CN117725468B

CN117725468B - Intelligent medical electric guarantee method and system

Info

Publication number: CN117725468B
Application number: CN202410171129.2A
Authority: CN
Inventors: 吴宇; 刘川; 张明赟; 饶沣菊; 冯豪
Original assignee: Sichuan Honglin Technology Co ltd
Current assignee: Sichuan Honglin Technology Co ltd
Priority date: 2024-02-06
Filing date: 2024-02-06
Publication date: 2024-04-26
Anticipated expiration: 2044-02-06
Also published as: CN117725468A

Abstract

According to the intelligent medical electric guarantee method and system, as the intelligent medical electric information queue of the intelligent medical electric information can be subjected to multi-aspect feature extraction to obtain a large amount of information such as important indication local features and important indication global features of the intelligent medical electric information, the method and system can further perform feature extraction on a reference event result corresponding to the intelligent medical electric information based on the important indication local features and the important indication global features to obtain key features, and the important indication local features, the important indication global features and the key features are used for obtaining target event features with deeper content representation between the intelligent medical electric information and the reference event result, so that the medical electric guarantee result of the intelligent medical electric information can be accurately regressed and analyzed based on the target event features.

Description

Intelligent medical electric guarantee method and system

Technical Field

The application relates to the technical field of data processing, in particular to an intelligent medical electric guarantee method and system.

Background

The invention provides an intelligent medical electric energy guarantee system which comprises an operating room, ICU, CCU, UPS power supplies, a storage battery, an IT isolation power supply, at least one area power supply unit for supplying power to medical equipment, an air conditioner automatic control system and a digital operating room system, wherein the area power supply unit and the digital operating room system are transmitted to a central controller through data connection, and the intelligent medical electric energy guarantee system can perfectly display all area data and can perfectly show and check the area data on a mobile phone or a computer through remote WIFI and 5G data.

In practice, a line abnormality or a lamp shortage may occur, and a power failure may occur, so that it is a technical problem that is difficult to solve at present, how to ensure the safety of the device and how to seamlessly connect the storage battery to ensure that the device continues to work.

Disclosure of Invention

In order to improve the technical problems in the related art, the application provides an intelligent medical electric guarantee method and system.

In a first aspect, an intelligent medical electrical assurance method is provided, including:

Obtaining intelligent medical electrical information to be processed and a reference event result corresponding to the intelligent medical electrical information, and classifying the intelligent medical electrical information to obtain at least one intelligent medical electrical information queue corresponding to the intelligent medical electrical information;

Extracting multiple aspects of characteristics from the intelligent medical electric information queue to obtain important indication local characteristics of each intelligent medical electric information and important indication global characteristics of the intelligent medical electric information;

combining the important indication local feature and the important indication global feature, extracting the features of the reference event result to obtain key features, wherein the key features represent the association relationship between the intelligent medical electrical information and the reference event result;

splicing the important indication local feature, the important indication global feature and the key feature to obtain a target event feature of the intelligent medical electrical information;

And generating a medical electric guarantee result corresponding to the intelligent medical electric information according to the target event characteristics and the reference event result.

In the application, the extracting of the multiple aspects of the intelligent medical electric information queue to obtain the important indication local feature of each intelligent medical electric information and the important indication global feature of the intelligent medical electric information includes:

Analyzing the positioning data of each intelligent medical electric information queue in the intelligent medical electric information, and adopting an intelligent medical electric information analysis thread to perform characteristic extraction on the positioning data to obtain the positioning characteristics of each intelligent medical electric information queue;

The intelligent medical information analysis thread is adopted to conduct feature extraction on the intelligent medical information queues, and original important indication local features of all the intelligent medical information queues and original important indication global features of the intelligent medical information are obtained;

splicing the positioning features and the original important indication local features to obtain important indication local features of each intelligent medical electric information queue;

and splicing the positioning feature and the original important indication global feature to obtain the important indication global feature of the intelligent medical electric information.

In the present application, before the intelligent medical electric information analysis thread is adopted to perform feature extraction on the intelligent medical electric information queue, the method further includes:

Obtaining a configuration data set, wherein the configuration data set comprises a plurality of intelligent medical electric information examples and power utilization abnormal examples corresponding to the intelligent medical electric information examples;

Classifying the intelligent medical electricity information examples and the electricity utilization abnormal examples respectively to obtain at least one intelligent medical electricity information range corresponding to the intelligent medical electricity information examples and at least one event cluster corresponding to the electricity utilization abnormal examples;

Adopting a specified intelligent medical electricity analysis thread to respectively perform feature extraction on the intelligent medical electricity information range and the event cluster to obtain important indication features of at least one important indication node corresponding to the intelligent medical electricity information example and event features of at least one event node corresponding to the electricity utilization abnormal example;

Performing feature recognition on the important indication feature and the event feature in at least one recognition node to obtain feature recognition results corresponding to all the recognition nodes, wherein the feature recognition results represent the recognition degree between the important indication feature and the event feature;

And converging the appointed intelligent medical information analysis thread according to the characteristic recognition result, the important indication characteristic and the event characteristic to obtain the intelligent medical information analysis thread.

In the present application, the feature recognition is performed on at least one recognition node by the important indication feature and the event feature to obtain feature recognition results corresponding to each recognition node, including:

Selecting target important indication nodes corresponding to all the identification nodes from the important indication nodes, and determining target event nodes corresponding to all the identification nodes from the event nodes;

Selecting a target important indication feature corresponding to the target important indication node from the important indication features, and extracting a target event feature corresponding to the target event node from the event features;

And identifying the target important indication features and the target event features to obtain feature identification results corresponding to all the identification nodes.

In the present application, the identifying the target important indication feature and the target event feature to obtain feature identification results corresponding to each identification node includes:

Determining at least one target identification node in the identification nodes; selecting undetermined important indication features corresponding to the target identification nodes from the target important indication features, and extracting undetermined event features corresponding to the target identification nodes from the target event features;

performing feature recognition on the undetermined important indication feature and the undetermined event feature at the target recognition node to obtain a target feature recognition result corresponding to the target recognition node;

And returning to the step of determining the target recognition node in the recognition nodes until each recognition node is the target recognition node, and obtaining the feature recognition result corresponding to each recognition node.

In the application, the undetermined important indication features comprise global abnormal features corresponding to the intelligent medical electrical information examples, and the undetermined event features comprise global event features corresponding to the electrical abnormality examples; the step of performing feature recognition on the undetermined important indication feature and the undetermined event feature at the target recognition node to obtain a target feature recognition result corresponding to the target recognition node, including:

When the target identification node is a global identification node, calculating an example commonality coefficient between the intelligent medical electricity information example and the electricity utilization abnormal example according to the global abnormal characteristics and the global event characteristics, wherein the global identification node is an identification node between the intelligent medical electricity information example and the electricity utilization abnormal example;

And splicing the example commonality coefficients to generate a global feature recognition result corresponding to the global recognition node, and determining the global feature recognition result as the target feature recognition result.

In the application, the undetermined important indication features further comprise range abnormal features corresponding to the intelligent medical electrical information range, and the undetermined event features further comprise cluster features corresponding to event clusters; the step of performing feature recognition on the undetermined important indication feature and the undetermined event feature at the target recognition node to obtain a target feature recognition result corresponding to the target recognition node, including:

when the target identification node is a range identification node, selecting a target intelligent medical information range from the intelligent medical information range, and selecting a target event cluster from the event clusters, wherein the range identification node is an identification node between the target intelligent medical information range and the target event cluster;

Calculating the confidence coefficient between the target intelligent medical electric information range and the target event cluster according to the range abnormal characteristics and the cluster characteristics;

And determining a range characteristic recognition result corresponding to the target recognition node by combining the confidence coefficient, and determining the range characteristic recognition result as a target characteristic recognition result.

In the present application, the determining, in combination with the confidence, a range feature recognition result corresponding to the target recognition node includes:

Selecting a range confidence coefficient corresponding to the target intelligent medical information range from the confidence coefficient according to the confidence coefficient, wherein the range confidence coefficient represents the confidence coefficient between the target intelligent medical information range and a undetermined event cluster corresponding to the target intelligent medical information range in the target event cluster;

Analyzing a confidence coefficient cluster corresponding to the target event cluster in the confidence coefficient by combining the confidence coefficient, wherein the confidence coefficient cluster represents the confidence coefficient between the target event cluster and a undetermined intelligent medical information range corresponding to the target event cluster in the target intelligent medical information range;

and splicing the range confidence coefficient and the confidence coefficient cluster to obtain a spliced confidence coefficient, and determining the spliced confidence coefficient as a range feature recognition result corresponding to the range recognition node.

In the present application, the selecting a target intelligent medical electric information range from the intelligent medical electric information ranges includes: classifying the range abnormal characteristics to obtain range types corresponding to each abnormal range, and determining a target abnormal range corresponding to the range type in the intelligent medical information range; the selecting the target event cluster from the event clusters comprises the following steps: classifying the cluster features to obtain cluster types corresponding to each event cluster, and determining a target event cluster from the event clusters.

In the present application, the calculating the confidence between the target intelligent medical electrical information range and the target event cluster according to the range abnormality feature and the cluster feature includes:

Integrating range abnormal characteristics corresponding to the same range type to obtain integrated range characteristics corresponding to the target intelligent medical electric information range, and splicing the integrated range characteristics and the range abnormal characteristics to obtain target range abnormal characteristics corresponding to the target intelligent medical electric information range;

Integrating cluster features corresponding to the same unit type to obtain integrated event features corresponding to the target event cluster, and splicing the integrated event features with the unit event features corresponding to the target event cluster to obtain target event features corresponding to the target event cluster;

And calculating the confidence coefficient between the target intelligent medical electric information range and the target event cluster according to the target range abnormal characteristics and the target cluster characteristics.

