A comprehensive supervision method and system for health management

Abstract

A comprehensive supervision method and system for health management proposed in the present application involves the field of big data analysis, makes full use of big data to analyze health anomalies, and determines the information acquisition frequency of the monitored institution by comprehensively utilizing the average values ​​of the monitored institution and similar institutions with the highest degree of comprehensive relevance. The determined information acquisition frequency is more in line with actual needs, and information acquisition methods are separately formulated for different items to be monitored, thereby ensuring the accuracy of health management data. In addition, based on frequent item sets, the correlation between data can be accurately utilized to accurately discover abnormal problems.

Description

Technical Field

[0001] The present invention belongs to the field of big data analysis, and in particular relates to a comprehensive supervision method and system for health management. Background Art

[0002] Based on the Internet, artificial intelligence, cloud computing, big data and other technologies, we can innovate work models, improve management efficiency, and further expand the scope of information technology applications, so as to create all-round and all-domain smart health monitoring and accelerate the deep integration of the Internet and health supervision. In the existing technology, health monitoring based on big data is usually only based on the analysis of the acquired data, and the timeliness and accuracy of health monitoring needs to be improved. Summary of the invention

[0003] The purpose of the present invention is to provide a comprehensive supervision method and system for health management to solve the technical problem in the prior art that the timeliness and accuracy of health condition monitoring need to be improved.

[0004] A comprehensive supervision method for health management, characterized in that the method comprises:

[0005] S1: Determine a first profile of the organization to be monitored based on basic information of the organization to be monitored;

[0006] S2: using the first portrait to perform similarity traversal in a first portrait set to obtain a second portrait having the highest similarity to the first portrait; the first portrait set includes portraits of several monitored institutions of the same type as the institution to be monitored;

[0007] S3: Acquire first historical information of a first monitored organization corresponding to the second portrait, and acquire a first abnormality probability according to the first historical information;

[0008] S4: determining a first information acquisition frequency according to the first abnormality probability;

[0009] S5: obtaining first historical monitoring abnormality information of the monitored institution, and inputting the first historical monitoring abnormality information and the first information acquisition frequency into an information acquisition frequency determination model to obtain a second information acquisition frequency;

[0010] S6: Determine M types of items to be monitored according to the type of the organization to be monitored, obtain the second information of each type of item to be monitored at the second information acquisition frequency, and upload it to the blockchain network; M is a natural number greater than or equal to 1;

[0011] S7: Inputting the second information into an abnormal data analysis model at a preset time interval to obtain a first abnormal data frequent item set;

[0012] S8: Determine, according to the first mapping relationship, a first target event corresponding to the first abnormal data frequent itemset.

[0013] Preferably, the S1 further includes preprocessing of the basic information, and the preprocessing specifically includes:

[0014] S11: removing obviously abnormal data from the basic information;

[0015] S12: input each first label in the basic information into the sensitive data recognition model in sequence to obtain N marker words; wherein N is a natural number greater than or equal to 1;

[0016] S13: For each of the first tags, calculating the first number of the tag words;

[0017] S14: If the first number is less than a first threshold, the marking word in the first label is anonymized; if the first number is greater than or equal to the first threshold, the first label is deleted.

[0018] Preferably, S2 comprises the following sub-steps:

[0019] S21: Acquire the first portrait set according to the first category of the organization to be monitored;

[0020] S22: representing the first portrait as a first vector, and representing multiple portraits in the first portrait set as multiple second vectors;

[0021] S23: Calculate the cosine similarity between the first vector and each of the second vectors respectively, and determine the portrait corresponding to the second vector with the highest cosine similarity as the second portrait.

[0022] Preferably, S3 includes the following sub-steps:

[0023] S31: Acquire first historical information of a first monitored organization corresponding to the second portrait;

[0024] S32: Inputting the first historical information into an abnormality probability prediction model to obtain the first abnormality probability.

