largest bias may be determined as a candidate resource, or a second resource with a bias ranked in the top few may be determined as a candidate resource, etc.; and resources that make the structure diagram relatively compact may also be determined to a certain extent.
  • the method comprises:
  • the step S13 comprises:
  • the first candidate structure subgraph including at least two of the resources is obtained from the second original structure graph.
  • the first original structure diagram includes 10 resources; the 10 resources are resources related to disease information, treatment information or payment information, among which 5 resources are resources related to disease information or treatment information; the terminal can determine that the structure diagram composed of the 5 resources is the second original structure diagram.
  • the selecting of the disease information from the resource of the first original structure map The resources related to the information and/or treatment information constitute a second original structure diagram, including:
  • the second original structure graph is determined based on the connection relationship between each of the source nodes and each of the destination nodes, and the weight values between each of the source nodes and each of the destination nodes.
  • the resources related to disease information and/or treatment information are third resources; the third resources can be divided into source nodes and destination nodes.
  • the source node and the destination node can be any one of the third resources; the same third resource can be a source node or a destination node at the same time.
  • the first original structure diagram may include resources corresponding to “1", “2", “3”, “4", “5" and “6", respectively, and "1", “2", “3”, “4", “5" and “6” correspond to physical sign record resources, patient resources, diagnosis record resources, medical event resources, service application resources and physical sign record resources, respectively.
  • “1", “3”, “5" and “6” may be resources related to disease information or treatment information
  • “2" and “4" are resources not related to disease information or treatment information
  • resources “1", “3", “5" and “6” related to disease information or treatment information are marked in the first original structure diagram.
  • the terminal can reconnect to the reserved resources according to the accessibility and shortest path principle of "1", "3", "5" and “6” in the first original structure diagram, and the connection method may be shown in FIG. 6.
  • connection method based on Figure 6 can be: if the source node can reach the destination node, then add a directed edge from the source node to the destination node; assign a weight value to the directed edge; the weight value can be determined based on the resources that the source node passes through to reach the destination node.
  • the weight value of the source node is determined to be ⁇ ; for another example, if the source node "1" reaches the destination node "3" in the first original structure diagram without passing through other resources, that is, passing through 0 resources, then the weight value from the source node "1" to the destination node “3” is determined to be 1; for another example, if the source node "3" reaches the destination node "1" in the first original structure diagram and needs to pass through 2 resources, then the weight value from the source node "3" to the destination node "1” is determined.
  • each diagonal element is the weight value of each resource being both a source node and a destination node; the elements outside the diagonal elements are the weight values of each resource (ie, source node) to the adjacent resource (ie, destination node).
  • some resources irrelevant to disease information and/or treatment information can be removed from the first original structure diagram, while resource information related to disease information and/or treatment information is retained; and the connection relationship of these resources in the second original structure diagram is determined by the connection relationship of the retained resources in the first original structure diagram, so that the first original structure diagram can be simplified to obtain an equivalent second original structure diagram, which is conducive to the subsequent simplification of the determination of the selection matrix and the calculation of selecting the target structure diagram.
  • the equivalent adjacency matrix of the obtained second original structure graph is determined based on the weight values determined by the resources through which the resources of the retained disease information and/or treatment information are connected in the first original structure graph, the topological information of the first original structure graph can be retained.
  • determining the selection matrix based on the first original structure graph and the first candidate structure subgraph in step S15 includes:
  • the selection matrix is determined based on the second original structure graph and the first candidate structure subgraph.
  • the equivalent adjacency matrix of the second original structure graph is From the second primitive structure
  • the first candidate structure subgraph with resources “1" and "3” is selected;
  • the equivalent adjacency matrix of the second candidate structure subgraph pair is
  • the selection matrix can be determined from the equivalent adjacency matrix Madj2 as follows: Among them, the diagonal elements in the selection matrix are used to describe the first-order similarity of each resource; the elements outside the diagonal elements in the selection matrix are used to describe the second-order similarity between the resources.
  • the diagonal elements in the 1st row and 1st column of the selection matrix are used to describe the first-order similarity of resource "1".
  • the corresponding selection matrix part of the first candidate structure subgraph is The second resources included in the selection matrix are resources "5" and "6".
  • the deflection degree of each second resource can be calculated as Among them, m is the number of resources in the first candidate structure subgraph, Di is the i-th resource
  • the terminal can determine the second candidate structure subgraph based on the first candidate structure subgraph and the candidate resource "5".
  • the embodiment of the present disclosure is based on a selection matrix determined based on a second original structure graph and a first candidate structure subgraph.
  • the second candidate structure subgraph determined based on the candidate matrix can reduce resources with weak correlation while retaining the topological information of the original first original structure graph to a certain extent, thereby greatly simplifying calculations on the basis of a relatively complete data model.
  • determining the support of the second candidate structure subgraph includes:
  • the second original structure graph contains all the resources contained in the second candidate structure subgraph, determining that the second original structure graph is a second similar structure graph
  • the support degree of the second candidate structure subgraph is determined according to the ratio of the number of the second similar structure graphs to the number of the second original structure graphs.
  • the second original structure graph includes resources “1”, “2”, “5" and “6”
  • the second candidate structure subgraph includes resources “1” and “3”
  • the second original structure graph includes all resources included in the second candidate structure subgraph
  • the second original structure graph is the second similar structure graph.
  • the second original structure graph includes resources “1”, “3”, “5" and “6”
  • the second candidate structure subgraph includes resources “1", "3” and “4"
  • the second original structure graph does not include all resources included in the second candidate structure subgraph, and the second original structure graph is not the second similar structure graph.
  • the support degree may be the ratio of the number of second similar structure graphs to the number of second original structure graphs; or the support degree may be a value different from the ratio of the number of second similar structure graphs to the number of second original structure graphs by a predetermined value.
  • the selection matrix and the first candidate structure subgraph are determined based on the second original structure graph, so that the second original structure graph can be determined as a comparison sample; based on the support of the candidate structure graph subgraph determined based on the second original structure graph, a relatively accurate support can also be determined.
  • the present disclosure provides a medical information processing method, which is executed by a terminal and includes the following steps:
  • Step S21 the terminal obtains at least one first original structure diagram
  • the terminal can use FHIR to model the medical information and establish a first original structure diagram for each patient; the first original structure diagram can be shown in Figure 2 as the vital sign record 1 resource, patient resource, diagnosis record resource, medical event resource, service application resource and vital sign record 2 resource.
  • the terminal converts the first original structure diagram into a topological model as shown in Figure 3, and the vital sign record 1 resource, patient resource, diagnosis record resource, medical event resource, service application resource and vital sign record 2 resource are indicated by "1", "2", “3", "4", "5" and "6” in Figure 3 respectively.
  • the connection relationship between resources in Figure 2 is the association between the references of resources.
  • Step S22 the terminal selects the resources related to disease information and/or treatment information from the resources of the first original structure diagram to form a second original structure diagram;
  • the terminal selects resources related to disease information and/or treatment information from the physical sign record 1 resource, patient resource, diagnosis record resource, medical event resource, service application resource and physical sign record 2 resource in the first original structure diagram: physical sign record 1 resource, diagnosis record resource, service application resource and physical sign record 2 resource; the second original structure diagram of the physical sign record 1 resource, diagnosis record resource, service application resource and physical sign record 2 resource is obtained as shown in Figure 6.
  • the second original structure diagram if the source node can reach the destination node, a directed edge is added from the source node to the destination node; and a weight value is assigned to the directed edge.
  • the terminal may determine that the equivalent adjacency matrix of the second original structure graph is:
  • Step S23 The terminal determines the selection matrix and determines the added candidate resources
  • the method of determining the candidate resources to be added can be considered as a biased node growth method.
  • the terminal determines the first candidate structure subgraph as a structure graph including resources "1" and "3" based on the second original structure graph; the terminal determines the selection matrix as follows based on the equivalent adjacency matrix Madj2 of the second original structure graph: Among them, the diagonal elements in the selection matrix are used to describe the first-order similarity of each resource; the elements outside the diagonal elements in the selection matrix are used to describe the second-order similarity between the resources. It can be seen from the selection matrix M select that resources "5" and "6" are two candidate directions for node growth, that is, the resources "5" and "6" can be selected as candidate resources in the second candidate structural subgraph.
  • Step S24 the terminal determines a second candidate structure subgraph according to the first candidate structure subgraph and the candidate resources.
  • Step S25 the terminal determines the support degree of the second candidate structure subgraph
  • the terminal determines the support of the second candidate structure subgraph