In the application, the target important indication features comprise global abnormal features corresponding to the intelligent medical electrical information examples, and the target event features comprise global event features corresponding to the electrical information examples; the step of performing feature recognition on the undetermined important indication feature and the undetermined event feature at the target recognition node to obtain a target feature recognition result corresponding to the target recognition node, including:

When the target identification node is a secondary identification node, secondary fault information corresponding to the intelligent medical electric information example is obtained, and characteristic extraction is carried out on the secondary fault information to obtain secondary fault characteristics, wherein the secondary identification node is an identification node among the secondary fault information, the intelligent medical electric information example and the electric abnormality example;

And carrying out feature recognition on the secondary fault feature, the global abnormal feature and the global event feature at the secondary recognition node to obtain a secondary feature recognition result corresponding to the secondary recognition node, and determining the secondary feature recognition result as the target feature recognition result.

In the present application, the feature recognition of the secondary fault feature, the global abnormal feature and the global event feature at the secondary recognition node to obtain a secondary feature recognition result corresponding to the secondary recognition node includes:

calculating abnormal matching degree between the secondary fault information and the intelligent medical electricity information example according to the secondary fault characteristics and the global abnormal characteristics;

combining the secondary fault characteristics and the global event characteristics, and calculating event matching degree between the secondary fault information and the abnormal electricity utilization event;

And generating secondary feature recognition results corresponding to the secondary recognition nodes according to the abnormal matching degree and the event matching degree.

In the application, the identification nodes comprise global identification nodes and range identification nodes; the method comprises the steps of converging the appointed intelligent medical information analysis thread according to the characteristic recognition result, the important indication characteristic and the event characteristic to obtain an intelligent medical information analysis thread, and comprises the following steps:

Respectively picking a global feature recognition result corresponding to the global recognition node and a range feature recognition result corresponding to the range recognition node from the feature recognition results;

Calculating an example performance evaluation between the intelligent medical electricity information example and the electricity consumption abnormal example according to the global feature recognition result, wherein the example performance evaluation represents a sharing performance evaluation between the intelligent medical electricity information example and the electricity consumption abnormal example;

Determining range association performance evaluation corresponding to the intelligent medical electrical information range according to the range characteristic identification result and range abnormal characteristics and cluster characteristics corresponding to the range identification node, wherein the range association performance evaluation represents association performance evaluation between the intelligent medical electrical information range and an event cluster;

Carrying out regression analysis on the intelligent medical information example according to the important indication features and the event features to obtain event regression analysis possibility, and projecting the event regression analysis possibility to obtain a regression analysis performance evaluation file corresponding to the intelligent medical information example;

And converging the appointed intelligent medical electric information analysis thread according to the example performance evaluation, the range-associated performance evaluation and the regression analysis performance evaluation file to obtain an intelligent medical electric information analysis thread.

In the present application, the determining the range association performance evaluation corresponding to the intelligent medical electrical information range according to the range feature recognition result and the range abnormal feature and the cluster feature corresponding to the range recognition node includes:

Combining the range feature recognition result to determine a range configuration catalog corresponding to the range recognition node;

Calculating the to-be-determined confidence coefficient between the intelligent medical electrical information range and the event cluster according to the range abnormal characteristics and the cluster characteristics;

And calculating target performance evaluation between the range configuration catalog and the to-be-determined confidence coefficient, and determining the target performance evaluation as range association performance evaluation corresponding to the intelligent medical electrical information range.

In the present application, the determining, in combination with the range feature recognition result, a range configuration directory corresponding to the range recognition node includes:

Selecting a reference abnormal range from the abnormal ranges, and selecting a reference event cluster from the event clusters;

Fusing the reference abnormal range and the reference event cluster to obtain at least one reference data set, wherein the reference data set comprises at least one target reference abnormal range and at least one target reference event cluster corresponding to the target reference abnormal range;

Selecting target confidence coefficient corresponding to the reference data set from the range feature recognition result, and splicing the target confidence coefficient to obtain target splicing confidence coefficient;

Determining a range configuration catalog corresponding to the range identification node by combining the target splicing confidence; calculating the confidence coefficient to be determined between the intelligent medical electrical information range and the event cluster according to the range abnormality characteristic and the cluster characteristic, wherein the confidence coefficient to be determined comprises the following steps: and extracting a real-time intelligent medical information range and a real-time event cluster from the target data set, and calculating the to-be-determined confidence coefficient between the real-time intelligent medical information range and the real-time event cluster according to the range abnormal characteristics corresponding to the real-time abnormal range and the cluster characteristics corresponding to the real-time event cluster.

In the application, the identification node further comprises a secondary identification node; the step of converging the specified intelligent medical information analysis thread according to the example performance evaluation, the range-associated performance evaluation and the regression analysis performance evaluation file to obtain an intelligent medical information analysis thread, comprising:

Selecting a secondary characteristic identification result corresponding to the secondary identification node from the characteristic identification result, wherein the secondary characteristic identification result comprises abnormal matching degree between secondary fault information corresponding to the secondary identification node and an intelligent medical electric information example and event matching degree between secondary fault information and the power utilization abnormal event;

Determining a secondary configuration catalog corresponding to the secondary identification node according to the secondary feature identification result;

Calculating secondary performance evaluation corresponding to the intelligent medical electrical information example according to the secondary configuration catalog, secondary fault characteristics corresponding to secondary fault information, the important indication characteristics and event characteristics, wherein the secondary performance evaluation represents correlation performance evaluation among the secondary fault information, the intelligent medical electrical information example and an electricity consumption abnormal example;

And converging the appointed intelligent medical electricity analysis thread by combining the secondary performance evaluation, the example performance evaluation, the range-associated performance evaluation and the regression analysis performance evaluation file to obtain an intelligent medical electricity analysis thread.

In the present application, the feature extraction is performed on the reference event result by combining the important indication local feature and the important indication global feature to obtain a key feature, including:

extracting features of the reference event result to obtain original reference event features;

And splicing the important indication local feature, the important indication global feature and the original reference event feature to obtain the key feature.

In the present application, the generating the medical electric guarantee result corresponding to the intelligent medical electric information according to the target event feature and the reference event result includes:

Performing event regression analysis on the intelligent medical electrical information according to the target event characteristics to obtain a target event;

Splicing the reference event result and the target event to obtain a spliced event, and determining the spliced event as a reference event result; returning to execute the step of extracting the features of the reference event result by combining the important indication local features and the important indication global features until the regression analysis ending requirement is reached, and obtaining a plurality of target events;

and splicing the plurality of target events to generate the medical electric guarantee result.

In a second aspect, a smart medical electrical assurance system is provided, comprising a processor and a memory in communication with each other, the processor being configured to read a computer program from the memory and execute the computer program to implement the method described above.

According to the intelligent medical electricity guarantee method and system provided by the embodiment of the application, the to-be-processed intelligent medical electricity information and the reference event result corresponding to the intelligent medical electricity information are obtained, the intelligent medical electricity information is classified, and at least one intelligent medical electricity information queue corresponding to the intelligent medical electricity information is obtained; extracting multiple aspects of characteristics from the intelligent medical electric information queue to obtain important indication local characteristics of each intelligent medical electric information and important indication global characteristics of the intelligent medical electric information; extracting features of the reference event result according to the important indication local features and the important indication global features to obtain key features, wherein the key features represent the association relationship between the intelligent medical electric information and the reference event result; splicing the important indication local feature, the important indication global feature and the key feature to obtain a target event feature of the intelligent medical electric information; and generating a medical electric guarantee result corresponding to the intelligent medical electric information based on the target event characteristics and the reference event result. The method and the system can perform multi-aspect feature extraction on the intelligent medical electric information queue of the intelligent medical electric information to obtain a large amount of information such as important indication local features and important indication global features of the intelligent medical electric information, so that the method and the system can further perform feature extraction on the reference event result corresponding to the intelligent medical electric information based on the important indication local features and the important indication global features to obtain key features, and perform feature extraction on the important indication local features, the important indication global features and the key features to obtain target event features with deeper content representation between the intelligent medical electric information and the reference event result, thereby accurately regression-analyzing the medical electric guarantee result of the intelligent medical electric information based on the target event features.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a flowchart of an intelligent medical electric guarantee method provided by an embodiment of the application.

Detailed Description

In order to better understand the above technical solutions, the following detailed description of the technical solutions of the present application is made by using the accompanying drawings and specific embodiments, and it should be understood that the specific features of the embodiments and the embodiments of the present application are detailed descriptions of the technical solutions of the present application, and not limiting the technical solutions of the present application, and the technical features of the embodiments and the embodiments of the present application may be combined with each other without conflict.

Referring to fig. 1, an intelligent medical electrical assurance method is shown, which may include the following technical solutions described in steps S101-S105.

S101, obtaining the to-be-processed intelligent medical electric information and a reference event result corresponding to the intelligent medical electric information, and classifying the intelligent medical electric information to obtain at least one intelligent medical electric information queue corresponding to the intelligent medical electric information.

Wherein, the reference event result may refer to an event for secondarily generating a medical electric assurance result of the intelligent medical electric information.

Aiming at step S101, the manner of classifying the intelligent medical electrical information to obtain at least one intelligent medical electrical information queue corresponding to the intelligent medical electrical information in step "may be: obtaining abnormal data volume of intelligent medical electric information; determining target segmentation positioning of intelligent medical electrical information based on the abnormal data volume; and dividing the intelligent medical electric information according to the target dividing and positioning to obtain at least one intelligent medical electric information queue.

The step of classifying the intelligent medical electrical information to obtain at least one intelligent medical electrical information queue corresponding to the intelligent medical electrical information may also include other modes, and specifically, the following manner of performing content segmentation on an intelligent medical electrical information example to obtain at least one intelligent medical electrical information range corresponding to the intelligent medical electrical information example may be referred to, which will not be described herein.

