An emergency monitoring management system based on artificial intelligence

Abstract

The present invention discloses an emergency monitoring management system based on artificial intelligence, including a data acquisition module, a data preprocessing module, a dynamic clustering processing module and an emergency monitoring management module. The present invention belongs to the field of medical information processing technology, and specifically refers to an emergency monitoring management system based on artificial intelligence. This solution uses multi-scale feature processing and Gaussian fuzzy membership function to calculate the local disease density of patients, and dynamically adjusts the clustering center and noise parameters, adaptively adjusts the noise processing, and improves the accuracy of patient classification; introduces dynamic parameter optimization and iterative optimization mechanisms to efficiently cluster patients; introduces local abnormal factors to ensure that the impact of noise data on the overall system is minimized; uses high-density area recognition to determine patients in disease-intensive areas, dynamically clusters based on disease density and distance, and quickly classifies patients through historical data labels; ensures that each patient receives reasonable disease classification and treatment.

Description

Technical Field

[0001] The present invention relates to the technical field of medical information processing, and in particular to an emergency monitoring management system based on artificial intelligence. Background Art

[0002] The emergency monitoring management system is a medical information system used to monitor, manage and support patient treatment in real time in the emergency department environment. Its core goal is to collect and process patients' vital signs and condition data in real time, and provide functions such as condition classification, emergency warning, diagnosis and treatment recommendations and medical resource scheduling based on intelligent algorithms and medical rules, so as to improve the efficiency and quality of the emergency department and ensure that critically ill patients can receive timely and effective treatment. However, the general emergency monitoring management system cannot fully consider the diversity of patients' conditions, resulting in the omission of key patients, and the insufficient processing capacity of noisy patient data, which leads to misjudgment and missed judgment; the general emergency monitoring management system has the problem of improper processing of fuzzy patient data, resulting in unclear classification or misdiagnosis, and it is difficult to flexibly respond to the complexity of different conditions, resulting in insufficient classification accuracy. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an emergency monitoring management system based on artificial intelligence. In view of the fact that the general emergency monitoring management system cannot fully consider the diversity of patients' conditions, resulting in the omission of key patients, and the insufficient processing ability of noisy patient data, which leads to misjudgment and missed judgment, this solution uses multi-scale feature processing and Gaussian fuzzy membership function to calculate the local disease density of patients, and dynamically adjusts the clustering center and noise parameters, adaptively adjusts the noise processing, and improves the patient classification accuracy; introduces dynamic parameter optimization and iterative optimization mechanisms to efficiently cluster patients; introduces local abnormal factors to ensure that the impact of noise data on the overall system is minimized; thereby ensuring the efficiency and accuracy of the emergency monitoring management system; in view of the fact that the general emergency monitoring management system has improper processing of fuzzy patient data, resulting in unclear classification or misdiagnosis, and it is difficult to flexibly respond to the complexity of different diseases, resulting in insufficient classification accuracy, this solution uses high-density area recognition to determine patients in disease-intensive areas, dynamically clusters based on disease density and distance, and quickly classifies patients through historical data labels; ensuring that each patient receives reasonable disease classification and treatment.

[0004] The technical solution adopted by the present invention is as follows: An artificial intelligence-based emergency monitoring management system provided by the present invention includes a data acquisition module, a data preprocessing module, a dynamic clustering processing module and an emergency monitoring management module;

[0005] The data acquisition module acquires historical emergency monitoring patient data sets and real-time emergency monitoring patient data sets;

[0006] The data preprocessing module performs data cleaning, data conversion and standardization on the collected data;

[0007] The dynamic clustering processing module uses multi-scale feature processing and Gaussian fuzzy membership function to calculate the local disease density of patients, and dynamically adjusts the cluster center and noise parameters; introduces dynamic parameter optimization and iterative optimization mechanism and local abnormal factors, and finally realizes dynamic clustering processing;

[0008] The emergency monitoring management module uses high-density area recognition to determine patients in disease-intensive areas, and quickly classifies patients through historical data tags, thereby realizing emergency monitoring management.