  • G'-SUP(g i ) is the support of the second candidate structure subgraph
  • is the number of the second similar structure graphs
  • is the number of the second original structure graphs.
  • Step S26 The terminal determines whether the support of the second candidate structural subgraph is less than or equal to the first predetermined value, if not, executing step S27; if yes, executing step S28;
  • the terminal determines the first predetermined value to be G-SUP min ; if G′-SUP( gi ) ⁇ G-SUP min , execute step S27 ; otherwise, execute step S28 .
  • Step S27 the terminal determines the second candidate structure subgraph as the first candidate structure subgraph
  • the terminal uses the second candidate structure subgraph as the first candidate structure subgraph that needs to add candidate resources next time.
  • Step S28 The terminal determines the target structure graph based on the second candidate structure sub-graph.
  • the terminal can use the second candidate structural subgraph when the support is less than or equal to the predetermined value as the target structural graph.
  • the second candidate structural subgraph determines the structural graph of the last added candidate resource (i.e., the first candidate structural subgraph corresponding to the second candidate structural subgraph) as the target structural graph; the target structural graph at this time is the maximum frequent subgraph including the largest number of resources.
  • the maximum frequent subgraph can not only fit the course mining target in terms of content, but also ensure the tightness and focus of the structure.
  • resources irrelevant to disease information and/or treatment information can be removed, and the topological relationship of the original treatment information can be retained to a certain extent through the equivalent adjacency matrix, which can simplify the calculation while providing a relatively more complete data model for the subsequent mining of the target structure graph.
  • a selection matrix can be defined to restrict the direction of resource growth, and the biased ability of the second candidate structure subgraph can be controlled; by selecting the resource with the largest bias as the candidate node, the resource that is more closely coupled with the first candidate structure subgraph can be selected and added to the second candidate structure subgraph.
  • the target structure graph to be mined can be biased towards a tightly structured topology, and the focus of the mined structure can be improved.
  • the present invention can also realize the conversion of equivalent topological structures (i.e., conversion from a first original structure graph to an equivalent second original structure graph) and a special node growth method (i.e., controlling the candidate resources with the largest or relatively large bias to be the additional resources required to transform the first candidate structure subgraph into the second candidate structure subgraph), thereby realizing the effective integration of the two.
  • equivalent topological structures i.e., conversion from a first original structure graph to an equivalent second original structure graph
  • a special node growth method i.e., controlling the candidate resources with the largest or relatively large bias to be the additional resources required to transform the first candidate structure subgraph into the second candidate structure subgraph
  • an embodiment of the present disclosure provides a medical information processing device, the device comprising:
  • the acquisition module 41 is configured to acquire at least one first original structure diagram, wherein the first original structure diagram is used to describe the topological relationship between resources; wherein one of the resources is used to describe a type of medical information;
  • a processing module 42 is configured to obtain a first candidate structure subgraph including at least two of the resources from the first original structure graph;
  • the processing module 42 is configured to determine a second candidate structure subgraph based on the first candidate structure subgraph and a selection matrix, wherein the selection matrix is used to determine the candidate resources added by the first candidate structure subgraph to obtain the second candidate structure subgraph;
  • the determination module 43 is configured to determine a target structure graph based on the second candidate structure subgraph if the second candidate structure subgraph meets a predetermined condition.
  • the apparatus comprises:
  • the processing module 42 is configured to determine the selection matrix based on the first original structure graph and the first candidate structure subgraph; wherein the diagonal elements in the selection matrix are used to describe the first-order similarity of each of the resources; and the elements other than the diagonal elements in the selection matrix are used to describe the second-order similarity between the resources.
  • the resource in the first candidate structure subgraph is a first resource; the resources in the selection matrix include the first resource and the second resource;
  • the processing module 42 is configured to determine, based on the bias of each second resource in the selection matrix, the second resource with the largest bias as the candidate resource; the bias is the sum of the second-order similarities between the second resources and the first resources respectively;
  • the second candidate structure subgraph is determined.
  • the acquisition module 41 is configured to select the resources related to disease information and/or treatment information from the resources of the first original structure diagram to form a second original structure diagram;
  • the processing module 42 is configured to obtain the first candidate structure subgraph including at least two of the resources from the second original structure graph;
  • the processing module 42 is further configured to determine the selection matrix based on the second original structure graph and the first candidate structure subgraph.
  • the processing module 42 is configured to perform the following steps:
  • the second original structure graph is determined based on the connection relationship between each of the source nodes and each of the destination nodes, and the weight values between each of the source nodes and each of the destination nodes.
  • the processing module 42 is configured to determine the support of the second candidate structure subgraph
  • the determination module 43 is configured to determine the target structure graph based on the second candidate structure subgraph if the support of the second candidate structure subgraph is less than or equal to a first predetermined value.
  • the processing module 42 is configured to determine that the first original structure graph is a first similar structure graph if the first original structure graph contains all the resources contained in the second candidate structure subgraph;
  • the processing module 42 is further configured to determine the support degree of the second candidate structure subgraph according to the ratio of the number of the first similar structure graphs to the number of the first original structure graphs.
  • the processing module 42 is configured to determine that the second original structure graph is a second similar structure graph if the second original structure graph contains all the resources contained in the second candidate structure subgraph;
  • the processing module 42 is further configured to determine the support degree of the second candidate structure subgraph according to the ratio of the number of the second similar structure graphs to the number of the second original structure graphs.
  • the present disclosure also provides a device, wherein the device includes a processor 61 and a memory 62 for storing a computer program that can be run on the processor 61; wherein the processor 61 is used to implement the medical information processing method of any embodiment of the present disclosure when running the computer program.
  • the memory in the disclosed embodiments may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories.
  • the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory.
  • the volatile memory may be a random access memory (RAM), which is used as an external cache.
  • RAM Static RAM
  • DRAM Dynamic RAM
  • SDRAM Synchronous DRAM
  • DDRSDRAM Double Data Rate SDRAM
  • ESDRAM Enhanced SDRAM
  • SLDRAM Synchlink DRAM
  • DRRAM Direct Rambus RAM
  • the processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit in the processor or the instructions in the form of software.
  • the above processor may be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.
  • DSP Digital Signal Processor
  • ASIC Application Specific Integrated Circuit
  • FPGA Field Programmable Gate Array
  • the various methods, steps and logic block diagrams disclosed in the embodiments of the present disclosure can be implemented or executed.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
  • the steps of the method disclosed in conjunction with the embodiments of the present disclosure can be directly embodied as being executed by a hardware decoding processor, or can be executed by a combination of hardware and software modules in the decoding processor.
  • the software module can be located in a random access memory, a flash memory, a read-only memory, Programmable read-only memory or electrically erasable programmable memory, registers and other mature storage media in the art.
  • the storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.
  • the embodiments described herein may be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof.
  • the processing unit may be implemented in one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), general purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions of the invention, or a combination thereof.
  • ASICs application specific integrated circuits
  • DSPs digital signal processors
  • DSPs digital signal processing devices
  • PLDs programmable logic devices
  • FPGAs field programmable gate arrays
  • general purpose processors controllers, microcontrollers, microprocessors, other electronic units for performing the functions of the invention, or a combination thereof.
  • the techniques described herein can be implemented by modules (e.g., procedures, functions, etc.) that perform the functions described herein.
  • the software code can be stored in a memory and executed by a processor.
  • the memory can be implemented in the processor or outside the processor.
  • Yet another embodiment of the present disclosure provides a computer storage medium, wherein the computer-readable storage medium stores an executable program, and when the executable program is executed by a processor, the steps of the medical information processing method of any embodiment of the present disclosure can be implemented.
  • the computer storage medium may include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk and other media that can store program code.
  • the medical information processing method, apparatus, device and storage medium can obtain at least one first original structure graph, and obtain a first candidate structure subgraph including at least two resources from the first original structure graph; and determine the second candidate structure subgraph based on the first candidate structure subgraph and the selection matrix, and if the second candidate structure subgraph meets the predetermined conditions, determine the target structure graph based on the second candidate structure subgraph; because the selection matrix can be used to determine the candidate resources added to the second candidate structure subgraph by obtaining the first candidate structure subgraph, the node addition direction can be restricted by the selection matrix, so that a second candidate structure subgraph with strong directionality can be obtained; in this way, a target structure graph with a relatively tight structure of the second candidate structure subgraph can be obtained, which can provide a more directional data basis in the field of medical applications and improve the focus of the mining results of medical information.