[0025] Preferably, the abnormal probability prediction model is obtained based on historical data training, and the specific training process is as follows:

[0026] S321: Acquire historical monitoring sample data of multiple monitored institutions; the historical monitoring sample data includes category information, basic information, the number of data anomalies within a preset time interval, and the first degree of anomalies of the monitored institutions;

[0027] S322: Perform K-MEANS clustering on the multiple historical monitoring sample data to obtain multiple first categories, and count the ratio of the number of abnormalities in each of the first categories to the total number of monitoring times as the probability that the sample data in the first category will be abnormal within a future preset time period;

[0028] S323: Take the category information, basic information, number of data anomalies within a preset time interval, and the first degree of the anomaly of the monitored organization in each sample data as input, and take the probability of the anomaly occurring within a future preset time period as output, train a convolutional neural network, and obtain the anomaly probability prediction model.

[0029] Preferably, the S4 comprises the following sub-steps:

[0030] S41: acquiring an abnormal probability and an information acquisition frequency corresponding to each historical monitoring sample data according to the plurality of historical monitoring sample data;

[0031] S42: fitting a first relationship between the abnormal probability and the information acquisition frequency according to the plurality of abnormal probabilities and information acquisition frequencies: F=A*Y, where F is the information acquisition frequency, Y is the abnormal probability, and A is a fitting coefficient;

[0032] S43: Substitute the first abnormal probability into the first relationship to obtain the first information acquisition frequency.

[0033] Preferably, S5 comprises the following sub-steps:

[0034] S51: Determine the information acquisition frequency determination model according to the category of the organization to be monitored;

[0035] S52: inputting the first historical monitoring abnormality information, the average monitoring abnormality information of the same category, and the first information acquisition frequency into the information acquisition frequency determination model to obtain a first information acquisition frequency adjustment value;

[0036] S53: Obtain the second information acquisition frequency according to the first information acquisition frequency and the first information acquisition frequency adjustment value.

[0037] Preferably, S6 comprises the following sub-steps:

[0038] S61: Determine M types of items to be monitored according to the type of the organization to be monitored, where M is 1 or 2;

[0039] S62: If the value of M is 2, go to S63; if the value of M is 1, go to S64;

[0040] S63: for the first type of monitored items, respectively obtain the first monitoring data of the first monitoring node and the first verification data of the first verification node; for the second type of monitored items, obtain the second monitoring data of the second monitoring node, and go to S65;

[0041] S64: Acquire third monitoring data from the second monitoring node, and go to S65;

[0042] S65: Determine the second information based on the first monitoring data, the first verification data, the second monitoring data, and the third monitoring data, and upload it to the blockchain.

[0043] Preferably, the anomaly analysis model includes multiple frequent item sets of abnormal data of similar monitored institutions, and the second information is semantically recognized through the anomaly analysis model, so that one or more frequent item sets of the first abnormal data contained in the second information can be identified.

[0044] This application also proposes a comprehensive supervision system for health management, which is used to implement the above-mentioned comprehensive supervision method for health management.

[0045] A comprehensive supervision method and system for health management proposed in this application makes full use of big data analysis, and comprehensively uses the average values ​​of the monitored institutions and similar institutions with the highest correlation to determine the information acquisition frequency of the monitored institutions. The determined information acquisition frequency is more in line with actual needs, and information acquisition methods are separately formulated for different monitored items, thereby ensuring the accuracy of health problem monitoring. In addition, based on frequent item sets, the correlation between data can be accurately utilized to accurately discover health abnormalities. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0047] Figure 1 It is an execution flow chart of a comprehensive supervision method for health management of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments, wherein the illustrative embodiments and descriptions are only used to explain the present invention but are not intended to limit the present invention.

[0050] The following is a detailed description of a comprehensive supervision method and system for health management of the present invention.

[0051] This embodiment proposes a comprehensive supervision method for health management, the method flow is as follows Figure 1 As shown, the specific steps include:

[0052] S1: Determine a first profile of the organization to be monitored based on basic information of the organization to be monitored.

[0053] The institutions to be monitored refer to institutions where health monitoring is to be carried out, such as canteens, hospitals and other public places.