[0009] Furthermore, in the data acquisition module, the historical emergency monitoring patient data set and the real-time emergency monitoring patient data set both include patient identification, medical history information, vital signs and laboratory test results; the historical emergency monitoring patient data set also includes patient type; the patient type includes high-risk patients, ordinary patients and abnormal patients; the patient type is used as a label, and the label does not participate in the clustering operation.

[0010] Furthermore, in the data preprocessing module, the data cleaning is to delete data containing missing values ​​and duplicate values; the data conversion is to convert the data into a vector form based on encoding processing; and the standardization processing is to standardize the data based on normalization.

[0011] Furthermore, the dynamic clustering processing module specifically includes the following contents:

[0012] Local density calculation unit: use Gaussian fuzzy membership function to calculate the local disease density of each patient, and select K patients with the largest local disease density as the initial clustering centers; the formula used is as follows:

[0013] ;

[0014] In the formula, is the local density of the ith patient; S is the total number of scales, and s is the scale index; is the characteristic value of the ith patient at the sth scale; α s is the weight of the sth scale; c s is the cluster center value at the sth scale; is the standard deviation of the sth scale;

[0015] The disease distance calculation unit between patients; introduce feature weights to use Euclidean distance to calculate the disease distance between patients and cluster centers, and allocate patients based on the minimum disease distance principle. The formula used to calculate the disease distance is as follows:

[0016] ;

[0017] In the formula, is the disease distance of the i-th patient; is the total number of features, k is the feature index; is the weight of the kth feature; and are the values ​​of the i-th patient and the j-th patient on the k-th feature, respectively;

[0018] Patient type identification unit; ordinary patients and borderline patients are processed normally; transitional patients and noise patients are not processed; specifically including:

[0019] Identify common patients; select patients whose local density is greater than the average density of all patients as common patients, and calculate the average density. The formula used is as follows:

[0020] ;

[0021] ;

[0022] In the formula, is the average density; n is the total number of patients;

[0023] Dynamically adjust the noise parameters at each iteration, ;in, and are the noise parameters at the t-th iteration and the t-1-th iteration respectively; is the learning rate; is the number of historical patient data in the cluster; is the number of labels of the largest number of historical patient data in the cluster; is to select the minimum value in the brackets ;

[0024] The local anomaly factor is introduced to measure the local density difference between the patient and his neighbors to detect the noise patient. The noise patient is expressed as: ;in, ; LOF(·) is the local outlier factor; KNN(·) is the neighbor set; is the local density of the jth patient; x i is the i-th patient; x j is the jth patient;

[0025] The noise parameters are used to identify the boundary patients, and the remaining patients are regarded as transition patients, and the transition patients are not processed; the boundary points are expressed as: ;in, ; ; is the rate of change; is the variance of the patient's disease distance; is the average disease distance of the patients;

[0026] Iteration unit; clustering patients based on K-means clustering principle;

[0027] Clustering decision unit; set a decision threshold, when the cluster converges, the average cluster When it is higher than the judgment threshold, clustering ends; if the maximum number of iterations is reached, it goes to the parameter optimization unit; otherwise, it continues to iterate;

[0028] Parameter optimization unit; specifically including:

[0029] Establish parameter optimization space based on initial cluster center, feature weight, scale weight, total number of scales, initial noise parameter and learning rate;

[0030] Individual initialization: Initialize and optimize individual positions, expressed as: ; where x i1+1,j1 and x i1,j1 are the initialization positions of the j1th dimension of the i1+1th and i1th individuals respectively; r is a random number between 1 and 2; the average cluster convergence of the historical emergency care patient data set and the real-time emergency care patient data set based on the individual position As individual fitness value;

[0031] Calculate the dynamic adaptive weight; expressed as: ;in, is the dynamic adaptive weight; t1max is the maximum number of searches, which is set to 50 here; t1 is the current number of searches; e is a natural constant;