S102, extracting multiple aspects of characteristics from the intelligent medical electric information queue to obtain important indication local characteristics of each intelligent medical electric information and important indication global characteristics of the intelligent medical electric information.

The important indication global feature may be a feature representing the entire content of the intelligent medical electric information. By important indicating local features may be meant features that represent the content in the intelligent medical electrical information queue.

After the intelligent medical electric information queue is obtained, the intelligent medical electric information queue can be subjected to multi-aspect feature extraction, and the method for obtaining the important indication local feature of each intelligent medical electric information and the important indication global feature of the intelligent medical electric information according to the multi-aspect feature extraction of the intelligent medical electric information queue in the step S1021 to S1024 can be shown as follows:

S1021, analyzing the positioning data of each intelligent medical electric information queue in the intelligent medical electric information, and adopting an intelligent medical electric information analysis thread to perform feature extraction on the positioning data to obtain the positioning features of each intelligent medical electric information queue.

The positioning data can be related information of the positioning of the intelligent medical electric information in the intelligent medical electric information queue.

For step S1021, the manner of analyzing the positioning data of each intelligent medical electrical information queue in the intelligent medical electrical information may be: analyzing key feature points in the intelligent medical electric information queue, and inquiring target feature points matched with the key feature points in the intelligent medical electric information; and determining target positioning data of the target feature points in the intelligent medical electric information, and determining the target positioning data as positioning data of an intelligent medical electric information queue.

Or the method of analyzing the positioning data of each intelligent medical electric information queue in the intelligent medical electric information can be as follows: respectively extracting the characteristics of the intelligent medical electric information queue and the intelligent medical electric information to obtain block characteristics corresponding to the intelligent medical electric information queue and characteristics of the intelligent medical electric information; calculating a commonality coefficient between the block feature and the feature of the intelligent medical electrical information; determining target features corresponding to each block feature in the features of the intelligent medical electrical information based on the commonality coefficient; and determining the to-be-determined positioning data corresponding to the target characteristics in the intelligent medical electric information, and determining the to-be-determined positioning data as the positioning data of the intelligent medical electric information queue.

Aiming at step S1021, the manner of performing feature extraction on the positioning data by using the intelligent medical electric information analysis thread to obtain the positioning features of each intelligent medical electric information queue in the step may be as follows: and adopting a target compression unit of the intelligent medical electric information analysis thread to perform feature extraction on the positioning data to obtain the positioning features of each intelligent medical electric information queue.

For example, an original feature extraction layer of the target compression unit may be used to perform feature extraction on the positioning data, so as to obtain the positioning features of each intelligent medical electric information queue.

And S1022, performing feature extraction on the intelligent medical electric information queues by adopting an intelligent medical electric information analysis thread to obtain original important indication local features of each intelligent medical electric information queue and original important indication global features of the intelligent medical electric information.

For step S1022, the exemplary manner of "extracting features from the intelligent medical information queues by using the intelligent medical information analysis thread to obtain the original important indication local features of each intelligent medical information queue and the original important indication global features of the intelligent medical information" may be: obtaining target global indication information corresponding to intelligent medical electric information; and adopting an original feature extraction layer of a target compression unit of the intelligent medical information analysis thread to perform feature extraction on the intelligent medical information queues and the target global indication information to obtain original important indication local features of the intelligent medical information queues and original important indication global features of the target global indication information, and determining the original important indication global features of the target global indication information as original important indication global features corresponding to the target global indication information.

The target global indication information can be information indicating an intelligent medical electric information analysis thread to generate important indication global characteristics of intelligent medical electric information.

S1023, splicing the positioning features and the original important indication local features to obtain the important indication local features of each intelligent medical electric information queue.

And S1024, splicing the positioning features and the original important indication global features to obtain the important indication global features of the intelligent medical electric information.

For step S1024, the manner of "splicing the positioning feature and the original important indication global feature to obtain the important indication global feature of the intelligent medical electrical information" in step S may be: and splicing the positioning features and the original important indication global features by adopting a target compression unit to obtain the important indication global features of the intelligent medical electric information.

By way of example, the present application may employ a first target feature extraction layer of the target compression unit to splice the positioning feature and the original important indication global feature to obtain the undetermined important indication global feature; splicing the global feature to be determined, the original important indication global feature and the original important indication local feature by adopting a second target feature extraction layer to obtain a transition important indication global feature; then, a third target feature extraction layer is adopted to carry out feature extraction on the transition important indication global features, so as to obtain reference important indication global features; and splicing the reference important indication global feature and the transition important indication global feature by adopting a fourth target feature extraction layer to obtain the important indication global feature of the intelligent medical electric information.

S103, extracting features of the reference event result according to the important indication local features and the important indication global features to obtain key features.

The key features represent the association relationship between the intelligent medical electric information and the reference event result.

For step S103, the manner of extracting features of the reference event result according to the important indication local features and the important indication global features to obtain the key features in step S103 may be as shown in step S1031 to step S1032:

S1031, extracting features of the reference event result to obtain original reference event features.

The method and the device can adopt the first target decoding layer to extract the characteristics of the reference event result to obtain the original event characteristics; and splicing the original event characteristics and the reference event results by adopting a second target decoding layer to obtain the original reference event characteristics.

S1032, splicing the important indication local feature, the important indication global feature and the original reference event feature to obtain the key feature.

And S104, splicing the important indication local feature, the important indication global feature and the key feature to obtain the target event feature of the intelligent medical electric information.

For step S104, the manner of "splicing the important indication local feature, the important indication global feature and the key feature to obtain the target event feature of the intelligent medical electrical information" in step may be: extracting the key features to obtain target key features; and splicing the target key features, the important indication local features and the important indication global features to obtain target event features of the intelligent medical electric information.

The method for extracting the key features to obtain the target key features in the step' can be as follows: extracting the key features by adopting a first to-be-determined decoding layer to obtain to-be-determined key features; and splicing the undetermined key features and the key features by adopting a second undetermined decoding layer to obtain target key features.

S105, based on the target event characteristics and the reference event results, medical electric guarantee results corresponding to the intelligent medical electric information are generated.

For step S105, the manner of generating the medical electric assurance result corresponding to the intelligent medical electric information in step "based on the target event feature and the reference event result" may be: based on the target event characteristics, carrying out event regression analysis on the intelligent medical electrical information to obtain a target event; splicing the reference event result and the target event to obtain a spliced event, and determining the spliced event as a reference event result; the step of extracting the features of the reference event results according to the important indication local features and the important indication global features to obtain key features is carried out, and a plurality of target events are obtained until the regression analysis ending requirement is met; and splicing the plurality of target events to generate a medical electric guarantee result.

For example, for the step of performing an event regression analysis on the intelligent medical electrical information based on the target event feature, the manner of obtaining the target event may be: and carrying out event regression analysis on the intelligent medical electric information based on the target event characteristics by adopting an intelligent medical electric information analysis thread to obtain a target event.

For example, the manner of "stitching several target events to generate a medical electrical assurance result" for the step may be: extracting a target reference event result from a plurality of reference event results; and splicing the target reference event result and a plurality of target events to obtain a medical electric guarantee result.

In the present application, the adopted intelligent medical information analysis thread may be a thread obtained by converging a specified intelligent medical information analysis thread, and in an exemplary embodiment, before the step of "adopting the intelligent medical information analysis thread to perform feature extraction on an intelligent medical information queue", the present application may perform thread parameter convergence on the specified intelligent medical information analysis thread, and in an exemplary embodiment, steps S201 to S205 are shown:

S201, a configuration data set containing a plurality of intelligent medical electric information examples and abnormal electricity utilization examples corresponding to the intelligent medical electric information examples is obtained.

The abnormal electricity consumption examples can comprise normal abnormal electricity consumption examples and abnormal electricity consumption examples; the normal electricity use abnormality example may be an electricity use abnormality example in which the diagnosis result is normal; the abnormal electricity usage abnormality example may refer to an electricity usage abnormality example in which the diagnosis result is abnormal.

Examples of the intelligent medical information can include normal examples of the intelligent medical information and abnormal examples of the intelligent medical information; accordingly, the normal intelligent medical electricity information example can be an example corresponding to a normal electricity abnormal example; the abnormal intelligent medical electricity information example can be an example corresponding to an abnormal electricity consumption abnormal example.

In step S201, the manner of "obtaining the configuration data set including the plurality of intelligent medical electrical information examples and the electrical anomaly examples corresponding to the intelligent medical electrical information examples" may be: and obtaining a configuration data set containing a plurality of intelligent medical electric information examples and abnormal electricity utilization examples corresponding to the intelligent medical electric information examples from a local database of the electronic equipment.

S202, respectively classifying the intelligent medical electric information examples and the electric abnormality examples to obtain at least one intelligent medical electric information range corresponding to the intelligent medical electric information examples and at least one event cluster corresponding to the electric abnormality.

For step 202, content segmentation is performed on the intelligent medical electrical information example, so that there are various ways of obtaining at least one intelligent medical electrical information range corresponding to the intelligent medical electrical information example: for example, an important indicative data amount of an example of intelligent medical electrical information may be obtained; determining segmentation positioning of the intelligent medical electric information example based on the important indication data quantity; and according to the segmentation positioning, the intelligent medical electric information example is segmented to obtain at least one intelligent medical electric information range.

For example, the manner of determining the segmented localization of the smart medical electrical information example based on the amount of important indication data may be: obtaining a specified number of divisions; and splicing the important indication data quantity and the appointed segmentation number to obtain segmentation positioning of the intelligent medical electric information example.

Wherein the specified number of divisions may include a specified number of intelligent medical electrical information ranges obtained by dividing the intelligent medical electrical information example.