Landscapes

  • Engineering & Computer Science (AREA)
  • Public Health (AREA)
  • Data Mining & Analysis (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • Epidemiology (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Theoretical Computer Science (AREA)
  • Pathology (AREA)
  • Biomedical Technology (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本公开实施例公开了一种医疗信息处理方法以及装置、设备及存储介质；方法包括：获取至少一个第一原始结构图，所述第一原始结构图用于描述资源与资源之间的拓扑关系；一种所述资源用于描述一种医疗信息；从所述第一原始结构图中获取包括至少两种所述资源的第一候选结构子图；基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，其中，所述选型矩阵用于确定第一候选结构子图得到第二候选结构子图所增加的候选资源；若所述第二候选结构子图满足预定条件，基于所述第二候选结构子图，确定目标结构图。本公开实施例可以得到由指向性较强的第二候选结构子图，得到结构相对紧密的目标结构图，提高了医疗信息的挖掘结果的聚焦性。

Description

医疗信息处理方法以及装置、设备及存储介质
相关申请的交叉引用
本公开基于申请号为202211429239.1、申请日为2022年11月15日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本公开作为参考。
技术领域
本公开涉及但不限于医疗应用技术领域或者大数据技术领域，尤其涉及一种医疗信息处理方法以及装置、设备及存储介质。
背景技术
病历是医务人员对患者疾病的发生、发展和/或转归等进行的检查、诊断、和/或治疗等医疗活动过程的记录；该病历对于医疗、预防、教学、科研和/或医院管理等都有重要的作用。
随着社会信息化的发展，电子病历应运而生。电子病历的出现使得可整合和/或可利用的医疗记录数据呈爆炸式增长；而如何采用科学有效的方法，挖掘出病历数据中隐藏的有价值的信息，对医学研究具有重要意义。
发明内容
有鉴于此，本公开实施例提供一种医疗信息处理方法以及装置、设备及存储介质，以至少解决上述部分技术问题。
本公开的技术方案是这样实现的：
第一方面，本公开实施例提供一种医疗信息处理方法，所述方法包括：
获取至少一个第一原始结构图，其中，所述第一原始结构图用于描述资源与资源之间的拓扑关系；其中，一种所述资源用于描述一种医疗信息；
从所述第一原始结构图中获取包括至少两种所述资源的第一候选结构子图；
基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，其中，所述选型矩阵用于确定第一候选结构子图得到第二候选结构子图所增加的候选资源；
若所述第二候选结构子图满足预定条件，基于所述第二候选结构子图，确定目标结构图。
上述方案中，所述方法包括：
基于所述第一原始结构图及所述第一候选结构子图，确定所述选型矩阵；其中，所述选型矩阵中对角线元素用于描述每个所述资源的一阶相似度；所述选型矩阵中对角线元素之外的元素用于描述所述资源与所述资源之间的二阶相似度。
上述方案中，所述第一候选结构子图中所述资源为第一资源；所述选型矩阵中所述资源包括所述第一资源和第二资源；
所述基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，包括：
基于所述选型矩阵中各所述第二资源的偏向度，确定所述偏向度最大的所述第二资源为所述候选资源；所述偏向度为所述第二资源分别与所述第一资源之间的所述二阶相似度之和；
基于所述第一候选结构子图及所述候选资源，确定所述第二候选结构子图。
上述方案中，所述方法包括：
从所述第一原始结构图的所述资源中选取与疾病信息和/或治疗信息相关的所述资源，构成第二原始结构图；
所述从所述第一原始结构图中获取包括至少两种所述资源的第一候选结构子图，包括：
从所述第二原始结构图中获取包括至少两种所述资源的所述第一候选结构子图；
所述基于所述第一原始结构图及所述第一候选结构子图，确定所述选型矩 阵，包括：
基于所述第二原始结构图及所述第一候选结构子图，确定所述选型矩阵。
上述方案中，所述从所述第一原始结构图的所述资源中选取与疾病信息和/或治疗信息相关的所述资源，构成第二原始结构图，包括：
基于所述第一原始结构图中所述资源，获取与所述疾病信息和/或治疗信息相关的所述资源为第三资源；
将所述第三资源确定为源节点以及与所述源节点具有连接关系的目的节点，其中，所述源节点在所述第一原始结构图中可达所述目的节点；
基于所述源节点在所述第一原始结构图中达到所述目的节点所经过资源数量，确定所述源节点与所述目的节点的权重值；
基于各所述源节点和各所述目的节点的连接关系，以及各所述源节点与所述各目的节点的权重值，确定所述第二原始结构图。
上述方案中，所述方法包括：
确定所述第二候选结构子图的支持度；
所述若所述第二候选结构子图满足预定条件，基于所述第二候选结构子图，确定目标结构图，包括：
若所第二候选结构子图的所述支持度小于或等于第一预定值，基于所述第二候选结构子图确定所述目标结构图。
上述方案中，所述确定所述第二候选结构子图的支持度，包括：
若所述第一原始结构图中包含所述第二候选结构子图中所包含的所有所述资源，确定所述第一原始结构图为第一相似结构图；根据所述第一相似结构图的数量与所述第一原始结构图的数量的比值，确定所述第二候选结构子图的所述支持度；
或者，
若第二原始结构图中包含所述第二候选结构子图中所包含的所有所述资源，确定所述第二原始结构图为第二相似结构图；根据所述第二相似结构图的数量与所述第二原始结构图的数量的比值，确定所述第二候选结构子图的所述支持 度。
第二方面，本公开实施例提供一种医疗信息处理装置，所述装置包括：
获取模块，配置为获取至少一个第一原始结构图，其中，所述第一原始结构图用于描述资源与资源之间的拓扑关系；其中，一种所述资源用于描述一种医疗信息；
处理模块，配置为从所述第一原始结构图中获取包括至少两种所述资源的第一候选结构子图；
所述处理模块，配置为基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，其中，所述选型矩阵用于确定第一候选结构子图得到第二候选结构子图所增加的候选资源；
确定模块，配置为若所述第二候选结构子图满足预定条件，基于所述第二候选结构子图，确定目标结构图。
第三方面，本公开实施例提供一种设备，所述设备包括处理器和用于存储能够在所述处理器上运行的计算机程序的存储器；其中，所述处理器用于运行计算机程序时，实现本公开任意实施例所述医疗信息处理方法。
第四方面，本公开实施例还提供了计算机一种存储介质，所述计算机存储介质中有计算机可执行指令，所述计算机可执行指令被处理器执行实现本公开任意实施例所述医疗信息处理方法。
在本公开实施例中，终端获取至少一个第一原始结构图，从第一原始结构图中获取包括至少两种资源的第一候选结构子图；并基于第一候选结构子图及选型矩阵，确定第二候选结构子图，若第二候选结构子图满足预定条件，基于第二候选结构子图确定目标结构图。由于该选型矩阵可以是确定第一候选结构子图得到第二候选结构子图所增加的候选资源，因此可以通过选型矩阵对节点增方向进行限制，从而可以得到具有指向性较强的第二候选结构子图；如此可以由第二候选结构子图结构相对紧密的目标结构图，能够提供在医学应用领域更有指向性的数据基础，且提高了医疗信息的挖掘结果的聚焦性。
附图说明
此处的附图被并入说明书中构成本说明书的一部分，这些附图示出了符合本申请的实施例，并与说明书一起用于说明本申请的技术方案。
图1为本公开实施例提供的一种医疗信息处理方法的流程示意图。
图2为本公开实施例提供的第一种第一原始结构图的示意图。
图3为本公开实施例提供的第二种第一原始结构图的示意图。
图4为本公开实施例提供的第三种第一原始结构图的示意图。
图5为本公开实施例提供的第四种第一原始结构图的示意图。
图6为本公开实施例提供的一种第二原始结构图的示意图。
图7为本公开实施例提供的另一种医疗信息处理方法的流程示意图。
图8为本公开实施例提供的一种医疗信息处理装置的结构示意图；
图9为本公开实施例提供的一种设备的硬件结构示意图。
具体实施方式
以下结合附图及实施例，对本公开进行进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本公开，并不用于限定本公开。
在后续的描述中，使用用于表示元件的诸如“模块”、“部件”或“单元”的后缀仅为了有利于本公开的说明，其本身没有特定的意义。因此，“模块”、“部件”或“单元”可以混合地使用。另外，在后续的表述中，使用用于标识信息的诸如“第一”或者“第二”等的前缀仅为了有利于本公开的说明，其本身没有特定的意义。另外，在后续的描述中，“多个”是指两个或者两个以上；“多种”是指两种或者两种以上。
如图1所示，本公开实施例提供了一种医疗信息处理方法，包括以下步骤：
步骤S11：获取至少一个第一原始结构图，其中，所述第一原始结构图用于描述资源与资源之间的拓扑关系；其中，一种所述资源用于描述一种医疗信息；
步骤S13：从所述第一原始结构图中获取包括至少两种所述资源的第一候 选结构子图；
步骤S15：基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，其中，所述选型矩阵用于确定第一候选结构子图得到第二候选结构子图所增加的候选资源；
步骤S17：若所述第二候选结构子图满足预定条件，基于所述第二候选结构子图，确定目标结构图。
本公开实施例提供的医疗信息处理方法，可由终端执行；该终端可以是任意一种移动终端或者固定终端。例如，该终端可以是但不限于是以下至少之一：手机、计算机、服务器、医疗设备、工业设备和/或穿戴式设备等。
这里，资源可以是任一种医疗信息或者与任意一种医疗信息相关的数据结构；例如，该资源可以是但不限于是以下医疗信息或者与医疗信息相关的数据结构：疾病信息、治疗信息、管理信息、支付信息及工作流程信息。这里，资源可以是一种最小的数据单元。
示例性的，第一原始结构图、第一候选结构子图、第二候选结构子图以及下述涉及的第二原始结构图中包括的资源均可以是两种或者两种以上。
在一个实施例中，步骤S11中获取的多个第一原始结构图中，任意两个第一原始结构图的至少部分资源不同，和/或任意两个第一原始结构图中至少部分资源所描述的医疗信息不同。