[0054] The basic information includes information used to describe the scale and working conditions of the institution to be monitored, and the basic information is used to match the monitored institution with the same magnitude as the institution to be monitored in S2. Taking a hospital as an example, the basic information may include the number of departments, the number of doctors, the average number of patients per day, the average proportion of health and health abnormalities in the region, etc.

[0055] After preprocessing the basic information, a number of first labels can be obtained, and the first labels are combined into the first portrait.

[0056] The preprocessing includes abnormal data removal and data desensitization.

[0057] The pre-processing specifically includes:

[0058] S11: removing obviously abnormal data from the basic information.

[0059] The obviously abnormal data includes empty data.

[0060] S12: Input each first label in the basic information into the sensitive data recognition model in sequence to obtain N tag words.

[0061] The N is a natural number greater than or equal to 1.

[0062] S13: For each of the first tags, calculate the first number of the tag words.

[0063] S14: If the first number is less than a first threshold, the marking word in the first label is anonymized; if the first number is greater than or equal to the first threshold, the first label is deleted.

[0064] S2: Performing a similarity traversal in a first portrait set using the first portrait to obtain a second portrait having the highest similarity to the first portrait.

[0065] The first portrait set includes portraits of several monitored organizations.

[0066] The S2 comprises the following sub-steps:

[0067] S21: Acquire the first portrait set according to the first category of the organization to be monitored.

[0068] Since there are many institutions to be subject to health monitoring, covering all aspects of society, and different types of institutions are usually not comparable in monitoring, it is necessary in this step to first obtain a first set of portraits of the same type as the first category based on the first category of the institution to be monitored.

[0069] S22: Represent the first portrait as a first vector, and represent multiple portraits in the first portrait set as multiple second vectors.

[0070] In this step, the first portrait or multiple portraits in the first portrait set are represented in vector form.

[0071] S23: Calculate the cosine similarity between the first vector and each of the second vectors respectively, and determine the portrait corresponding to the second vector with the highest cosine similarity as the second portrait.

[0072] S3: Obtain first historical information of the first monitored organization corresponding to the second portrait, and obtain a first abnormal probability according to the first historical information.

[0073] The first abnormality probability corresponds to the first monitored institution and is determined according to the number of abnormalities detected, the degree of abnormality and average monitoring data during the monitoring of the first monitored institution.

[0074] The S3 comprises the following sub-steps:

[0075] S31: Acquire first historical information of the first monitored organization corresponding to the second portrait.

[0076] The first historical information includes the first number of times data anomalies are found in the first monitored institution and the first degree of the anomalies in the historical monitoring behavior, wherein the first degree refers to the deviation of the data anomaly value from the average data anomaly value.

[0077] For example, when the institution to be monitored is a hospital, the usage quantity of radioactive medical consumables is one of the data to be monitored. When the usage quantity is monitored to deviate from the average value by more than a threshold, it is determined that data abnormality is monitored. At this time, it is also necessary to monitor the degree of deviation between the data abnormality value and the average data abnormality value, so as to determine the first degree, which is proportional to the degree of deviation. The average data abnormality value refers to the average data abnormality value when the usage data of radioactive medical consumables is abnormal.

[0078] S32: Inputting the first historical information into an abnormality probability prediction model to obtain the first abnormality probability.

[0079] The abnormal probability prediction model is obtained based on historical data training, and the specific training process is as follows:

[0080] First, historical monitoring sample data of multiple monitored institutions are obtained, wherein the historical monitoring sample data includes category information, basic information, the number of data anomalies within a preset time interval, and the first degree of anomalies of the monitored institutions.

[0081] Secondly, K-MEANS clustering is performed on the multiple historical monitoring sample data to obtain multiple first categories, and the ratio of the number of abnormalities occurring in each of the first categories to the total number of monitoring times is counted as the probability that the sample data in the first category will be abnormal within a preset time period in the future.

[0082] Finally, the category information, basic information, the number of data anomalies within a preset time interval, and the first degree of the anomaly of the monitored organization in each sample data are taken as input, and the probability of the anomaly occurring within a preset time period in the future is taken as output to train the convolutional neural network to obtain the anomaly probability prediction model.