For step S202, the method for classifying the electricity consumption abnormality examples to obtain at least one event cluster corresponding to the electricity consumption abnormality may be: and classifying the electricity utilization abnormal examples by using the word segmentation neural network thread to obtain at least one event cluster corresponding to the electricity utilization abnormal. Or carrying out event analysis on the electricity utilization abnormal examples by using a specified dictionary to obtain an analysis result; and classifying the electricity utilization abnormal examples based on the analysis result to obtain at least one event cluster corresponding to the electricity utilization abnormal.

S203, adopting a specified intelligent medical electric information analysis thread to respectively extract characteristics of an intelligent medical electric information range and an event cluster, and obtaining important indication characteristics of at least one important indication node corresponding to an intelligent medical electric information example and event characteristics of at least one event node corresponding to an electric abnormality example.

The important indication node may refer to a node that indicates how much information the feature contains. The information amount may refer to an information amount of an example of intelligent medical electrical information. For example, the importance indicating nodes may include a global importance indicating node and a scope importance indicating node; the global important indication node may refer to a node whose important indication feature contains the information amount of the intelligent medical electric information example global; the scope importance indicating node may refer to a node of an information amount of an intelligent medical electrical information scope in which the importance indicating feature includes an example of the intelligent medical electrical information.

After the intelligent medical information range and the event cluster are obtained, the feature extraction can be performed on the intelligent medical information range and the event cluster respectively, and for example, in step 203, the manner of "adopting the designated intelligent medical information analysis thread to perform feature extraction on the intelligent medical information range and the event cluster respectively to obtain the important indication feature of at least one important indication node corresponding to the intelligent medical information example and the event feature of at least one event node corresponding to the electricity utilization abnormal example" may be as follows: performing feature extraction on the intelligent medical electrical information range by adopting an appointed intelligent medical electrical information analysis thread to obtain global abnormal features corresponding to the global important indication nodes and range abnormal features corresponding to the range important indication nodes, and determining the global abnormal features and the range abnormal features as important indication features; and extracting the characteristics of the event clusters by adopting a designated intelligent medical electrical information analysis thread to obtain global event characteristics corresponding to the global event nodes and cluster characteristics corresponding to the event cluster nodes, and determining the cluster characteristics and the important indication characteristics as important indication characteristics.

The compression unit of the appointed intelligent medical electric information analysis thread can be used for extracting the characteristics of the intelligent medical electric information range to obtain the global abnormal characteristics corresponding to the global important indication nodes and the range abnormal characteristics corresponding to the range important indication nodes; and adopting a report encoder of a designated intelligent medical electrical information analysis thread to perform feature extraction on the event cluster to obtain global event features corresponding to the global event nodes and cluster features corresponding to the event cluster nodes.

The application can also obtain the range positioning data of each intelligent medical electric information range; and carrying out feature extraction on the range positioning data by designating an original feature extraction layer to obtain range positioning features.

Then, the second feature extraction layer can be adopted to splice the abnormal features of each original range based on the reference attention weight, so as to obtain the characteristics of the undetermined abnormal range corresponding to each intelligent medical electric information range; and splicing the global abnormal feature to be determined, the original range abnormal feature and the original global abnormal feature by adopting a second feature extraction layer to obtain the transitional global abnormal feature.

Then, the method can adopt a third feature extraction layer to extract the feature of the to-be-determined abnormal range and the transition global abnormal feature to obtain the reference abnormal range feature corresponding to the to-be-determined abnormal range feature and the reference global abnormal feature corresponding to the transition global abnormal feature.

S204, carrying out feature recognition on the important indication features and the event features in at least one recognition node to obtain feature recognition results corresponding to all the recognition nodes.

Wherein the feature recognition result represents a degree of recognition between the important indication feature and the event feature. Wherein the degree of recognition can be represented by a degree of matching, a coefficient of commonality, and a degree of confidence.

The identifying node may refer to a node that identifies the important indicating feature and the event feature. The so-called identification node may comprise at least one of a global identification node, a range identification node and a secondary identification node.

The global identification node is an identification node between an intelligent medical electrical information example and an electrical abnormality example. The range identification node is an identification node between the target intelligent medical electric information range and the target event cluster. The secondary identification node is an identification node among secondary fault information, intelligent medical electricity information examples and electricity utilization abnormality examples.

After the important indication features and the event features are obtained, the important indication features and the event features can be identified, and for the step S204, the method of "identifying the important indication features and the event features in at least one identification node to obtain the feature identification results corresponding to each identification node" in the step may be as shown in steps S1 to S3:

S1, selecting target important indication nodes corresponding to all the identification nodes from the important indication nodes, and determining target event nodes corresponding to all the identification nodes from the event nodes.

For step S1, the manner of selecting the target important indication node corresponding to each identification node from the important indication nodes in step S1 may be: obtaining a projection relation set, wherein the projection relation set comprises projection relations between appointed identification nodes and appointed important indication nodes; and determining target important indication nodes corresponding to the identification nodes in the important indication nodes based on the projection relation set.

For step S1, the manner of determining the target event node corresponding to each identified node in the event nodes in step "may be: obtaining a corresponding relation set, wherein the corresponding relation set comprises a corresponding relation between a designated identification node and a designated event node; and determining target event nodes corresponding to the identification nodes in the event nodes based on the corresponding relation.

S2, selecting a target important indication feature corresponding to the target important indication node from the important indication features, and extracting a target event feature corresponding to the target event node from the event features.

Each target important indication node has a corresponding target important indication feature, for example, when the target important indication node is a global important indication node, the target important indication feature may be a global abnormal feature; when the target importance indicating node is a range importance indicating node, the target importance indicating feature may be a range abnormality feature.

For step S2, the manner of selecting the target important indication feature corresponding to the target important indication node from the important indication features in step S may be: obtaining the ordering of the important indication features; in the sorting, determining a target sorting corresponding to a target important indication node; extracting target important indication features corresponding to the target sequence from the important indication features, and determining the target important indication features corresponding to the target sequence as target important indication features corresponding to the target important indication nodes.

It can be understood that the important indication features ranked as the first bits can be global abnormal features corresponding to the global important indication nodes; the important indication features not ranked first may be range abnormality features corresponding to the range important indication nodes.

Wherein each target event node has a corresponding target event feature, e.g., when the target event node is a global event node, the target event feature may be a global event feature; when the target event node is an event cluster node, the target event feature may be a cluster feature.

For step S2, the manner of "extracting the target event feature corresponding to the target event node from the event features" in the step may refer to the manner of "selecting the target important indication feature corresponding to the target important indication node from the important indication features", which is not described herein.

And S3, identifying the target important indication features and the target event features to obtain feature identification results corresponding to all the identification nodes.

After the target important indication feature and the target event feature are obtained, the target important indication feature and the target event feature can be identified, and the method for identifying the target important indication feature and the target event feature to obtain the feature identification result corresponding to each identification node can be multiple in an exemplary manner, for example, the target important indication feature and the target event feature can be determined to be feature sets, and then the feature sets corresponding to different identification nodes are identified in parallel to obtain the feature identification result corresponding to each identification node.

For another example, the manner of identifying the target important indication feature and the target event feature to obtain the feature identification result corresponding to each identified node may be as shown in step S31 to step S34:

S31, determining at least one target identification node in the identification nodes.

For step S31, the manner of determining the target identification node from the identification nodes may be: randomly extracting a target identification node from the identification nodes; or the target identification node can be determined from the identification nodes according to the designated sequence corresponding to the identification nodes.

It may be understood herein that when there are at least two target recognition nodes determined in the recognition nodes, the undetermined feature sets corresponding to different target recognition nodes may be recognized in parallel, where the undetermined feature sets include undetermined important indication features and undetermined event features.

S32, selecting undetermined important indication features corresponding to the target identification nodes from the target important indication features, and extracting undetermined event features corresponding to the target identification nodes from the target event features.

After the target identification node is determined, the method can pick out the undetermined important indication characteristic corresponding to the target identification node and the undetermined event characteristic corresponding to the target identification node.

It can be understood herein that each identification node has a corresponding target important indication feature, the present application can obtain a target correspondence between the identification node and the important indication feature, based on which, for step S32, a manner of "selecting a pending important indication feature corresponding to the target identification node from the target important indication features" may be: and selecting undetermined important indication features corresponding to the target identification nodes from the target important indication features based on the target corresponding relation.

Similarly, the method of extracting the undetermined event feature corresponding to the target identification node from the target event feature may be: and extracting undetermined event features corresponding to the target identification node from the target event features based on undetermined corresponding relations between the identification node and the event features.

And S33, carrying out feature recognition on the undetermined important indication features and undetermined event features at the target recognition nodes to obtain target feature recognition results corresponding to the target recognition nodes.

For step S33, there may be multiple cases, for example, a case one, a case two, and a case three, in which the feature recognition is performed on the important indication feature to be determined and the event feature to be determined by the present application:

(one) case one: when the target identification node is a global identification node, the undetermined important indication features comprise global abnormal features corresponding to intelligent medical electric information examples, and the undetermined event features comprise global event features corresponding to electric abnormal examples; based on this, for step S33, "feature recognition is performed on the undetermined important indication feature and undetermined event feature at the target recognition node, and the specific implementation manner of obtaining the target feature recognition result corresponding to the target recognition node" may be: based on the global abnormal characteristics and the global event characteristics, calculating an example commonality coefficient between the intelligent medical electric information example and the electric abnormal example, wherein the global identification node is an identification node between the intelligent medical electric information example and the electric abnormal example; and splicing the example commonality coefficients to generate a global feature recognition result corresponding to the global recognition node, and determining the global feature recognition result as a target feature recognition result.