示例性的，快速医疗互操作资源(Fast Healthcare Interoperability Resources，FHIR)将医疗信息解耦重构为资源；该FHIR Release 4一共发布了145种资源，一种资源对应一种医疗信息的数据结构。例如，体征记录(Observation)资源定义了描述症状信息的数据结构；治疗计划(CarePlan)资源定义了描述治疗计划的数据结构等等。
示例性的，各类医疗信息可关联一种或者多种资源；例如，如下表1所示，疾病信息可关联但不限于以下至少之一的资源：过敏反应(AllergyIntolerance)资源、诊断记录(Condition)资源、体征记录(Observation)资源、诊断报告(DiagnosticReport)资源、问卷答复(QuestionnaireResponse)资源、家族病史 (FamilyMemberHistory)资源、免疫接种(Immunization)资源及风险评估(RiskAssessment)资源；治疗信息可以关联但不限于以下至少之一的资源：医疗申请(MedicationRequest)资源、服务申请(ServiceRequest)资源、营养订单(NutritionOrder)资源、程序(Procedure)资源、设备申请(DeviceRequest)资源以及治疗计划(CarePlan)资源。这里，表1中“释义”用于表示各资源所描述的含义；例如，家族病史资源，用于描述家族病史；又如，服务申请资源，用于描述检验和/或检查处方。
表1
可以理解的是，上述表1中的每一个元素都是独立存在的，这些元素被示 例性的列在同一张表格中，但是并不代表表格中所有元素必须根据表格所示的同时存在。其中每一个元素的值，是不依赖表1中任何其它元素值。因此本领域技术人员可以理解，该表1的每一个元素的取值都是一个独立的实施例。
上述资源的描述仅是示例，并不对资源的种类及个数等作限制。
这里，第一原始结构图、第一候选结构子图、第二候选结构子图、目标结构图及下述的第二原始结构图均可以是任一种描述资源与资源之间连接关系或者拓扑关系的结构图。该连接关系或者拓扑关系可用于指示资源与资源之间是相关联的。
示例性的，第一原始结构图可以是资源与资源之间具有连接关系的结构图。例如，第一原始结构图可以如图2所示，该第一原始结构图中资源包括体征记录1资源、患者资源、服务请求资源、就诊事件资源、诊断记录资源以及体征记录2资源；该第一原始结构图还包括观察资源与患者资源的连接关系、以及服务申请资源与患者资源的连接关系等。
示例性的，第二原始结构图可以是资源与资源之间具有拓扑关系的结构图。例如，第一原始结构图可以如图3所示，该第一原始结构图可以是一个拓扑结构图；该拓扑结构图中“1”、“2”、“3”、“4”、“5”及“6”可以分别指示一种资源，例如可以分别用于指示如图2中的体征记录1资源、患者资源、诊断记录资源、就诊事件资源、服务申请资源以及体征记录2资源。该第一候选结构子图中箭头可以用于描述资源与资源之间的连接关系；例如“1”与“2”之间的箭头用于描述“1”与“2”所指示的资源是相关联的。
在一个实施例中，第一原始结构图、第一候选结构子图、第二候选结构子图、目标结构图及第二原始结构图均可以是有向结构图。例如，如图2和图3所示，资源与资源之间的连接的箭头是有方向的；图3中“1”与“2”之间的箭头用于描述“1”所指示的资源是可达“2”所指示的资源的。
在一个实施例中，步骤S13，包括：从所述第一原始结构图中获取任意至少两种资源以构成第一候选结构子图。
示例性的，如图3所示的第一原始结构图，可以从中选取资源“1”和资源 “2”构成第一候选结构子图。
在一个实施例中，所述方法包括：确定选型矩阵。这里，终端可以基于各资源与资源之间的相关性确定选型矩阵，或者，终端可以基于历史结构图确定选型矩阵；或者，终端可以基于第一原始结构图确定选型矩阵。这里，只需确定的选型矩阵中对角元素用于描述各资源自身的关联性，以及选型矩阵中对角元素之外的元素用于描述资源与资源之间的关联性。该关联性可以用权重值表示。
示例性的，第一原始结构图中包括资源“1”、资源“2”、资源“3”、资源“4”、资源“5”及资源“6”；该选型矩阵可以是如下公式所示：其中，对角元素均为“∞”，分别用于指示资源“1”、资源“2”、资源“3”、资源“4”、资源“5”及资源“6”自身的关联性。对角元素以外的元素用于指资源与资源之间的关联性；例如，第1行第2列的元素用于指示资源“1”与资源“2”之间的关联性，资源“1”到资源“2”直接连接；则可确定资源“1”与资源“2”之间关联性的权重值为1。又如，第3行第5列的元素用于指示资源“3”与资源“5”之间的关联性，资源“3”到资源“5”需经过一个资源；则可确定资源“1”与资源“5”之间关联性的权重值为再如，第5行第1列的元素用于指示资源“6”与资源“1”之间的关联性，资源“6”不可连接到资源“1”，则可确定资源“6”与资源“1”之间关联性的权重值为0。
在一个实施例中，第一候选结构子图中包括的资源为第一资源；候选矩阵 中资源包括第一资源和第二资源；候选矩阵用于从第二资源中选择一个与第一候选结构子图关联性最强的第二资源作为候选资源。
示例性的，在上述第一原始结构图包括资源“1”、资源“2”、资源“3”、资源“4”、资源“5”及资源“6”示例中，若确定第一候选结构子图为由资源“1”及资源“2”确定的结构图，则选型矩阵用于在资源“3”、资源“4”、资源“5”及资源“6”中确定候选资源。由于资源“3”分别与资源“1”和资源“2”的关联性的权重值分别为和资源“4”分别与资源“1”和资源“2”的关联性的权重值分别为1和资源“5”分别与资源“1”和资源“2”的关联性的权重值分别为和1，资源“6”分别与资源“1”和资源“2”的关联性的权重值分别为1和因此确定资源“4”为候选资源。
在本公开实施例中，可以重复执行步骤S15，不断得到新第一候选结构子图及第二候选结构子图，直到得到的第二候选结构子图满足预定条件后，基于该第二候选结构子图，确定目标结构图。例如，在重复执行步骤S15时，可以是：第一次执行步骤S15后，得到第一个第二候选结构子图；将第一个第二候选结构子图作为第二次执行步骤S15的第一候选结构子图；以此类推，将第i-1次执行步骤S15后得到的第i个第二候选结构子图，作为第i次执行步骤S15的第一候选结构子图；其中i为大于1的整数。
在一个实施例中，步骤S17中第二候选结构子图满足预定条件可以是但不限于是至少之一：
第二候选结构子图的支持度小于或等于第一预定值；
第二候选结构子图中包含的资源数量大于或等于第二预定值。
在一个实施例中，所述方法包括：确定所述第二候选结构子图的支持度；
所述步骤S17，包括：若所第二候选结构子图的所述支持度小于或等于第一预定值，基于所述第二候选结构子图确定所述目标结构图。
在一个实施例中，所述确定所述第二候选结构子图的支持度，包括：
若所述第一原始结构图中包含所述第二候选结构子图中所包含的所有所述资源，确定所述第一原始结构图为第一相似结构图；
根据所述第一相似结构图的数量与所述第一原始结构图的数量的比值，确定所述第二候选结构子图的所述支持度。
这里，若第一原始结构图中包括资源“1”、“2”、“3”及“4”，且第二候选结构子图包括资源“1”及“2”；则第一原始结构图中包含第二候选结构子图中所包含的所有资源，该第一原始结构图为第一相似结构图。或者，若第一原始结构图中包括资源“1”、“2”、“3”及“4”，且第二候选结构子图包括资源“1”、“4”及“5”；则第一原始结构图中不包含第二候选结构子图中所包含的所有资源，该第一原始结构图不是第一相似结构图。
这里，支持度可以是第一相似结构图的数量与第一原始结构图的数量的比值；或者支持度也可以是：与第一相似结构图的数量与第一原始结构图的数量的比值相差预定值的值。
这里，第一预定值及第二预定值均可以是终端基于用户输入操作确定，或者基于历史经验信息确定或者预先设置的。该第一预定值可以是90％或者80％或者70％等。该第二预定值可以是100、200、1000或者10000等。
在一个实施例中，步骤S17中基于所述第二候选结构子图，确定目标结构图，包括：
确定满足所述预定条件的第二候选结构子图为所述目标结构图；
或者，
确定满足所述预定条件的第二候选结构子图所对应的第一候选结构子图为所述目标结构图。
示例性的，提供一种确定第二候选结构子图的支持度的计算公式：其中，G-SUP(gi)为第二候选结构子图的支持度，|{Gi|g∈Gi}|为第一相似结构图的数量；|G|为第一原始结构图的数量。
在本公开实施例中，终端获取至少一个第一原始结构图，从第一原始结构 图中获取包括至少两种资源的第一候选结构子图；并基于第一候选结构子图及选型矩阵，确定第二候选结构子图，若第二候选结构子图满足预定条件，基于第二候选结构子图确定目标结构图。由于该选型矩阵可以是确定第一候选结构子图得到第二候选结构子图所增加的候选资源，因此可以通过选型矩阵对节点增方向进行限制，从而可以得到具有指向性较强的第二候选结构子图；如此可以由第二候选结构子图结构相对紧密的目标结构图，能够提供在医学应用领域更有指向性的数据基础，且提高了医疗信息的挖掘结果的聚焦性。
并且，在本公开实施例中，若预定条件为支持度小于或等于第一预定值，才基于第二候选结构子图确定目标结构图；如此可以挖掘出相对较大或者最大的频繁子图作为目标结构图的，该相对较大的或者最大的频繁子图能够在内容上贴合病程挖掘目标，还能进一步保证结构的紧密性及聚焦性；如此得到的目标结构图可为医学应用领域提供重要的数据支撑和思维引导。
在一些实施例中，所述方法包括：基于所述第一原始结构图及所述第一候选结构子图，确定所述选型矩阵；其中，所述选型矩阵中对角线元素用于描述每个所述资源的一阶相似度；所述选型矩阵中对角线元素之外的元素用于描述所述资源与所述资源之间的二阶相似度。
在一个实施例中，一阶相似度可以基于资源的入度与出度之和确定；二阶相似度是指资源与邻接资源的一阶相似度确定。这里，一个资源的邻接资源是指在结构图中与该资源直接相连的资源；例如，如图3所示，“1”所指示的资源可以为“2”所指示的资源的邻接资源。
在另一个实施例中，一阶相似度可以基于资源与第一邻接资源权重值及资源与第二邻接资源的权重值之和确定；二阶相似度可以基于资源与第一邻接资源的权重值确定或者基于资源与第二邻接资源的权重值确定；其中，第一邻接资源为进入资源的邻接资源，第二邻接资源为离开资源的邻接资源。示例性的，如图4所示，“2”所指示的资源为“1”所指示的资源的第二邻接资源；“4”所指示的资源为“1”所指示的资源的第一邻接资源。
在本公开实施例中，通过确定选型矩阵中对角元素为各资源的一阶相似度， 以及对角元素之外的元素为资源与资源的二阶相似度，可以确定出结构紧密的选型矩阵，使得可增长的候选资源为使得结构图更加紧密的资源。