[0032] Position movement; for non-global optimal individuals, ; The global optimal individual is the individual with the largest individual fitness value; the global optimal individual is ;in, and are the positions of the i1th individual in the j1th dimension at the t1+1th and t1th searches respectively; Levy is a Lévy random number, which refers to a random variable that obeys the Lévy distribution; , and are the optimal individual positions during the t1-1th, t1th, and t1+1th searches respectively; rand is a random number between 0 and 1; is the worst individual position; is the individual empirical optimal position; exp(·) is the exponential function; i1 is the individual index;

[0033] Search judgment; when there is an individual fitness value higher than the judgment threshold, the individual position is the optimal parameter, and the clustering result is obtained at this time; if the maximum search number is reached, return to individual initialization; otherwise, return to calculate the dynamic adaptive weight.

[0034] Furthermore, the emergency monitoring management module specifically includes the following contents:

[0035] To identify high-density areas, the formula used is as follows:

[0036] ;

[0037] In the formula, B c It is a high-density area; is the disease distance of the jth patient;

[0038] Label assignment: Starting from the cluster center, assign labels to patients in all high-density areas; assign labels to real-time patient data based on the labels in the historical emergency care patient dataset in the high-density area. The formula used for label assignment is as follows:

[0039] ;

[0040] Where L(·) is the assigned label, L is the label; p j2 is the probability of the j2th label category; N is the total number of categories, which is 3 here; is to select L that takes the minimum value in the brackets;

[0041] For border patients, the distance between them and different high-density areas is calculated, the high-density area with the shortest distance is selected, and labels are assigned. The formula used to calculate the distance between the patient and the high-density area is as follows:

[0042] ;

[0043] In the formula, SB(x b ,B c ) is patient x b With high density area B c The distance between KNN(x b ,B c ) is the set of neighbors of the patient in the high-density area; x b is the bth patient;

[0044] If a patient is located between multiple high-density areas, the high-density area with the shortest distance is selected and a label is assigned to the patient, expressed as:

[0045] ;

[0046] In the formula, p c(·) is the probability of belonging to a high-density region; C1 is the set of high-density regions, c is the high-density region index; B(·) is the selection of high-density regions; It is to select Bc that takes the minimum value in the brackets;

[0047] After the label allocation is completed, for high-risk patients, an early warning will be immediately issued to medical staff through the emergency management system; for abnormal patients, high-frequency monitoring will be carried out, that is, high-frequency sampling of patients will be performed for clustering processing. If they are continuously judged as abnormal patients during the high-frequency monitoring process, an early warning will also be issued to medical staff through the emergency management system.

[0048] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0049] (1) In view of the problems that general emergency monitoring management systems cannot fully consider the diversity of patients' conditions, resulting in the omission of key patients, and the insufficient processing capacity of noisy patient data, which leads to misjudgment and missed judgment, this solution uses multi-scale feature processing and Gaussian fuzzy membership function to calculate the local condition density of patients, and dynamically adjusts the clustering center and noise parameters, adaptively adjusts the noise processing, and improves the patient classification accuracy; introduces dynamic parameter optimization and iterative optimization mechanism to efficiently cluster patients; introduces local abnormal factors to ensure that the impact of noise data on the overall system is minimized; thereby ensuring the efficiency and accuracy of the emergency monitoring management system.

[0050] (2) In view of the problem that general emergency monitoring management systems improperly handle fuzzy patient data, resulting in unclear classification or misdiagnosis, and are unable to flexibly respond to the complexity of different conditions, resulting in insufficient classification accuracy, this solution uses high-density area recognition to identify patients in areas with high disease concentrations, and quickly classifies patients through historical data labels, ensuring that each patient receives a reasonable disease classification and treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A flowchart of an emergency monitoring management system based on artificial intelligence provided by the present invention;

[0052] Figure 2 Schematic diagram of the flow of the dynamic clustering processing module.