It may be appreciated herein that, after the example commonality coefficients are spliced to generate the global feature recognition result corresponding to the global recognition node, the global feature recognition result may be made to include the example commonality coefficients between each intelligent medical electrical information example and each electrical anomaly example.

(II) case two: when the target identification node is a range identification node, the undetermined important indication feature further comprises a range abnormal feature corresponding to the intelligent medical electric information range, and the undetermined event feature further comprises a cluster feature corresponding to the event cluster; based on this, for step S33, "feature recognition is performed on the undetermined important indication feature and undetermined event feature at the target recognition node, and the specific implementation manner of obtaining the target feature recognition result corresponding to the target recognition node" may be: selecting a target intelligent medical information range from the intelligent medical information range, and selecting a target event cluster from the event clusters, wherein a range identification node is an identification node between the target intelligent medical information range and the target event cluster; calculating the confidence coefficient between the target intelligent medical electric information range and the target event cluster based on the range abnormality feature and the cluster feature; and determining a range characteristic recognition result corresponding to the target recognition node according to the confidence level, and determining the range characteristic recognition result as a target characteristic recognition result.

Wherein, each intelligent medical electric information range can be determined as a target intelligent medical electric information range; or determining part of the intelligent medical electric information range as a target intelligent medical electric information range; or integrating at least two intelligent medical electric information ranges to obtain a target intelligent medical electric information range.

It can be understood herein that, since the number of the range abnormality features and the cluster features is large, the related computation of the range abnormality features and the cluster features, such as computing the range performance evaluation corresponding to the range identification node, is time-consuming under the range identification node, so that the number of the range abnormality features and the number of the cluster features can be reduced to reduce the data redundancy, thereby improving the computing efficiency.

In order to reduce the number of the range abnormal features and the number of the cluster features, the application can respectively classify the range abnormal features and the cluster features to reduce the number of the intelligent medical electric information range and the number of the event clusters, and further can reduce the number of the range abnormal features and the number of the cluster features.

Aiming at the range abnormal characteristics, the method can adopt the sub-characteristic integration layer to classify the range abnormal characteristics to obtain range types corresponding to each abnormal range, determine a target abnormal range corresponding to the range type in the intelligent medical electric information range, and determine a target abnormal range corresponding to the range type in the intelligent medical electric information range; then, adopting a sub-feature integration layer to integrate the range abnormal features corresponding to the same range type to obtain integrated range features corresponding to the target intelligent medical electric information range; then, the integrated range features and the range abnormal features are spliced by adopting a cross attention mechanism in the sub-feature splicing layer, so that the target range abnormal features corresponding to the target intelligent medical electric information range are obtained

The method for calculating the confidence between the target intelligent medical electric information range and the target event cluster based on the target range abnormal characteristic and the target cluster characteristic in the step can be as follows: performing transposition processing on the target range abnormal characteristics to obtain transposition of the target range abnormal characteristics; and calculating a target feature commonality coefficient between the transpose of the target range abnormal features and the target cluster features to obtain the confidence coefficient between the target intelligent medical electric information range and the target event cluster.

Then, the application can normalize the target characteristic commonality coefficient to obtain the confidence coefficient between the target intelligent medical information range and the target event cluster

For the second case, the method of determining the range feature recognition result corresponding to the target recognition node according to the confidence level in the step "may be: selecting range confidence corresponding to the target intelligent medical information range from the confidence degrees based on the confidence degrees, wherein the range confidence degrees represent the confidence degrees between the target intelligent medical information range and undetermined event clusters corresponding to the target intelligent medical information range in the target event clusters; according to the confidence level, a confidence coefficient cluster corresponding to the target event cluster is analyzed in the confidence coefficient, and the confidence coefficient cluster represents the confidence coefficient between the target event cluster and the undetermined intelligent medical information range corresponding to the target event cluster in the target intelligent medical information range; and splicing the range confidence coefficient and the confidence coefficient clusters to obtain a spliced confidence coefficient, and determining the spliced confidence coefficient as a range feature recognition result corresponding to the range recognition node.

It may be appreciated herein that, for a range confidence, a maximum confidence for a target event cluster for a target intelligent medical electrical range may be obtained in the confidence based on the magnitude of the confidence, and the maximum confidence may be determined as a range confidence.

Similarly, for the confidence coefficient cluster, the maximum target confidence coefficient of the target cluster for the target intelligent medical electric information range can be obtained in the confidence coefficient based on the magnitude of the confidence coefficient, and the maximum target confidence coefficient is determined to be the confidence coefficient cluster.

(III) case three: when the target identification node is a secondary identification node, the target important indication feature comprises a global abnormal feature corresponding to the intelligent medical electric information example, and the target event feature comprises a global event feature corresponding to the electric abnormal example; for step S33, the manner of performing feature recognition on the undetermined important indication feature and the undetermined event feature at the target recognition node to obtain the target feature recognition result corresponding to the target recognition node may be: secondary fault information corresponding to the intelligent medical electric information example is obtained, and feature extraction is carried out on the secondary fault information to obtain secondary fault features, wherein secondary identification nodes are identification nodes among the secondary fault information, the intelligent medical electric information example and the electricity utilization abnormal example; and carrying out feature recognition on the secondary fault feature, the global abnormal feature and the global event feature in the secondary recognition node to obtain a secondary feature recognition result corresponding to the secondary recognition node, and determining the secondary feature recognition result as a target feature recognition result.

For the third case, the manner of performing feature recognition on the secondary fault feature, the global abnormal feature and the global event feature at the secondary recognition node to obtain a secondary feature recognition result corresponding to the secondary recognition node may be: calculating abnormal matching degree between the secondary fault information and the intelligent medical electric information example based on the secondary fault characteristics and the global abnormal characteristics; calculating the event matching degree between the secondary fault information and the abnormal electricity consumption event according to the secondary fault characteristics and the global event characteristics; and generating secondary feature recognition results corresponding to the secondary recognition nodes based on the abnormal matching degree and the event matching degree.

Wherein, the abnormal matching degree can represent the correlation degree between the minor fault information and the intelligent medical electric information example; the event matching degree may represent a degree of correlation between the secondary fault information and the power consumption abnormality event.

The abnormal matching degree and the event matching degree can be determined as secondary feature recognition results corresponding to the secondary recognition nodes.

For example, the manner of calculating the abnormal matching degree between the secondary failure information and the intelligent medical electrical information example based on the secondary failure feature and the global abnormal feature may be: simplifying the global abnormal characteristics to obtain simplified abnormal characteristics; and calculating the commonality coefficient between the simplified abnormal characteristic and the minor fault characteristic to obtain the abnormal matching degree between the minor fault information and the intelligent medical electric information example.

The commonality coefficient between the simplified abnormal feature and the secondary fault feature can be determined as abnormal matching degree between the secondary fault information and the intelligent medical electric information example.

For example, the manner of "calculating the event matching degree between the secondary fault information and the abnormal electricity consumption event according to the secondary fault feature and the global event feature" may refer to the manner of "calculating the abnormal matching degree between the secondary fault information and the intelligent medical electricity consumption example based on the secondary fault feature and the global abnormal feature", which will not be described herein.

S34, returning to the step of determining the target recognition node in the recognition nodes until each recognition node is the target recognition node, and obtaining the feature recognition result corresponding to each recognition node.

S205, based on the feature recognition result, the important indication feature and the event feature, converging the appointed intelligent medical electric information analysis thread to obtain an intelligent medical electric information analysis thread, and analyzing the intelligent medical electric information by adopting the intelligent medical electric information analysis thread to obtain a medical electric guarantee result.

After the feature recognition result, the important indication feature and the event feature are obtained, the method can converge the appointed intelligent medical information analysis thread, and the method can be shown in the steps S2051 to S2055, wherein the appointed intelligent medical information analysis thread is converged based on the feature recognition result, the important indication feature and the event feature in the step S205, and the method for obtaining the intelligent medical information analysis thread is exemplified by the global recognition node and the range recognition node:

S2051, respectively picking out a global feature recognition result corresponding to the global recognition node and a range feature recognition result corresponding to the range recognition node from the feature recognition results.

S2052, calculating an example performance evaluation between the intelligent medical electric information example and the electric abnormality example based on the global feature recognition result.

Wherein the example performance evaluation represents a shared performance evaluation between the smart medical electrical information example and the electrical anomaly example.

For S2052, since the global feature recognition result includes an example common coefficient between each intelligent medical electrical information example and each electrical anomaly example, based on this, the manner of step "calculating an example performance evaluation between an intelligent medical electrical information example and an electrical anomaly example based on the global feature recognition result" may be: selecting a target example commonality coefficient from example commonality coefficients included in the global feature recognition result; and projecting the example commonality coefficient and the target example commonality coefficient to obtain an example performance evaluation between the intelligent medical electricity information example and the electricity utilization abnormality example.

The target example commonality coefficients comprise first target example commonality coefficients corresponding to intelligent medical electric information examples and second target example commonality coefficients corresponding to abnormal electric information examples; the example commonality coefficients include a first example commonality coefficient for the smart medical electrical information example for the electrical anomaly example and a second example commonality coefficient for the electrical anomaly example for the smart medical electrical information example. Based on this, the step of "projecting the example commonality coefficient and the target example commonality coefficient with the example performance evaluation function, to obtain the example performance evaluation between the intelligent medical electrical information example and the electrical anomaly example" may be: splicing the first target example commonality coefficient and the first example commonality coefficient to obtain a first spliced commonality coefficient; splicing the second target example common coefficient and the second example common coefficient to obtain a second spliced common coefficient; and obtaining performance evaluation reference parameters, and projecting the first splicing commonality coefficient, the second splicing commonality coefficient and the performance evaluation reference parameters by adopting an example performance evaluation function to obtain an example performance evaluation between the intelligent medical electricity information example and the electricity consumption abnormality example.