[0083] S4: Determine a first information acquisition frequency according to the first abnormality probability.

[0084] The first information acquisition frequency refers to the frequency of acquiring the monitoring information of the monitored institution, and the first information acquisition frequency is positively correlated with the first abnormality probability.

[0085] The S4 comprises the following sub-steps:

[0086] S41: According to the plurality of historical monitoring sample data, obtain the abnormal probability and information acquisition frequency corresponding to each historical monitoring sample data.

[0087] The abnormal probability corresponding to each historical monitoring sample data is obtained by statistically analyzing the abnormal conditions of the historical monitoring sample within a preset time period, specifically, the ratio of the number of abnormalities occurring within the preset time period to the total number of monitoring times.

[0088] S42: According to the plurality of abnormal probabilities and information acquisition frequencies, a first relationship between the abnormal probability and the information acquisition frequency is obtained by fitting: F=A*Y, wherein F is the information acquisition frequency, Y is the abnormal probability, and A is the fitting coefficient.

[0089] S43: Substitute the first abnormal probability into the first relationship to obtain the first information acquisition frequency.

[0090] S5: Acquire first historical monitoring abnormality information of the monitored institution, and input the first historical monitoring abnormality information and the first information acquisition frequency into an information acquisition frequency determination model to obtain a second information acquisition frequency.

[0091] In S4, the first information acquisition frequency is determined based on the first abnormal probability corresponding to the first monitored institution that has a higher similarity to the monitored institution, but the first information acquisition frequency is only inferred based on the data of similar institutions. In this step, the first information acquisition frequency also needs to be adjusted based on the historical monitoring abnormality information of the monitored institution.

[0092] The S5 comprises the following sub-steps:

[0093] S51: Determine the information acquisition frequency determination model according to the category of the organization to be monitored.

[0094] Different types of institutions to be monitored have different speeds of generating information, and thus different average information monitoring frequencies. For example, the frequency of information monitoring in a hospital is significantly higher than that in a canteen.

[0095] Therefore, in this step, a plurality of information acquisition frequency determination models are trained respectively according to the different types of the institutions to be monitored.

[0096] The information acquisition frequency determination model is obtained by training in the following manner:

[0097] First, multiple pieces of sample data are obtained. The monitored institutions to which the multiple pieces of sample data belong are of the same category. Each piece of sample data includes historical monitoring abnormality information, average monitoring abnormality information of the same category, average information acquisition frequency, and information acquisition frequency adjustment value.

[0098] Among them, the historical monitoring abnormality information is the historical monitoring abnormality information corresponding to a piece of the sample data, which may specifically include the frequency of occurrence of abnormal information within a preset time period; the average monitoring abnormality information refers to the average frequency of occurrence of abnormal information of the same category of monitored institutions within a preset time period; the average information acquisition frequency refers to the average information acquisition frequency of the same category of monitored institutions; the information acquisition frequency adjustment value refers to the difference between the information acquisition frequency actually determined by the sample data and the average information acquisition frequency.

[0099] Secondly, the historical monitoring abnormality information in each sample, the average monitoring abnormality information under the same category, and the average information acquisition frequency are taken as input, and the information acquisition frequency adjustment value is taken as output to train a convolutional neural network model to obtain the information acquisition frequency determination model.

[0100] The information acquisition frequency determination model is trained for each type of monitored organization respectively, so that a plurality of information acquisition frequency determination models can be obtained.

[0101] S52: Inputting the first historical monitoring abnormality information, the average monitoring abnormality information of the same category, and the first information acquisition frequency into the information acquisition frequency determination model to obtain a first information acquisition frequency adjustment value.

[0102] S53: Obtain the second information acquisition frequency according to the first information acquisition frequency and the first information acquisition frequency adjustment value.

[0103] Preferably, the first information acquisition frequency can be obtained by adding the first information acquisition frequency and the first information acquisition frequency adjustment value.