在一些实施例中，所述第一候选结构子图中所述资源为第一资源；所述选型矩阵中所述资源包括所述第一资源和第二资源；
所述步骤S15中基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，包括：
基于所述选型矩阵中各所述第二资源的偏向度，确定所述偏向度最大的所述第二资源为所述候选资源；所述偏向度为所述第二资源分别与所述第一资源之间的所述二阶相似度之和；
基于所述第一候选结构子图及所述候选资源，确定所述第二候选结构子图。
这里，二阶相似度可以是资源到资源的关联性的权重值之和。
在本公开实施例中，可以选择偏向度最大的第二资源为候选资源，从而可以使得增长的候选资源为使得结构图相对最为紧密的资源，从而有利于确定更为紧密结构的第二候选结构子图。
当然，在其它实施例中，也可以是确定偏向度次大的第二资源为候选资源，或者确定偏向度为排名前几位的一个第二资源为候选资源等；也能在一定程度上确定使得结构图相对较为紧密资源。
在一些实施例中，所述方法包括：
从所述第一原始结构图的所述资源中选取与疾病信息和/或治疗信息相关的所述资源，构成第二原始结构图；
所述步骤S13，包括：
从所述第二原始结构图中获取包括至少两种所述资源的所述第一候选结构子图。
示例性的，第一原始结构图中包括10种资源；该10种资源是与疾病信息、治疗信息或者支付信息等相关的资源，其中，有5种资源是与疾病信息或者治疗信息相关的资源；终端可以确定该5中资源构成的结构图为第二原始结构图。
在一些实施例中，所述从所述第一原始结构图的所述资源中选取与疾病信 息和/或治疗信息相关的所述资源，构成第二原始结构图，包括：
基于所述第一原始结构图中所述资源，获取与所述疾病信息和/或治疗信息相关的所述资源为第三资源；
将所述第三资源确定为源节点以及与所述源节点具有连接关系的目的节点，其中，所述源节点在所述第一原始结构图中可达所述目的节点；
基于所述源节点在所述第一原始结构图中达到所述目的节点所经过资源数量，确定所述源节点与所述目的节点的权重值；
基于各所述源节点和各所述目的节点的连接关系，以及各所述源节点与所述各目的节点的权重值，确定所述第二原始结构图。
这里，疾病信息和/或治疗信息相关的资源为第三资源；第三资源可分为源节点及目的节点。这里，源节点及目的节点均可以是第三资源中的任意一个资源；同一个第三资源可同时为源节点或者目的节点。
示例性的，如图2及图3所示，第一原始结构图可包括“1”、“2”、“3”、“4”、“5”及“6”分别对应的资源，该“1”、“2”、“3”、“4”、“5”及“6”分别对应体征记录资源、患者资源、诊断记录资源、就诊事件资源、服务申请资源以及体征记录资源。其中，“1”、“3”、“5”及“6”可以是与疾病信息或者治疗信息相关的资源，“2”及“4”则是与疾病信息或者治疗信息不相关的资源；如图5所示，在第一原始结构图中标注出与疾病信息或者治疗信息相关的资源“1”、“3”、“5”及“6”。终端可以根据“1”、“3”、“5”及“6”在第一原始结构图中的可达性和最短路径原则重新连接保留资源，连接方式可如图6所示。
基于图6的连接方式可以是：若源节点可达目的节点，则从源节点到目的节点增加一条有向边；给有向边赋予权重值；该权重值可基于源节点到达目的节点所经过的资源确定。如图5所示，例如，若源节点达到的目的节点为自身，则确定该源节点的权重值为∞；又如，若源节点“1”在第一原始结构图中达到目的节点“3”无需经过其它资源，即经过0个资源，则确定源节点“1”到目的节点“3”的权重值为1；再如，若源节点“3”在第一原始结构图中达到目的节点“1”需要经过2个资源，则确定源节点“3”到目的节点“1”的权重值 为再如，若源节点“5”在第一原始结构图中达到目的节点“3”需经过3个资源，则确定源节点“3”到目的节点“1”的权重值为这里，该赋予权重值后的第二原始结构的等价邻接矩阵可如下矩阵所示：其中，各对角元素为各资源同时为源节点及目的节点的权重值；对角元素之外的元素为各资源(即源节点)到邻接资源(即目的节点)的权重值。
如此，在本公开实施例中，可以从第一原始结构图中去掉一些与疾病信息和/或治疗信息无关的资源，保留疾病信息和/或治疗信息相关的资源信息；并通过保留下来的资源在第一原始结构图中的连接关系确定该些资源在第二原始结构图中连接关系，如此可以由第一原始结构图简化得到等价的第二原始结构图，有利于后续简化确定选型矩阵以及选取目标结构图的计算。
在本公开实施例中，由于得到的第二原始结构图的等价邻接矩阵是基于保留的疾病信息和/或治疗信息的资源在第一原始结构图中连接所经过的资源确定的权重值确定，从而可以保留第一原始结构图的拓扑信息。
在一些实施例中，所述步骤S15中基于所述第一原始结构图及所述第一候选结构子图，确定所述选型矩阵，包括：
基于所述第二原始结构图及所述第一候选结构子图，确定所述选型矩阵。
在上述包括资源“1”、“3”、“5”及“6”的第二原始结构图的示例中，该第二原始结构图的等价邻接矩阵为可以从第二原始结构制 图中选取资源包括资源“1”及“3”的第一候选结构子图；第二候选结构子图对的等价邻接矩阵为由该等价邻接矩阵Madj2可确定选型矩阵为其中，该选型矩阵中对角线元素用于描述每个所述资源的一阶相似度；所述选型矩阵中对角线元素之外的元素用于描述所述资源与所述资源之间的二阶相似度。这里，选型矩阵中第1行第1列的对角元素用于描述资源“1”的一阶相似度，基于第二原始结构图(如图6)及等价邻接矩阵可知，资源“1”与离开资源“1”的第二邻接资源“3”、“5”及“6”的权重值分别为1、及以及资源“1”与进入资源“1”的第一邻接资源“3”、“5”的权重值分别为可得到第1行第1列的描述资源“1”的一阶相似度为这里，选型矩阵第1行第2列的元素用于描述资源“1”到资源“2”的二阶相似度，该二阶相似度可为资源“2”的一阶相似度以及选型矩阵第2行第1列的元素用于描述资源“2”到资源“1”的二阶相似度，该二阶相似度可为资源“1”的一阶相似度
该第一候选结构子图的对应的选型矩阵部分为选型矩阵中包括的第二资源为资源“5”及“6”；计算各第二资源的偏向度可以是其中，m为第一候选结构子图的资源的数量，Di为第i个资源 的偏向度。终端可以确定资源“5”的偏向度以及确定资源“6”的偏向度D6＝Mselect(6,1)+Mselect(6,3)＝0+0＝0。终端确定最大的偏向度Vnext＝maxiDi；这里，第二资源“5”及“6”的偏向度中最大的偏向度为第二资源“5”的偏向度，则确定第二资源“5”为候选资源。终端可基于第第一候选结构子图及候选资源“5”确定第二候选结构子图。
本公开实施例是基于第二原始结构图及第一候选结构子图，确定的选型矩阵，如此基于该候选矩阵确定的第二候选结构子图，是可以在一定程度上保留原第一原始结构图的拓扑信息的前提下能够减少相关性不强的资源，从而能够相对较为完整的数据模型的基础上大大简化计算。
在一些实施例中，所述确定所述第二候选结构子图的支持度，包括：
若第二原始结构图中包含所述第二候选结构子图中所包含的所有所述资源，确定所述第二原始结构图为第二相似结构图；
根据所述第二相似结构图的数量与所述第二原始结构图的数量的比值，确定所述第二候选结构子图的所述支持度。
这里，若第二原始结构图中包括资源“1”、“2”、“5”及“6”，且第二候选结构子图包括资源“1”及“3”；则第二原始结构图中包含第二候选结构子图中所包含的所有资源，该第二原始结构图为第二相似结构图。或者，若第二原始结构图中包括资源“1”、“3”、“5”及“6”，且第二候选结构子图包括资源“1”、“3”及“4”；则第二原始结构图中不包含第二候选结构子图中所包含的所有资源，该第二原始结构图不是第二相似结构图。
这里，支持度可以是第二相似结构图的数量与第二原始结构图的数量的比值；或者支持度也可以是：与第二相似结构图的数量与第二原始结构图的数量的比值相差预定值的值。
本公开实施例中是基于第二原始结构图确定选型矩阵及第一候选结构子图，如此可确定为第二原始结构图为比对样本；基于该第二原始结构图确定的候选结构图子图的支持度，也能确定相对较为准确的支持度。
为了进一步解释本公开任意实施例，以下提供一个具体实施例。
如图7所示，本公开实施例提供一种医疗信息处理方法，由终端执行，包括以下步骤：
步骤S21：终端获取至少一个第一原始结构图；
这里，终端可利用FHIR对医疗信息进行建模，建立每个患者的一个第一原始结构图；该第一原始结构图可如图2所示的体征记录1资源、患者资源、诊断记录资源、就诊事件资源、服务申请资源以及体征记录2资源。终端该第一原始结构图转换为如图3所示的拓扑模型，该体征记录1资源、患者资源、诊断记录资源、就诊事件资源、服务申请资源以及体征记录2资源分别用图3中的“1”、“2”、“3”、“4”、“5”及“6”指示。该图2中资源与资源连接关系，即资源与资源的引用(Reference)的关联性。
步骤S22：终端从所述第一原始结构图的所述资源中选取与疾病信息和/或治疗信息相关的所述资源，构成第二原始结构图；
这里，终端从第一原始结构图中体征记录1资源、患者资源、诊断记录资源、就诊事件资源、服务申请资源以及体征记录2资源中，选取与疾病信息和/或治疗信息相关的资源：体征记录1资源、诊断记录资源、服务申请资源以及体征记录2资源；得到该体征记录1资源、诊断记录资源、服务申请资源以及体征记录2资源的第二原始结构图可如图6所示。这里，该第二原始结构图中，若源节点可达目的节点，则从源节点到目的节点增加一条有向边；给有向边赋予权重值。
终端可基于可确定该第二原始结构图的等价邻接矩阵为：
步骤S23：终端确定选型矩阵及确定增加的候选资源；
这里，确定增加的候选资源的方法，可认为是偏置型节点增长方法。
这里，终端基于第二原始结构图，确定第一候选结构子图为包括资源“1”及“3”的结构图；终端基于第二原始结构图的等价邻接矩阵Madj2确定选型矩阵为其中，该选型矩阵中对角线元素用于描述每个所述资源的一阶相似度；所述选型矩阵中对角线元素之外的元素用于描述所述资源与所述资源之间的二阶相似度。由选型矩阵Mselect可知，资源“5”及“6”是节点增长的两个待选方向，即该资源“5”及“6”可以作为被选到作为第二候选结构子图中的候选资源。假设挖掘目标是结构紧密型子图，那么需要根据Mselect选出资源“5”及“6”中与第一候选结构子图耦合更紧密的资源，选择依旧可依据最大偏向度确定：Vnext＝maxiDi；其中，其中，m为第一候选结构子图的资源的数量，Di为第i个资源的偏向度。这里，可以确定资源“5”为候选资源，即资源“5”是节点增长需要的方向。
步骤S24：终端根据第一候选结构子图及候选资源，确定第二候选结构子图。
步骤S25：终端确定第二候选结构子图的支持度；