S2053, determining range association performance evaluation corresponding to the intelligent medical electric information range based on the range characteristic recognition result and the range abnormal characteristic and the cluster characteristic corresponding to the range recognition node.

The range association performance evaluation represents the association performance evaluation between the intelligent medical electric information range and the event cluster.

For step S2053, the manner of determining the range-associated performance evaluation corresponding to the intelligent medical electrical information range in step "based on the range feature recognition result and the range abnormality feature and the cluster feature corresponding to the range recognition node" may be as shown in steps S531 to S533:

s531, determining a range configuration catalog corresponding to the range identification node according to the range characteristic identification result.

The range feature recognition result includes a confidence level between a target intelligent medical information range of the intelligent medical information example and a target event cluster of the electricity utilization anomaly example, based on which, the method of determining a range configuration directory corresponding to the range recognition node according to the range feature recognition result may be: selecting a reference abnormal range from the abnormal ranges, and selecting a reference event cluster from the event clusters; fusing the reference abnormal range and the reference event cluster to obtain at least one reference data set, wherein the reference data set comprises at least one target reference abnormal range and at least one target reference event cluster corresponding to the target reference abnormal range; selecting target confidence coefficient corresponding to the reference data set from the range feature recognition result, and splicing the target confidence coefficient to obtain target splicing confidence coefficient; and determining a range configuration catalog corresponding to the range identification node according to the target splicing confidence.

Wherein, a plurality of reference abnormal ranges can be randomly selected from the abnormal ranges; several reference event clusters can be randomly selected from the event clusters.

Wherein each reference abnormality range may be determined as a target reference abnormality range; or a partial reference abnormality range may be determined as the target reference abnormality range.

Wherein each reference event cluster may be determined as a target reference event cluster; or a partial reference event cluster may be determined as the target reference event cluster.

For example, the step of "picking out the target confidence corresponding to the reference data set from the range feature recognition result" may be: extracting a first confidence coefficient corresponding to each target reference abnormal range in the reference data set from the range feature recognition result; selecting a second confidence coefficient corresponding to each target reference event cluster in the reference data set from the range feature recognition result; and determining the first confidence coefficient and the second confidence coefficient as target confidence coefficients corresponding to the reference data set.

For example, the manner of "stitching the target confidence level to obtain the target stitching confidence level" in the step may be: and adding the target confidence degrees corresponding to the reference data sets to obtain the target splicing confidence degrees.

For example, the step of determining the range configuration directory corresponding to the range identification node according to the target splicing confidence may be: performing projection processing on the spliced confidence coefficient of each target to obtain a projected confidence coefficient; normalizing the projected confidence coefficient to obtain a range configuration catalog corresponding to the range identification node.

S532, calculating the to-be-determined confidence between the intelligent medical electric information range and the event cluster based on the range abnormality feature and the cluster feature.

For step S532, based on the foregoing data set, the manner of calculating the confidence to be determined between the intelligent medical electrical information range and the event cluster in step "based on the range abnormality feature and the cluster feature" may be: and extracting a real-time intelligent medical electric information range and a real-time event cluster from the target data set, and calculating the to-be-determined confidence between the real-time intelligent medical electric information range and the real-time event cluster based on the range abnormality characteristics corresponding to the real-time abnormality range and the cluster characteristics corresponding to the real-time event cluster.

The to-be-determined confidence coefficient between the real-time intelligent medical electric information range and the real-time event cluster can be calculated by adopting a commonality coefficient function based on the range abnormal characteristics corresponding to the real-time abnormal range and the cluster characteristics corresponding to the real-time event cluster.

Exemplary, the range anomaly characteristic corresponding to the real-time anomaly range is transposed, and the target transposition corresponding to the range anomaly characteristic is obtained; calculating undetermined commonality coefficients between the target transpose and cluster features corresponding to the real-time event clusters by adopting a commonality coefficient function; normalizing the undetermined commonality coefficient to obtain undetermined confidence coefficient between the real-time intelligent medical electric information range and the real-time event cluster.

S533, calculating target performance evaluation between the range configuration catalog and the to-be-determined confidence coefficient, and determining the target performance evaluation as range association performance evaluation corresponding to the intelligent medical electric information range.

It will be appreciated herein that the scope-associated performance evaluation may bring the confidence to be determined into close proximity to the scope configuration catalog to strengthen the relationship between event clusters and intelligent medical electrical scope.

S2054, carrying out regression analysis on the intelligent medical electrical information example based on the important indication features and the event features to obtain event regression analysis possibility, and projecting the event regression analysis possibility to obtain a regression analysis performance evaluation file corresponding to the intelligent medical electrical information example.

For step S2054, a regression analysis performance evaluation file function may be used to project the probability of regression analysis of the event, so as to obtain a regression analysis performance evaluation file corresponding to the intelligent medical electrical information example.

S2055, converging the appointed intelligent medical electric information analysis thread based on the example performance evaluation, the range correlation performance evaluation and the regression analysis performance evaluation file to obtain the intelligent medical electric information analysis thread.

The identifying node further includes a secondary identifying node, based on which, for step S2055, the manner of step "converging the specified intelligent medical information analysis thread based on the example performance evaluation, the range-associated performance evaluation, and the regression analysis performance evaluation file to obtain the intelligent medical information analysis thread" may be as shown in steps S51 to S54:

s51, selecting secondary characteristic recognition results corresponding to the secondary recognition nodes from the characteristic recognition results.

The secondary feature recognition result comprises abnormal matching degree between secondary fault information corresponding to the secondary recognition node and the intelligent medical electricity information example, and event matching degree between the secondary fault information and the electricity utilization abnormal event.

S52, determining a secondary configuration catalog corresponding to the secondary identification node based on the secondary feature identification result.

For step S52, the manner of determining the secondary configuration directory corresponding to the secondary identification node in step "based on the secondary feature identification result" may be: obtaining secondary features corresponding to the secondary fault information; selecting reference secondary fault information from the secondary fault information; selecting a reference intelligent medical electric information example from the intelligent medical electric information examples; selecting a reference electricity consumption abnormal example from the electricity consumption abnormal examples; fusing the reference secondary fault information and the reference intelligent medical electrical information examples to obtain at least one reference secondary data set, wherein the reference secondary data set comprises at least one target reference secondary fault information and at least one target reference intelligent medical electrical information example corresponding to the target reference secondary fault information; fusing the reference secondary fault information and the reference electricity consumption abnormal examples to obtain at least one undetermined secondary data set, wherein the undetermined secondary data set comprises at least one undetermined reference secondary fault information and at least one reference electricity consumption abnormal example corresponding to the undetermined reference secondary fault information; selecting a target abnormal matching degree corresponding to the reference secondary data set from the secondary characteristic identification result, and splicing the target abnormal matching degree to obtain a spliced abnormal matching degree; determining a secondary abnormal catalog corresponding to the secondary identification node according to the spliced abnormal matching degree; selecting a target event matching degree corresponding to the undetermined secondary data set from the secondary characteristic identification result, and splicing the target event matching degree to obtain a spliced event matching degree; determining a secondary event catalog corresponding to the secondary identification node according to the matching degree of the spliced events; the secondary exception directory and the secondary event directory are determined to be secondary configuration directories.

The step of selecting the reference secondary fault information from the secondary fault information, selecting the reference intelligent medical electrical information example from the intelligent medical electrical information examples, and selecting the reference electrical anomaly example from the electrical anomaly examples may be specifically referred to the step of selecting the reference anomaly range from the anomaly range described above, and will not be repeated here.

The step of determining the secondary abnormal directory corresponding to the secondary identification node according to the matching degree of the splicing anomaly and the step of determining the secondary event directory corresponding to the secondary identification node according to the matching degree of the splicing event are specifically referred to the step of determining the range configuration directory corresponding to the range identification node according to the target splicing confidence, and are not described herein.

And S53, calculating secondary performance evaluation corresponding to the intelligent medical electric information example based on the secondary configuration catalog, the secondary fault characteristics corresponding to the secondary fault information, the important indication characteristics and the event characteristics.

Wherein the secondary performance evaluation represents a correlation performance evaluation among the secondary failure information, the smart medical electrical information example, and the electrical anomaly example.

For step S53, the important indication features include global exception features, the event features include global event features, and the secondary configuration directory includes a secondary exception directory and a secondary event directory; the step of calculating the secondary performance evaluation corresponding to the intelligent medical electrical information example based on the secondary configuration catalog, the secondary fault feature, the important indication feature and the event feature corresponding to the secondary fault information may be as follows: simplifying the global abnormal characteristics to obtain simplified abnormal characteristics; simplifying the global event features to obtain simplified event features; calculating a commonality coefficient between the simplified abnormal characteristic and the minor fault characteristic to obtain the to-be-determined abnormal matching degree between the minor fault information and the intelligent medical electric information example; calculating a commonality coefficient between the simplified event feature and the secondary fault feature to obtain the matching degree of the to-be-determined event between the secondary fault information and the electricity consumption abnormal example; and calculating secondary performance evaluation corresponding to the intelligent medical electric information example based on the secondary configuration catalog, the pending event matching degree and the pending abnormal matching degree.

The method of calculating the secondary performance evaluation corresponding to the intelligent medical electric information example based on the secondary configuration catalog, the predetermined event matching degree and the predetermined abnormal matching degree may be as follows: calculating a first coefficient of commonality between the secondary anomaly inventory and the coefficient of commonality of the pending event; calculating a second common coefficient between the secondary event catalogue and the predetermined abnormal matching degree; and splicing the first common coefficient and the second common coefficient by adopting a secondary performance evaluation function to obtain secondary performance evaluation corresponding to the intelligent medical electrical information example.

S54, converging the appointed intelligent medical electric information analysis thread according to the secondary performance evaluation, the example performance evaluation, the range correlation performance evaluation and the regression analysis performance evaluation file to obtain the intelligent medical electric information analysis thread.