[0104] S6: According to the category of the organization to be monitored, determine M categories of items to be monitored, obtain the second information of each category of the items to be monitored at the second information acquisition frequency, and upload it to the blockchain network.

[0105] Wherein, each of the items to be monitored has a different monitoring method.

[0106] The M is preferably 2, the first category of monitored items are items that need back-to-back verification, and the second category of monitored items are ordinary items. For example, the monitored institution is a hospital. The radioactive supplies it needs to use belong to the first category of monitored items. The corresponding data of the inventory consumption node and the smart trash can node need to be uploaded to the blockchain respectively. The smart trash can is equipped with a sensor that can detect radioactive consumables. Other non-severely contaminated supplies needed for surgery belong to the first category of monitored items and can be directly uploaded to the blockchain network through the inventory consumption node.

[0107] The S6 comprises the following sub-steps:

[0108] S61: Determine M-type items to be monitored according to the category of the organization to be monitored.

[0109] The M may be 1 or 2. For example, if the institution to be monitored is a swimming pool, there is no obvious difference in the various monitored items representing the hygienic conditions, so M may be 1; if the institution to be monitored is a hospital, since the hospital includes radiological medical consumables and ordinary medical consumables, the corresponding monitoring methods and strictness are different, so M may be 2.

[0110] S62: If the value of M is 2, go to S63; if the value of M is 1, go to S64.

[0111] S63: For the first type of monitored items, obtain the first monitoring data of the first monitoring node and the first verification data of the first verification node respectively; for the second type of monitored items, obtain the second monitoring data of the second monitoring node, and go to S65.

[0112] Taking the hospital as an example, the first type of monitored items are radioactive medical consumables, the first monitoring node is an inventory consumption node, and the first verification node is an intelligent trash can node; the second type of monitored items are ordinary medical consumables, and the second monitoring node is an inventory consumption node.

[0113] S64: Acquire third monitoring data from the second monitoring node, and go to S65.

[0114] S65: Determine the second information based on the first monitoring data, the first verification data, the second monitoring data, and the third monitoring data, and upload it to the blockchain.

[0115] Preferably, if the value of M is 2, the first monitoring data, the first verification data, and the second monitoring data are determined as the second information; if the value of M is 1, the third monitoring data is determined as the second information.

[0116] S7: Inputting the second information into an abnormal data analysis model at preset time intervals to obtain a first abnormal data frequent item set.

[0117] Typically, a specific hygiene problem corresponds to a combination of several abnormal monitoring data. For example, for radioactive consumables in a hospital, if the first monitoring data and the first verification data cannot be accurately matched, and the number of matching consumables used corresponding to the radioactive consumables obviously does not match the first monitoring data, then a set of abnormal monitoring data combinations [first monitoring data, first verification data, number of matching consumables used] can be formed.

[0118] The anomaly analysis model in this step includes multiple frequent itemsets of abnormal data of similar monitored institutions. The second information is semantically identified through the anomaly analysis model, so that one or more frequent itemsets of the first abnormal data contained in the second information can be identified.

[0119] S8: Determine, according to the second mapping relationship, a first target event corresponding to the first abnormal data frequent itemset.

[0120] The target event represents an event of hygiene abnormality.

[0121] The second mapping relationship includes a one-to-one correspondence between each of the first abnormal data frequent itemsets and a target event. Therefore, the first abnormal event corresponding to the first abnormal data frequent itemsets can be determined through the second mapping relationship.

[0122] This application also proposes a comprehensive supervision system for health management, which is used to implement the above-mentioned comprehensive supervision method for health management.

[0123] A comprehensive supervision method and system for health management proposed in this application makes full use of big data analysis, and comprehensively uses the average values ​​of the monitored institutions and similar institutions with the highest correlation to determine the information acquisition frequency of the monitored institutions. The determined information acquisition frequency is more in line with actual needs, and information acquisition methods are separately formulated for different monitored items, thereby ensuring the accuracy of health problem monitoring. In addition, based on frequent item sets, the correlation between data can be accurately utilized to accurately discover health abnormalities.