[0053] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0055] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0056] Example 1, see Figure 1 , an artificial intelligence-based emergency monitoring management system provided by the present invention includes a data acquisition module, a data preprocessing module, a dynamic clustering processing module and an emergency monitoring management module;

[0057] The data acquisition module collects historical emergency monitoring patient data sets and real-time emergency monitoring patient data sets; and sends the data to the data preprocessing module;

[0058] The data preprocessing module receives the data sent by the data acquisition module; performs data cleaning, data conversion and standardization on the collected data; and sends the data to the dynamic clustering processing module;

[0059] The dynamic clustering processing module receives the data sent by the data preprocessing module; uses multi-scale feature processing and Gaussian fuzzy membership function to calculate the local disease density of the patient, and dynamically adjusts the cluster center and noise parameters; introduces dynamic parameter optimization and iterative optimization mechanism and local abnormal factors, and finally realizes dynamic clustering processing; and sends the data to the emergency monitoring management module;

[0060] The emergency monitoring management module receives data sent by the dynamic clustering processing module; uses high-density area recognition to determine patients in disease-intensive areas, and quickly classifies patients through historical data tags, thereby realizing emergency monitoring management.

[0061] Example 2, see Figure 1This embodiment is based on the above embodiment. In the data acquisition module, the historical emergency monitoring patient data set and the real-time emergency monitoring patient data set both include patient identification, medical history information, vital signs and laboratory test results; the historical emergency monitoring patient data set also includes patient type; patient types include high-risk patients, ordinary patients and abnormal patients; the patient type is used as a label, and the label does not participate in clustering operations; the patient identification is a mark for the patient data and does not participate in data operations.

[0062] Example 3, see Figure 1 This embodiment is based on the above embodiment. In the data preprocessing module, data cleaning is to delete data containing missing values ​​and duplicate values; data conversion is to convert data into vector form based on encoding processing; and standardization processing is to standardize data based on normalization.

[0063] Example 4, see Figure 1 and Figure 2 This embodiment is based on the above embodiment, and the dynamic clustering processing module specifically includes the following contents:

[0064] Local density calculation unit; use Gaussian fuzzy membership function to calculate the local disease density of each patient, and select K patients with the largest local disease density as the initial clustering centers; the value range of K is 5-15; the formula used is as follows:

[0065] ;

[0066] In the formula, is the local density of the ith patient; S is the total number of scales, ranging from 2 to 10; s is the scale index; is the characteristic value of the ith patient at the sth scale; α s is the weight of the sth scale; c s is the cluster center value at the sth scale; is the standard deviation of the sth scale;

[0067] The disease distance calculation unit between patients; introduce feature weights to use Euclidean distance to calculate the disease distance between patients and cluster centers, and allocate patients based on the minimum disease distance principle. The formula used to calculate the disease distance is as follows:

[0068] ;

[0069] In the formula, is the disease distance of the i-th patient; is the total number of features, here set to 20, k is the feature index; is the weight of the kth feature, ranging from 0 to 1; and are the values ​​of the i-th patient and the j-th patient on the k-th feature, respectively;

[0070] Patient type identification unit; ordinary patients and borderline patients are processed normally; transitional patients and noise patients are not processed; specifically including:

[0071] Identify common patients; select patients whose local density is greater than the average density of all patients as common patients, and calculate the average density. The formula used is as follows:

[0072] ;

[0073] ;

[0074] In the formula, is the average density; n is the total number of patients, which is 1000 here;

[0075] Dynamically adjust the noise parameters at each iteration, ;in, and are the noise parameters at the t-th iteration and the t-1-th iteration respectively, and the initial noise parameter ranges from 0.1 to 0.5; is the learning rate, ranging from 0.01 to 0.1; is the number of historical patient data in the cluster; is the number of labels of the largest number of historical patient data in the cluster; is to select the minimum value in the brackets ;

[0076] The local anomaly factor is introduced to measure the local density difference between the patient and his neighbors to detect the noise patient. The noise patient is expressed as: ;in, ; LOF(·) is the local outlier factor; KNN(·) is the neighbor set; is the local density of the jth patient; x i is the i-th patient; x j is the jth patient;