Aiming at the step S54, a performance evaluation splicing function can be adopted to splice the secondary performance evaluation, the example performance evaluation, the range correlation performance evaluation and the regression analysis performance evaluation files to obtain splicing performance evaluation; and based on the splicing performance evaluation, converging the appointed intelligent medical electric information analysis thread to obtain the intelligent medical electric information analysis thread.

By way of example, the present application may obtain target performance evaluation parameters; splicing the example performance evaluation, the secondary performance evaluation and the range-associated performance evaluation based on the target performance evaluation parameter to obtain an original splicing performance evaluation; and adding the original splicing performance evaluation and regression analysis performance evaluation files to obtain the splicing performance evaluation.

It can be understood here that, for steps S201 to S205, the embodiment of the present application may identify the abnormal feature of the intelligent medical electrical information example and the event feature of the electrical abnormality example, so as to obtain the feature identification result of at least one identification node, so that the feature identification result, the important indication feature and the event feature of at least one identification node may be used to configure and designate the intelligent medical electrical information analysis thread, so that the intelligent medical electrical information analysis thread may learn the identification capability of any intelligent medical electrical information, and thus the accuracy of the intelligent medical electrical information analysis thread to generate the medical electrical guarantee result corresponding to the intelligent medical electrical information may be improved.

The application can obtain the intelligent medical electric information to be processed and the reference event result corresponding to the intelligent medical electric information, and classify the intelligent medical electric information to obtain at least one intelligent medical electric information queue corresponding to the intelligent medical electric information; extracting multiple aspects of characteristics from the intelligent medical electric information queue to obtain important indication local characteristics of each intelligent medical electric information and important indication global characteristics of the intelligent medical electric information; extracting features of the reference event result according to the important indication local features and the important indication global features to obtain key features, wherein the key features represent the association relationship between the intelligent medical electric information and the reference event result; splicing the important indication local feature, the important indication global feature and the key feature to obtain a target event feature of the intelligent medical electric information; and generating a medical electric guarantee result corresponding to the intelligent medical electric information based on the target event characteristics and the reference event result. The method and the system can perform multi-aspect feature extraction on the intelligent medical electric information queue of the intelligent medical electric information to obtain a large amount of information such as important indication local features and important indication global features of the intelligent medical electric information, so that the method and the system can further perform feature extraction on the reference event result corresponding to the intelligent medical electric information based on the important indication local features and the important indication global features to obtain key features, and perform feature extraction on the important indication local features, the important indication global features and the key features to obtain target event features with deeper content representation between the intelligent medical electric information and the reference event result, thereby accurately regression-analyzing the medical electric guarantee result of the intelligent medical electric information based on the target event features.

The specific flow of the intelligent medical electricity guaranteeing method is as follows from step S501 to step S508:

S501, the electronic equipment obtains a configuration data set comprising a plurality of intelligent medical electric information examples and abnormal electricity utilization examples corresponding to the intelligent medical electric information examples.

For step S501, for example, the electronic device sends an example obtaining request to the storage server through the internal network, so that the storage server extracts an intelligent medical electrical information example and an intelligent medical electrical information example from a storage space corresponding to the storage server, and returns the intelligent medical electrical information example and an electrical anomaly example to the electronic device; the electronic equipment receives the intelligent medical electric information examples and the electricity consumption abnormal examples returned by the storage server and generates a configuration data set based on the intelligent medical electric information examples and the electricity consumption abnormal examples.

S502, the electronic equipment classifies the intelligent medical electric information examples and the electric abnormality examples respectively to obtain at least one intelligent medical electric information range corresponding to the intelligent medical electric information examples and at least one event cluster corresponding to the electric abnormality.

For step S502, the content segmentation is performed on the intelligent medical electrical information example, and the manner of obtaining at least one intelligent medical electrical information range corresponding to the intelligent medical electrical information example may be: obtaining important indication data quantity of an intelligent medical electric information example; determining segmentation positioning of the intelligent medical electric information example based on the important indication data quantity; and according to the segmentation positioning, the intelligent medical electric information example is segmented to obtain at least one intelligent medical electric information range.

The manner of determining the segmentation positioning of the intelligent medical electric information example based on the important indication data amount can be as follows: obtaining a specified number of divisions; and splicing the important indication data quantity and the appointed segmentation number to obtain segmentation positioning of the intelligent medical electric information example. Wherein the specified number of divisions may include a specified number of intelligent medical electrical information ranges obtained by dividing the intelligent medical electrical information example.

Aiming at step S502, the method for classifying the electricity consumption abnormality examples to obtain at least one event cluster corresponding to the electricity consumption abnormality may be: carrying out event analysis on the electricity utilization abnormal examples by using a specified dictionary to obtain an analysis result; and classifying the electricity utilization abnormal examples based on the analysis result to obtain at least one event cluster corresponding to the electricity utilization abnormal.

S503, the electronic equipment adopts a specified intelligent medical electric information analysis thread to respectively conduct feature extraction on the intelligent medical electric information range and the event cluster, so that the important indication feature of at least one important indication node corresponding to the intelligent medical electric information example and the event feature of at least one event node corresponding to the electric abnormality example are obtained.

For step S503, the manner of "adopting the specified intelligent medical electrical information analysis thread to perform feature extraction on the intelligent medical electrical information range and the event cluster respectively to obtain the important indication feature of not less than one important indication node corresponding to the intelligent medical electrical information example and the event feature of not less than one event node corresponding to the electrical anomaly example" may be: the electronic equipment adopts a specified intelligent medical electric information analysis thread to perform feature extraction on the intelligent medical electric information range to obtain global abnormal features corresponding to the global important indication nodes and range abnormal features corresponding to the range important indication nodes, and the global abnormal features and the range abnormal features are determined to be important indication features; and extracting the characteristics of the event clusters by adopting a designated intelligent medical electrical information analysis thread to obtain global event characteristics corresponding to the global event nodes and cluster characteristics corresponding to the event cluster nodes, and determining the cluster characteristics and the important indication characteristics as important indication characteristics.

S504, the electronic equipment picks out target important indication nodes corresponding to the identification nodes from the important indication nodes, and determines target event nodes corresponding to the identification nodes from the event nodes.

S505, the electronic equipment selects the target important indication feature corresponding to the target important indication node from the important indication features, and extracts the target event feature corresponding to the target event node from the event features.

S506, the electronic equipment identifies the target important indication features and the target event features to obtain feature identification results corresponding to the identification nodes.

Wherein the feature recognition result represents a degree of recognition between the important indication feature and the event feature.

For example, when the target identification node is a global identification node, the undetermined important indication feature comprises a global abnormal feature corresponding to the intelligent medical electrical information example, and the undetermined event feature comprises a global event feature corresponding to the electrical abnormal example. On the basis, the electronic equipment can calculate an example common coefficient between the intelligent medical electric information example and the electricity consumption abnormal example based on the global abnormal characteristic and the global event characteristic, and the global identification node is an identification node between the intelligent medical electric information example and the electricity consumption abnormal example; and splicing the example commonality coefficients to generate a global feature recognition result corresponding to the global recognition node, and determining the global feature recognition result as a target feature recognition result.

For example, when the target identification node is a range identification node, the undetermined important indication feature further comprises a range abnormal feature corresponding to the intelligent medical electric information range, and the undetermined event feature further comprises a cluster feature corresponding to the event cluster; on the basis, the electronic equipment can select a target intelligent medical information range from the intelligent medical information range, select a target event cluster from the event clusters, and the range identification node is an identification node between the target intelligent medical information range and the target event cluster; calculating the confidence coefficient between the target intelligent medical electric information range and the target event cluster based on the range abnormality feature and the cluster feature; and determining a range characteristic recognition result corresponding to the target recognition node according to the confidence level, and determining the range characteristic recognition result as a target characteristic recognition result.

For another example, when the target identification node is a secondary identification node, the target important indication feature includes a global abnormal feature corresponding to the intelligent medical electrical information example, and the target event feature includes a global event feature corresponding to the electrical abnormality example. Based on the secondary fault information, the electronic equipment can acquire secondary fault information corresponding to the intelligent medical electric information example, and perform feature extraction on the secondary fault information to acquire secondary fault features, wherein the secondary identification node is an identification node between the secondary fault information, the intelligent medical electric information example and the electric abnormality example; and carrying out feature recognition on the secondary fault feature, the global abnormal feature and the global event feature in the secondary recognition node to obtain a secondary feature recognition result corresponding to the secondary recognition node, and determining the secondary feature recognition result as a target feature recognition result.

S507, the electronic equipment converges the appointed intelligent medical information analysis thread based on the feature recognition result, the important indication feature and the event feature to obtain the intelligent medical information analysis thread.

For S507, for example, the electronic device may select, from the feature recognition results, a global feature recognition result corresponding to the global recognition node and a range feature recognition result corresponding to the range recognition node; calculating an example performance evaluation between the intelligent medical electrical information example and the electricity consumption abnormality example based on the global feature recognition result, wherein the example performance evaluation represents a sharing performance evaluation between the intelligent medical electrical information example and the electricity consumption abnormality example; determining range association performance evaluation corresponding to the intelligent medical electric information range based on the range characteristic recognition result and the range abnormal characteristics and the cluster characteristics corresponding to the range recognition node, wherein the range association performance evaluation represents association performance evaluation between the intelligent medical electric information range and the event cluster; carrying out regression analysis on the intelligent medical electrical information example based on the important indication features and the event features to obtain event regression analysis possibility, and projecting the event regression analysis possibility to obtain a regression analysis performance evaluation file corresponding to the intelligent medical electrical information example; based on the example performance evaluation, the range correlation performance evaluation and the regression analysis performance evaluation file, the appointed intelligent medical information analysis thread is converged, and the intelligent medical information analysis thread is obtained.