这里，终端确定第二候选结构子图的支持度其中，G'-SUP(gi)为第二候选结构子图的支持度，|{G'i|g∈G'i}|为第二相似结构图的数量；|G'|为第二原始结构图的数量。
步骤S26：终端确定第二候选结构子图的支持度是否小于或等于第一预定值，若否，执行步骤S27；若是，执行步骤S28；
这里，终端确定第一预定值为G-SUPmin；若G'-SUP(gi)≤G-SUPmin，则执行步骤S27，否则，执行步骤S28。
步骤S27：终端确定第二候选结构子图为第一候选结构子图；
这里，终端将第二候选结构子图作为下次需要增加候选资源的第一候选结构子图。
步骤S28：终端基于第二候选结构子图确定目标结构图。
这里，终端可将支持度小于或等于预定值时的第二候选结构子图作为目标结构图。或者将该第二候选结构子图确定最后一个增加的候选资源的结构图(即该第二候选结构子图对应的第一候选结构子图)为目标结构图；此时的目标结构图为包括资源数最多的最大频繁子图。最大频繁子图不仅能够在内容上贴合病程挖掘目标，还能保证结构的紧密性及聚焦性。
在本公开实施例中，可以去除与疾病信息和/或治疗信息无关的资源，并通过等价邻接矩阵在一定程度上保留原来治疗信息的拓扑关系，可以在简化计算的同时为后续目标结构图的挖掘提供相对更为完整的数据模型。
在本公开实施例中，可以定义一个选型矩阵限制对资源的增长限制方向，可以控制第二候选结构子图的偏向性能力；可以通过选择偏向度最大的资源作为候选节点，从而可以选择与第一候选结构子图耦合更紧密的资源增加在第二候选结构子图中。如此可以使得挖掘的目标结构图能够偏向于结构紧密型拓扑，提高挖掘结构的聚焦性。
并且，本公开还能实现等价拓扑结构的转换(即由第一原始结构图转换为等价的第二原始结构图)以及特殊节点增长方式(即控制具有偏向度最大或者相对较大的候选资源为第一候选结构子图变换到第二候选结构子图所需增加的资源)，实现了二者的有效融合。
这里需要指出的是：以下信息处理装置项的描述，与上述医疗信息处理方法项描述是类似的，同方法的有益效果描述，不做赘述。对于本公开信息处理装置实施例中未披露的技术细节，请参照本公开医疗信息处理方法实施例的描 述。
如图8所示，本公开实施例提供一种医疗信息处理装置，所述装置包括：
获取模块41，配置为获取至少一个第一原始结构图，其中，所述第一原始结构图用于描述资源与资源之间的拓扑关系；其中，一种所述资源用于描述一种医疗信息；
处理模块42，配置为从所述第一原始结构图中获取包括至少两种所述资源的第一候选结构子图；
所述处理模块42，配置为基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，其中，所述选型矩阵用于确定第一候选结构子图得到第二候选结构子图所增加的候选资源；
确定模块43，配置为若所述第二候选结构子图满足预定条件，基于所述第二候选结构子图，确定目标结构图。
在一些实施例中，所述装置包括：
所述处理模块42，配置为基于所述第一原始结构图及所述第一候选结构子图，确定所述选型矩阵；其中，所述选型矩阵中对角线元素用于描述每个所述资源的一阶相似度；所述选型矩阵中对角线元素之外的元素用于描述所述资源与所述资源之间的二阶相似度。
在一些实施例中，所述第一候选结构子图中所述资源为第一资源；所述选型矩阵中所述资源包括所述第一资源和第二资源；
所述处理模块42，配置为基于所述选型矩阵中各所述第二资源的偏向度，确定所述偏向度最大的所述第二资源为所述候选资源；所述偏向度为所述第二资源分别与所述第一资源之间的所述二阶相似度之和；
基于所述第一候选结构子图及所述候选资源，确定所述第二候选结构子图。
在一些实施例中，所述获取模块41，配置为从所述第一原始结构图的所述资源中选取与疾病信息和/或治疗信息相关的所述资源，构成第二原始结构图；
所述处理模块42，配置为从所述第二原始结构图中获取包括至少两种所述资源的所述第一候选结构子图；
所述处理模块42，还配置为基于所述第二原始结构图及所述第一候选结构子图，确定所述选型矩阵。
在一些实施例中，所述处理模块42，配置为执行以下步骤：
基于所述第一原始结构图中所述资源，获取与所述疾病信息和/或治疗信息相关的所述资源为第三资源；
将所述第三资源确定为源节点以及与所述源节点具有连接关系的目的节点，其中，所述源节点在所述第一原始结构图中可达所述目的节点；
基于所述源节点在所述第一原始结构图中达到所述目的节点所经过资源数量，确定所述源节点与所述目的节点的权重值；
基于各所述源节点和各所述目的节点的连接关系，以及各所述源节点与所述各目的节点的权重值，确定所述第二原始结构图。
在一些实施例中，所述处理模块42，配置为确定所述第二候选结构子图的支持度；
所述确定模块43，配置为若所第二候选结构子图的所述支持度小于或等于第一预定值，基于所述第二候选结构子图确定所述目标结构图。
在一些实施例中，所述处理模块42，配置为若所述第一原始结构图中包含所述第二候选结构子图中所包含的所有所述资源，确定所述第一原始结构图为第一相似结构图；
所述处理模块42，还配置为根据所述第一相似结构图的数量与所述第一原始结构图的数量的比值，确定所述第二候选结构子图的所述支持度。
在一些实施例中，所述处理模块42，配置为若第二原始结构图中包含所述第二候选结构子图中所包含的所有所述资源，确定所述第二原始结构图为第二相似结构图；
所述处理模块42，还配置为根据所述第二相似结构图的数量与所述第二原始结构图的数量的比值，确定所述第二候选结构子图的所述支持度。
如图9所示，本公开实施例还提供了一种设备，所述设备包括处理器61 和用于存储能够在所述处理器61上运行的计算机程序的存储器62；其中，所述处理器61用于运行计算机程序时，实现本公开任意实施例的医疗信息处理方法。
在一些实施例中，本公开实施例中的存储器可以是易失性存储器或非易失性存储器，或可包括易失性和非易失性存储器两者。其中，非易失性存储器可以是只读存储器(Read-Only Memory，ROM)、可编程只读存储器(Programmable ROM，PROM)、可擦除可编程只读存储器(Erasable PROM，EPROM)、电可擦除可编程只读存储器(Electrically EPROM，EEPROM)或闪存。易失性存储器可以是随机存取存储器(Random Access Memory，RAM)，其用作外部高速缓存。通过示例性但不是限制性说明，许多形式的RAM可用，例如静态随机存取存储器(Static RAM，SRAM)、动态随机存取存储器(Dynamic RAM，DRAM)、同步动态随机存取存储器(Synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(Double Data Rate SDRAM，DDRSDRAM)、增强型同步动态随机存取存储器(Enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(Synchlink DRAM，SLDRAM)和直接内存总线随机存取存储器(Direct Rambus RAM，DRRAM)。本文描述的系统和方法的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
而处理器可能种集成电路芯片，具有信号的处理能力。在实现过程中，上述方法的各步骤可以通过处理器中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器可以是通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现成可编程门阵列(Field Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本公开实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本公开实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器， 可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器，处理器读取存储器中的信息，结合其硬件完成上述方法的步骤。
在一些实施例中，本文描述的这些实施例可以用硬件、软件、固件、中间件、微码或其组合来实现。对于硬件实现，处理单元可以实现在一个或多个专用集成电路(Application Specific Integrated Circuits，ASIC)、数字信号处理器(Digital Signal Processing，DSP)、数字信号处理设备(DSP Device，DSPD)、可编程逻辑设备(Programmable Logic Device，PLD)、现场可编程门阵列(Field-Programmable Gate Array，FPGA)、通用处理器、控制器、微控制器、微处理器、用于执行发明所述功能的其它电子单元或其组合中。
对于软件实现，可通过执行本文所述功能的模块(例如过程、函数等)来实现本文所述的技术。软件代码可存储在存储器中并通过处理器执行。存储器可以在处理器中或在处理器外部实现。
本公开又一实施例提供了一种计算机存储介质，该计算机可读存储介质存储有可执行程序，所述可执行程序被处理器执行时，可实现本公开任意实施例的医疗信息处理方法的步骤。
在一些实施例中，所述计算机存储介质可以包括：U盘、移动硬盘、只读存储器(ROM，Read Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
需要说明的是：本公开实施例所记载的技术方案之间，在不冲突的情况下，可以任意组合。
以上所述，仅为本公开的具体实施方式，但本公开的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本公开揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本公开的保护范围之内。因此，本公开的保护范围应以所述权利要求的保护范围为准。
工业实用性
本公开的实施例所提供的医疗信息处理方法以及装置、设备及存储介质，可以获取至少一个第一原始结构图，从第一原始结构图中获取包括至少两种资源的第一候选结构子图；并基于第一候选结构子图及选型矩阵，确定第二候选结构子图，若第二候选结构子图满足预定条件，基于第二候选结构子图确定目标结构图；由于该选型矩阵可以是确定第一候选结构子图得到第二候选结构子图所增加的候选资源，因此可以通过选型矩阵对节点增方向进行限制，从而可以得到具有指向性较强的第二候选结构子图；如此可以由第二候选结构子图结构相对紧密的目标结构图，能够提供在医学应用领域更有指向性的数据基础，且提高了医疗信息的挖掘结果的聚焦性。