[0077] The noise parameters are used to identify the boundary patients, and the remaining patients are regarded as transition patients, and the transition patients are not processed; the boundary points are expressed as: ;in, ; ; is the rate of change; is the variance of the patient's condition distance; is the average disease distance of the patients;

[0078] Iteration unit; clustering patients based on K-means clustering principle;

[0079] Clustering decision unit; set a decision threshold, here the value is 0.9, when the clustering converges, the average cluster When it is higher than the judgment threshold, clustering ends; if the maximum number of iterations is reached, which is 30, it will go to the parameter optimization unit; otherwise, it will continue to iterate;

[0080] Parameter optimization unit; specifically including:

[0081] Establish parameter optimization space based on initial cluster center, feature weight, scale weight, total number of scales, initial noise parameter and learning rate;

[0082] Individual initialization: Initialize and optimize individual positions, expressed as: ; where x i1+1,j1 and x i1,j1 are the initialization positions of the j1th dimension of the i1+1th and i1th individuals respectively; r is a random number between 1 and 2; the average cluster convergence of the historical emergency care patient data set and the real-time emergency care patient data set based on the individual position As individual fitness value;

[0083] Calculate the dynamic adaptive weight; expressed as: ;in, is the dynamic adaptive weight; t1max is the maximum number of searches, which is set to 50 here; t1 is the current number of searches; e is a natural constant;

[0084] Position movement; for non-global optimal individuals, ; The global optimal individual is the individual with the largest individual fitness value; the global optimal individual is ;in, and are the positions of the i1th individual in the j1th dimension at the t1+1th and t1th searches respectively; Levy is a Lévy random number, which refers to a random variable that obeys the Lévy distribution; , and are the optimal individual positions during the t1-1th, t1th, and t1+1th searches respectively; rand is a random number between 0 and 1; is the worst individual position; is the individual empirical optimal position; exp(·) is the exponential function; i1 is the individual index;

[0085] Search judgment; when there is an individual fitness value higher than the judgment threshold, the individual position is the optimal parameter, and the clustering result is obtained at this time; if the maximum search number is reached, return to individual initialization; otherwise, return to calculate the dynamic adaptive weight.

[0086] By performing the above operations, in order to address the problems that general emergency monitoring management systems cannot fully consider the diversity of patients' conditions, resulting in the omission of key patients, and the insufficient processing capacity of noisy patient data, which leads to misjudgment and missed judgments, this solution uses multi-scale feature processing and Gaussian fuzzy membership function to calculate the local condition density of patients, and dynamically adjusts the clustering center and noise parameters, adaptively adjusts the noise processing, and improves the patient classification accuracy; introduces dynamic parameter optimization and iterative optimization mechanisms to efficiently cluster patients; introduces local abnormal factors to ensure that the impact of noise data on the overall system is minimized; thereby ensuring the efficiency and accuracy of the emergency monitoring management system.

[0087] Example 5, see Figure 1 This embodiment is based on the above embodiment, and the emergency monitoring management module specifically includes the following contents:

[0088] To identify high-density areas, the formula used is as follows:

[0089] ;

[0090] In the formula, B c It is a high-density area; is the disease distance of the jth patient;

[0091] Label assignment: Starting from the cluster center, assign labels to patients in all high-density areas; assign labels to real-time patient data based on the labels in the historical emergency care patient dataset in the high-density area. The formula used for label assignment is as follows:

[0092] ;

[0093] Where L(·) is the assigned label, L is the label; p j2 is the probability of the j2th label category; N is the total number of categories, which is 3 here; is to select L that takes the minimum value in the brackets;

[0094] For border patients, by calculating the distance between them and different high-density areas, the high-density area with the shortest distance is selected, and a label is assigned to the patient. The formula used to calculate the distance between the patient and the high-density area is as follows:

[0095] ;

[0096] In the formula, SB(x b ,B c ) is patient x b With high density area B c The distance between KNN(x b ,B c ) is the set of neighbors of the patient in the high-density area; x bis the bth patient;

[0097] If a patient is located between multiple high-density areas, the high-density area with the shortest distance is selected and a label is assigned to the point, expressed as:

[0098] ;

[0099] In the formula, p c (·) is the probability of belonging to a high-density region; C1 is the set of high-density regions, c is the high-density region index; B(·) is the selection of high-density regions; It is to select Bc that takes the minimum value in the brackets;

[0100] After the label allocation is completed, for high-risk patients, an early warning will be immediately issued to medical staff through the emergency management system; for abnormal patients, high-frequency monitoring will be carried out, that is, high-frequency sampling of patients will be performed for clustering processing. If they are continuously judged as abnormal patients during the high-frequency monitoring process, an early warning will also be issued to medical staff through the emergency management system.

[0101] By performing the above operations, the general emergency monitoring management system has the problem of improper handling of fuzzy patient data, resulting in unclear classification or misdiagnosis, and difficulty in flexibly responding to the complexity of different diseases, resulting in insufficient classification accuracy. This solution uses high-density area recognition to identify patients in areas with dense disease conditions, and quickly classifies patients through historical data labels, ensuring that each patient receives reasonable disease classification and treatment.

[0102] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0103] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0104] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. An emergency monitoring management system based on artificial intelligence, characterized by: The system includes a data acquisition module, a data preprocessing module, a dynamic clustering processing module and an emergency monitoring management module; The data acquisition module acquires historical emergency monitoring patient data sets and real-time emergency monitoring patient data sets; The data preprocessing module performs data cleaning, data conversion and standardization on the collected data; The dynamic clustering processing module uses multi-scale feature processing and Gaussian fuzzy membership function to calculate the local disease density of patients, and dynamically adjusts the cluster center and noise parameters; introduces dynamic parameter optimization and iterative optimization mechanism and local abnormal factors, and finally realizes dynamic clustering processing; The emergency monitoring management module uses high-density area recognition to identify patients in densely populated areas, and quickly classifies patients through historical data tags, thereby achieving emergency monitoring management; The dynamic clustering processing module specifically includes the following contents: Local density calculation unit: use Gaussian fuzzy membership function to calculate the local disease density of each patient, and select K patients with the largest local disease density as the initial clustering centers; the formula used is as follows: ； In the formula, is the local density of the ith patient; S is the total number of scales, s is the scale index; is the characteristic value of the ith patient at the sth scale; α s is the weight of the sth scale; c s is the cluster center value at the sth scale; is the standard deviation of the sth scale; The disease distance calculation unit between patients; introduce feature weights to use Euclidean distance to calculate the disease distance between patients and cluster centers, and allocate patients based on the minimum disease distance principle. The formula used to calculate the disease distance is as follows: ； In the formula, is the disease distance of the i-th patient; is the total number of features, k is the feature index; is the weight of the kth feature; and are the values ​​of the i-th patient and the j-th patient on the k-th feature, respectively; Patient type identification unit; ordinary patients and borderline patients are processed normally; transitional patients and noise patients are not processed; specifically including: Identify common patients; select patients whose local density is greater than the average density of all patients as common patients, and calculate the average density. The formula used is as follows: ； ； In the formula, is the average density; n is the total number of patients; Dynamically adjust the noise parameters at each iteration, ;in, and are the noise parameters at the t-th iteration and the t-1-th iteration respectively; is the learning rate; is the number of historical patient data in the cluster; is the number of labels of the largest number of historical patient data in the cluster; is to select the minimum value in the brackets ; The local anomaly factor is introduced to measure the local density difference between the patient and his neighbors to detect the noise patient. The noise patient is expressed as: ;in, ; LOF(·) is the local outlier factor; KNN(·) is the neighbor set; is the local density of the jth patient; x i is the i-th patient; x j is the jth patient; The noise parameters are used to identify the boundary patients, and the remaining patients are regarded as transition patients, and the transition patients are not processed; the boundary points are expressed as: ;in, ; ; is the rate of change; is the variance of the patient's condition distance; is the average disease distance of the patients; Iteration unit; clustering patients based on K-means