S508, the electronic equipment analyzes the intelligent medical electric information by adopting an intelligent medical electric information analysis thread to obtain a medical electric guarantee result.

Aiming at step S508, for example, the electronic device may obtain the to-be-processed intelligent medical electrical information and a reference event result corresponding to the intelligent medical electrical information, and classify the intelligent medical electrical information to obtain at least one intelligent medical electrical information queue corresponding to the intelligent medical electrical information; the intelligent medical electric information analysis thread is adopted to conduct multi-aspect feature extraction on the intelligent medical electric information queue, and important indication local features of each intelligent medical electric information and important indication global features of the intelligent medical electric information are obtained; extracting features of the reference event result according to the important indication local features and the important indication global features to obtain key features, wherein the key features represent the association relationship between the intelligent medical electric information and the reference event result; splicing the important indication local feature, the important indication global feature and the key feature to obtain a target event feature of the intelligent medical electric information; and generating a medical electric guarantee result corresponding to the intelligent medical electric information based on the target event characteristics and the reference event result.

On the basis, an intelligent medical electric guarantee device is provided, and the device comprises:

the queue obtaining module is used for obtaining the intelligent medical electric information to be processed and a reference event result corresponding to the intelligent medical electric information, classifying the intelligent medical electric information and obtaining at least one intelligent medical electric information queue corresponding to the intelligent medical electric information;

The global feature extraction module is used for extracting multiple aspects of features of the intelligent medical electric information queue to obtain important indication local features of each intelligent medical electric information and important indication global features of the intelligent medical electric information;

The key feature obtaining module is used for carrying out feature extraction on the reference event result by combining the important indication local feature and the important indication global feature to obtain key features, wherein the key features represent the association relationship between the intelligent medical electric information and the reference event result;

The event feature splicing module is used for splicing the important indication local features, the important indication global features and the key features to obtain target event features of the intelligent medical electric information;

And the result generation module is used for generating a medical electric guarantee result corresponding to the intelligent medical electric information according to the target event characteristics and the reference event result.

On the above basis, an intelligent medical electrical assurance system is shown comprising a processor and a memory in communication with each other, the processor being adapted to read a computer program from the memory and execute it to implement the method described above.

On the basis of the above, there is also provided a computer readable storage medium on which a computer program stored which, when run, implements the above method.

In summary, based on the above scheme, obtaining the to-be-processed intelligent medical electrical information and a reference event result corresponding to the intelligent medical electrical information, and classifying the intelligent medical electrical information to obtain at least one intelligent medical electrical information queue corresponding to the intelligent medical electrical information; extracting multiple aspects of characteristics from the intelligent medical electric information queue to obtain important indication local characteristics of each intelligent medical electric information and important indication global characteristics of the intelligent medical electric information; extracting features of the reference event result according to the important indication local features and the important indication global features to obtain key features, wherein the key features represent the association relationship between the intelligent medical electric information and the reference event result; splicing the important indication local feature, the important indication global feature and the key feature to obtain a target event feature of the intelligent medical electric information; and generating a medical electric guarantee result corresponding to the intelligent medical electric information based on the target event characteristics and the reference event result. The method and the system can perform multi-aspect feature extraction on the intelligent medical electric information queue of the intelligent medical electric information to obtain a large amount of information such as important indication local features and important indication global features of the intelligent medical electric information, so that the method and the system can further perform feature extraction on the reference event result corresponding to the intelligent medical electric information based on the important indication local features and the important indication global features to obtain key features, and perform feature extraction on the important indication local features, the important indication global features and the key features to obtain target event features with deeper content representation between the intelligent medical electric information and the reference event result, thereby accurately regression-analyzing the medical electric guarantee result of the intelligent medical electric information based on the target event features.

It should be appreciated that the systems and modules thereof shown above may be implemented in a variety of ways. For example, in some embodiments, the system and its modules may be implemented in hardware, software, or a combination of software and hardware. Wherein the hardware portion may be implemented using dedicated logic; the software portions may then be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or special purpose design hardware. Those skilled in the art will appreciate that the methods and systems described above may be implemented using computer executable instructions and/or embodied in processor control code, such as provided on a carrier medium such as a magnetic disk, CD or DVD-ROM, a programmable memory such as read only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The system of the present application and its modules may be implemented not only with hardware circuitry such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but also with software executed by various types of processors, for example, and with a combination of the above hardware circuitry and software (e.g., firmware).

It should be noted that, the advantages that may be generated by different embodiments may be different, and in different embodiments, the advantages that may be generated may be any one or a combination of several of the above, or any other possible advantages that may be obtained.

Claims

1. An intelligent medical electrical assurance method, comprising:

generating a medical electric guarantee result corresponding to the intelligent medical electric information according to the target event characteristics and the reference event result;

The method for extracting the multi-aspect characteristics of the intelligent medical electric information queue to obtain important indication local characteristics of each intelligent medical electric information and important indication global characteristics of the intelligent medical electric information comprises the following steps:

2. The intelligent medical electrical assurance method of claim 1, wherein prior to the feature extraction of the intelligent medical electrical information queue with the intelligent medical electrical information analysis thread, the method further comprises:

3. The intelligent medical electrical assurance method according to claim 2, wherein the feature recognition of the important indication feature and the event feature at least one recognition node to obtain feature recognition results corresponding to the respective recognition nodes includes:

4. The intelligent medical electrical assurance method of claim 3, wherein identifying the target important indication feature and the target event feature to obtain feature identification results corresponding to each identification node includes:

5. The intelligent medical electrical assurance method of claim 4, wherein the pending significant indication features include global anomaly features corresponding to the intelligent medical electrical examples, and the pending event features include global event features corresponding to electrical anomaly examples; the step of performing feature recognition on the undetermined important indication feature and the undetermined event feature at the target recognition node to obtain a target feature recognition result corresponding to the target recognition node, including:

6. The intelligent medical electrical assurance method of claim 4, wherein the pending significant indication features further comprise range anomaly features corresponding to the intelligent medical electrical information range, the pending event features further comprising cluster features corresponding to event clusters; the step of performing feature recognition on the undetermined important indication feature and the undetermined event feature at the target recognition node to obtain a target feature recognition result corresponding to the target recognition node, including:

7. The intelligent medical electrical assurance method of claim 6, wherein the determining, in combination with the confidence level, a range feature recognition result corresponding to the target recognition node includes:

8. The intelligent medical electrical assurance method of claim 6, wherein selecting a target intelligent medical electrical scope from the intelligent medical electrical scope comprises: classifying the range abnormal characteristics to obtain range types corresponding to each abnormal range, and determining a target abnormal range corresponding to the range type in the intelligent medical information range; the selecting the target event cluster from the event clusters comprises the following steps: classifying the cluster features to obtain cluster types corresponding to each event cluster, and determining a target event cluster from the event clusters.

9. The intelligent medical electrical assurance method of claim 8, wherein the calculating confidence between the target intelligent medical electrical information range and the target event cluster based on the range anomaly feature and the cluster feature comprises:

10. The intelligent medical electrical assurance method of claim 4, wherein the target important indication features include global anomaly features corresponding to the intelligent medical electrical information examples, the target event features include global event features corresponding to electrical anomaly examples; the step of performing feature recognition on the undetermined important indication feature and the undetermined event feature at the target recognition node to obtain a target feature recognition result corresponding to the target recognition node, including:

11. The intelligent medical electrical assurance method of claim 10, wherein performing feature recognition on the secondary identification node on the secondary fault feature, the global anomaly feature and the global event feature to obtain a secondary feature recognition result corresponding to the secondary identification node comprises:

12. The intelligent medical electrical assurance method of claim 2, wherein the identification nodes include global identification nodes and range identification nodes; the method comprises the steps of converging the appointed intelligent medical information analysis thread according to the characteristic recognition result, the important indication characteristic and the event characteristic to obtain an intelligent medical information analysis thread, and comprises the following steps:

13. The intelligent medical electrical assurance method of claim 12, wherein determining the range associated performance rating corresponding to the intelligent medical electrical information range based on the range feature recognition result and the range anomaly feature and the cluster feature corresponding to the range recognition node comprises:

14. The intelligent medical electrical assurance method of claim 13, wherein the determining a range configuration catalog corresponding to the range identification node in combination with the range feature identification result comprises:

Determining a range configuration catalog corresponding to the range identification node by combining the target splicing confidence; calculating the confidence coefficient to be determined between the intelligent medical electrical information range and the event cluster according to the range abnormality characteristic and the cluster characteristic, wherein the confidence coefficient to be determined comprises the following steps: extracting a real-time intelligent medical electric information range and a real-time event cluster from a target data set, and calculating the to-be-determined confidence coefficient between the real-time intelligent medical electric information range and the real-time event cluster according to the range abnormal characteristics corresponding to the real-time abnormal range and the cluster characteristics corresponding to the real-time event cluster.

15. The intelligent medical electrical assurance method of claim 12, wherein the identification nodes further comprise secondary identification nodes; the step of converging the specified intelligent medical information analysis thread according to the example performance evaluation, the range-associated performance evaluation and the regression analysis performance evaluation file to obtain an intelligent medical information analysis thread, comprising:

16. The intelligent medical electrical assurance method of claim 1, wherein the combining the important indication local features and the important indication global features, performing feature extraction on the reference event results to obtain key features, includes:

17. The intelligent medical electrical assurance method of claim 1, wherein the generating the medical electrical assurance results corresponding to the intelligent medical electrical information according to the target event features and the reference event results comprises:

18. A smart medical electrical assurance system comprising a processor and a memory in communication with each other, the processor being configured to read a computer program from the memory and execute the computer program to implement the method of any one of claims 1-17.