Claims (10)

  1. 一种医疗信息处理方法，所述方法包括：
    获取至少一个第一原始结构图，其中，所述第一原始结构图用于描述资源与资源之间的拓扑关系；其中，一种所述资源用于描述一种医疗信息；
    从所述第一原始结构图中获取包括至少两种所述资源的第一候选结构子图；
    基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，其中，所述选型矩阵用于确定第一候选结构子图得到第二候选结构子图所增加的候选资源；
    若所述第二候选结构子图满足预定条件，基于所述第二候选结构子图，确定目标结构图。
  2. 根据权利要求1所述的方法，其中，所述方法包括：
    基于所述第一原始结构图及所述第一候选结构子图，确定所述选型矩阵；其中，所述选型矩阵中对角线元素用于描述每个所述资源的一阶相似度；所述选型矩阵中对角线元素之外的元素用于描述所述资源与所述资源之间的二阶相似度。
  3. 根据权利要求2所述的方法，其中，所述第一候选结构子图中所述资源为第一资源；所述选型矩阵中所述资源包括所述第一资源和第二资源；
    所述基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，包括：
    基于所述选型矩阵中各所述第二资源的偏向度，确定所述偏向度最大的所述第二资源为所述候选资源；所述偏向度为所述第二资源分别与所述第一资源之间的所述二阶相似度之和；
    基于所述第一候选结构子图及所述候选资源，确定所述第二候选结构子图。
  4. 根据权利要求2所述的方法，其中，所述方法包括：
    从所述第一原始结构图的所述资源中选取与疾病信息和/或治疗信息相关的所述资源，构成第二原始结构图；
    所述从所述第一原始结构图中获取包括至少两种所述资源的第一候选结构子图，包括：
    从所述第二原始结构图中获取包括至少两种所述资源的所述第一候选结构子图；
    所述基于所述第一原始结构图及所述第一候选结构子图，确定所述选型矩阵，包括：
    基于所述第二原始结构图及所述第一候选结构子图，确定所述选型矩阵。
  5. 根据权利要求4所述的方法，其中，所述从所述第一原始结构图的所述资源中选取与疾病信息和/或治疗信息相关的所述资源，构成第二原始结构图，包括：
    基于所述第一原始结构图中所述资源，获取与所述疾病信息和/或治疗信息相关的所述资源为第三资源；
    将所述第三资源确定为源节点以及与所述源节点具有连接关系的目的节点，其中，所述源节点在所述第一原始结构图中可达所述目的节点；
    基于所述源节点在所述第一原始结构图中达到所述目的节点所经过资源数量，确定所述源节点与所述目的节点的权重值；
    基于各所述源节点和各所述目的节点的连接关系，以及各所述源节点与所述各目的节点的权重值，确定所述第二原始结构图。
  6. 根据权利要求1或4所述的方法，其中，所述方法包括：
    确定所述第二候选结构子图的支持度；
    所述若所述第二候选结构子图满足预定条件，基于所述第二候选结构子图，确定目标结构图，包括：
    若所第二候选结构子图的所述支持度小于或等于第一预定值，基于所 述第二候选结构子图确定所述目标结构图。
  7. 根据权利要求6所述的方法，其中，所述确定所述第二候选结构子图的支持度，包括：
    若所述第一原始结构图中包含所述第二候选结构子图中所包含的所有所述资源，确定所述第一原始结构图为第一相似结构图；根据所述第一相似结构图的数量与所述第一原始结构图的数量的比值，确定所述第二候选结构子图的所述支持度；
    或者，
    若第二原始结构图中包含所述第二候选结构子图中所包含的所有所述资源，确定所述第二原始结构图为第二相似结构图；根据所述第二相似结构图的数量与所述第二原始结构图的数量的比值，确定所述第二候选结构子图的所述支持度。
  8. 一种医疗信息处理装置，所述装置包括：
    获取模块，配置为获取至少一个第一原始结构图，其中，所述第一原始结构图用于描述资源与资源之间的拓扑关系；其中，一种所述资源用于描述一种医疗信息；
    处理模块，配置为从所述第一原始结构图中获取包括至少两种所述资源的第一候选结构子图；
    所述处理模块，配置为基于所述第一候选结构子图及选型矩阵，确定第二候选结构子图，其中，所述选型矩阵用于确定第一候选结构子图得到第二候选结构子图所增加的候选资源；
    确定模块，配置为若所述第二候选结构子图满足预定条件，基于所述第二候选结构子图，确定目标结构图。
  9. 一种设备，所述设备包括处理器和用于存储能够在所述处理器上运行的计算机程序的存储器；其中，所述处理器用于运行计算机程序时，实现权利要求1至7任一项所述医疗信息处理方法。
  10. 一种计算机存储介质，所述计算机存储介质中有计算机可执行指令， 其中，所述计算机可执行指令被处理器执行实现权利要求1至7任一项所述医疗信息处理方法。
PCT/CN2023/123804 2022-11-15 2023-10-10 医疗信息处理方法以及装置、设备及存储介质 WO2024104005A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211429239.1 2022-11-15
CN202211429239.1A CN116959731A (zh) 2022-11-15 2022-11-15 医疗信息处理方法以及装置、设备及存储介质