clustering principle; Clustering decision unit; set a decision threshold, when the cluster converges, the average cluster When it is higher than the judgment threshold, clustering ends; if the maximum number of iterations is reached, it goes to the parameter optimization unit; otherwise, it continues to iterate; Parameter optimization unit; specifically includes the following contents: Establish parameter optimization space based on initial cluster center, feature weight, scale weight, total number of scales, initial noise parameter and learning rate; Individual initialization: Initialize and optimize individual positions, expressed as: ; where x i1+1,j1 and x i1,j1 are the initialization positions of the j1th dimension of the i1+1th and i1th individuals respectively; r is a random number between 1 and 2; the average cluster convergence of the historical emergency care patient data set and the real-time emergency care patient data set based on the individual position As individual fitness value; Calculate the dynamic adaptive weight; expressed as: ;in, is the dynamic adaptive weight; t1max is the maximum number of searches, which is set to 50 here; t1 is the current number of searches; e is a natural constant; Position movement; for non-global optimal individuals, ; The global optimal individual is the individual with the largest individual fitness value; the global optimal individual is ;in, and are the positions of the i1th individual in the j1th dimension at the t1+1th and t1th searches respectively; Levy is a Lévy random number, which refers to a random variable that obeys the Lévy distribution; , and are the optimal individual positions during the t1-1th, t1th, and t1+1th searches respectively; rand is a random number between 0 and 1; is the worst individual position; is the individual empirical optimal position; exp(·) is the exponential function; i1 is the individual index; Search judgment; when there is an individual fitness value higher than the judgment threshold, the individual position is the optimal parameter, and the clustering result is obtained at this time; if the maximum search number is reached, the individual initialization is returned; otherwise, the dynamic adaptive weight is calculated; The emergency monitoring management module specifically includes the following contents: To identify high-density areas, the formula used is as follows: ； In the formula, B c It is a high-density area; is the disease distance of the jth patient; Label assignment: Starting from the cluster center, assign labels to patients in all high-density areas; assign labels to real-time patient data based on the labels in the historical emergency care patient dataset in the high-density area. The formula used for label assignment is as follows: ； Where L(·) is the assigned label, L is the label; p j2 is the probability of the j2th label category; N is the total number of categories, which is 3 here; is to select L that takes the minimum value in the brackets; For border patients, the distance between them and different high-density areas is calculated, the high-density area with the shortest distance is selected, and labels are assigned. The formula used to calculate the distance between the patient and the high-density area is as follows: ； In the formula, SB(x b ,B c ) is patient x b With high density area B c The distance between KNN(x b ,B c ) is the set of neighbors of the patient in the high-density area; x b is the bth patient; If a patient is located between multiple high-density areas, the high-density area with the shortest distance is selected and a label is assigned to the patient, expressed as: ； In the formula, p c (·) is the probability of belonging to a high-density region; C1 is the set of high-density regions, c is the high-density region index; B(·) is the selection of high-density regions; It is to select Bc that takes the minimum value in the brackets; After the label allocation is completed, for high-risk patients, an early warning will be immediately issued to medical staff through the emergency management system; for abnormal patients, high-frequency monitoring will be carried out, that is, high-frequency sampling of patients will be performed for clustering processing. If they are continuously judged as abnormal patients during the high-frequency monitoring process, an early warning will also be issued to medical staff through the emergency management system.

2. The artificial intelligence-based emergency monitoring management system according to claim 1, characterized in that: In the data acquisition module, the historical emergency monitoring patient data set and the real-time emergency monitoring patient data set both include patient identification, medical history information, vital signs and laboratory test results; the historical emergency monitoring patient data set also includes patient type; the patient type includes high-risk patients, ordinary patients and abnormal patients; the patient type is used as a label, and the label does not participate in the clustering operation.

3. The artificial intelligence-based emergency monitoring management system according to claim 1, characterized in that: In the data preprocessing module, the data cleaning is to delete data containing missing values ​​and duplicate values; the data conversion is to convert the data into a vector form based on encoding processing; and the standardization processing is to standardize the data based on normalization.

Images