Publications (1)

Publication Number Publication Date
WO2024104005A1 true WO2024104005A1 (zh) 2024-05-23

Family

ID=88455353

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/123804 WO2024104005A1 (zh) 2022-11-15 2023-10-10 医疗信息处理方法以及装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN116959731A (zh)
WO (1) WO2024104005A1 (zh)

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102750263A (zh) * 2012-05-31 2012-10-24 常州工学院 互联网超链接网络图数据的简化方法
CN104239126A (zh) * 2013-06-19 2014-12-24 华为技术有限公司 一种有向图的最小割获取方法及设备
CN104516995A (zh) * 2013-09-29 2015-04-15 华为技术有限公司 网络流图缩减方法和装置
CN106682514A (zh) * 2016-12-15 2017-05-17 哈尔滨工程大学 基于子图挖掘的系统调用序列特征模式集生成方法
US20200192368A1 (en) * 2018-12-18 2020-06-18 Aptiv Technologies Limited Methods for generating maps using hyper-graph data structures
CN113326404A (zh) * 2021-06-25 2021-08-31 深圳前海微众银行股份有限公司 数据流图处理方法、装置、设备及存储介质
CN113609345A (zh) * 2021-09-30 2021-11-05 腾讯科技（深圳）有限公司 目标对象关联方法和装置、计算设备以及存储介质
US11228505B1 (en) * 2021-01-29 2022-01-18 Fujitsu Limited Explanation of graph-based predictions using network motif analysis
CN114239198A (zh) * 2021-12-06 2022-03-25 国网湖北省电力有限公司电力科学研究院 一种基于并行优化的电网子图划分方法及装置
CN114283099A (zh) * 2020-09-21 2022-04-05 华为技术有限公司 一种图处理的方法，系统以及装置

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102750263A (zh) * 2012-05-31 2012-10-24 常州工学院 互联网超链接网络图数据的简化方法
CN104239126A (zh) * 2013-06-19 2014-12-24 华为技术有限公司 一种有向图的最小割获取方法及设备
CN104516995A (zh) * 2013-09-29 2015-04-15 华为技术有限公司 网络流图缩减方法和装置
CN106682514A (zh) * 2016-12-15 2017-05-17 哈尔滨工程大学 基于子图挖掘的系统调用序列特征模式集生成方法
US20200192368A1 (en) * 2018-12-18 2020-06-18 Aptiv Technologies Limited Methods for generating maps using hyper-graph data structures
CN114283099A (zh) * 2020-09-21 2022-04-05 华为技术有限公司 一种图处理的方法，系统以及装置
US11228505B1 (en) * 2021-01-29 2022-01-18 Fujitsu Limited Explanation of graph-based predictions using network motif analysis
CN113326404A (zh) * 2021-06-25 2021-08-31 深圳前海微众银行股份有限公司 数据流图处理方法、装置、设备及存储介质
CN113609345A (zh) * 2021-09-30 2021-11-05 腾讯科技（深圳）有限公司 目标对象关联方法和装置、计算设备以及存储介质
CN114239198A (zh) * 2021-12-06 2022-03-25 国网湖北省电力有限公司电力科学研究院 一种基于并行优化的电网子图划分方法及装置

Also Published As

Publication number Publication date
CN116959731A (zh) 2023-10-27

Similar Documents

Publication Publication Date Title
US11728030B2 (en) Methods of treatment and diagnosis using enhanced patient-physician communication
Senthilkumar et al. Big data in healthcare management: a review of literature
JP5744058B2 (ja) 医療ガイドラインへの患者データのマッピング
US20160132645A1 (en) System and architecture for providing shared patient data notifications
US20170011182A1 (en) Systems and Methods of Providing Patient Apps
JP2016517595A (ja) 個別化医療における情報の電子配信
WO2011133543A1 (en) Diagnostic system and method
US20150100327A1 (en) Maintaining context between applications utilizing a prioritized patient list
US9304761B2 (en) System and method for collaborative programming of data entry workflows between system developers, end users, and third party developers
US20190103193A1 (en) Normalization of medical terms
US20140195168A1 (en) Constructing a differential diagnosis and disease ranking in a list of differential diagnosis
US20230245790A1 (en) Data driven predictive analysis of complex data sets for determining decision outcomes
US20190103173A1 (en) Techniques for managing access of user devices to third-party resources
Harding et al. Candidate coethnicity, rural/urban divides, and partisanship in Africa
US20180101649A1 (en) Family tree construction supporting method and family tree construction supporting device
CN111052259B (zh) 用于在用户设备上使用医疗术语表达进行搜索的方法和设备
US10824684B2 (en) Techniques for anonymized searching of medical providers
WO2023273455A1 (zh) 医疗知识图谱的构建方法、装置、设备及计算机可读介质
WO2024104005A1 (zh) 医疗信息处理方法以及装置、设备及存储介质
Jerrett et al. Nature exposure gets a boost from a cluster randomized trial on the mental health benefits of greening vacant lots
US20180277258A1 (en) System for remote guidance of health care examinations
KR102251373B1 (ko) 임상연구를 위한 cdw 연구검색 시스템과 방법
US20150220704A1 (en) Clinical Decision Support System over a bipartite graph
US20200176127A1 (en) Systems and methods for guideline concordance
Khairallah Are the Standardization of the Uveitis Nomenclature (SUN) Working Group criteria for codifying the site of inflammation appropriate for all uveitis problems? Limitations of the SUN Working Group classification