[0124] The above description is only a preferred embodiment of the present invention, so all equivalent changes or modifications made according to the structure, characteristics and principles described in the scope of the patent application of the present invention are included in the scope of the patent application of the present invention.

Claims

1. A comprehensive supervision method for health management, characterized in that: The method includes: S1: Determine a first profile of the organization to be monitored based on basic information of the organization to be monitored; the basic information includes information used to describe the scale and working conditions of the organization to be monitored, and the basic information is used to match a monitored organization of the same magnitude as the organization to be monitored in S2; S2: using the first portrait to perform similarity traversal in a first portrait set to obtain a second portrait having the highest similarity to the first portrait; the first portrait set includes portraits of several monitored institutions of the same type as the institution to be monitored; S3: Acquire first historical information of the first monitored organization corresponding to the second portrait, and acquire a first abnormality probability according to the first historical information; the first historical information includes a first number of times data abnormalities of the first monitored organization are found in historical monitoring behaviors and a first degree of abnormality; S4: determining a first information acquisition frequency according to the first abnormality probability; the first information acquisition frequency refers to a frequency of acquiring monitoring information of the monitored institution; S5: obtaining first historical monitoring abnormality information of the monitored institution, and inputting the first historical monitoring abnormality information and the first information acquisition frequency into an information acquisition frequency determination model to obtain a second information acquisition frequency; S6: According to the category of the institution to be monitored, determine M categories of items to be monitored, obtain the second information of each category of items to be monitored at the second information acquisition frequency, and upload it to the blockchain network; M is a natural number greater than or equal to 1; S6 includes: according to the category of the institution to be monitored, determine M categories of items to be monitored; M is 1 or 2; If the value of M is 2, for the first type of monitored items, the first monitoring data of the first monitoring node and the first verification data of the first verification node are obtained respectively; for the second type of monitored items, the second monitoring data of the second monitoring node is obtained; the first monitoring data, the first verification data, and the second monitoring data are determined as the second information; If the value of M is 1, third monitoring data is obtained from the second monitoring node, and the third monitoring data is determined as the second information; S7: Inputting the second information into an abnormal data analysis model at a preset time interval to obtain a first abnormal data frequent item set; S8: Determine a first target event corresponding to the first abnormal data frequent item set according to the first mapping relationship; the target event represents an event of abnormal hygiene.

2. A comprehensive supervision method for health management according to claim 1, characterized in that: The S1 also includes preprocessing of the basic information, and the preprocessing specifically includes: S11: removing obviously abnormal data from the basic information; S12: input each first label in the basic information into the sensitive data recognition model in sequence to obtain N marker words; wherein N is a natural number greater than or equal to 1; S13: For each of the first tags, calculating the first number of the tag words; S14: If the first number is less than a first threshold, the marking word in the first label is anonymized; if the first number is greater than or equal to the first threshold, the first label is deleted.

3. A comprehensive supervision method for health management according to claim 2, characterized in that: The S2 comprises the following sub-steps: S21: Acquire the first portrait set according to the first category of the organization to be monitored; S22: representing the first portrait as a first vector, and representing multiple portraits in the first portrait set as multiple second vectors; S23: Calculate the cosine similarity between the first vector and each of the second vectors respectively, and determine the portrait corresponding to the second vector with the highest cosine similarity as the second portrait.

4. A comprehensive supervision method for health management according to claim 3, characterized in that: The S3 comprises the following sub-steps: S31: Acquire first historical information of a first monitored organization corresponding to the second portrait; S32: Inputting the first historical information into an abnormality probability prediction model to obtain the first abnormality probability.

5. A comprehensive supervision method for health management according to claim 4, characterized in that: The abnormal probability prediction model is obtained based on historical data training, and the specific training process is as follows: S321: Acquire historical monitoring sample data of multiple monitored institutions; the historical monitoring sample data includes category information, basic information, the number of data anomalies within a preset time interval, and the first degree of anomalies of the monitored institutions; S322: Perform K-MEANS clustering on the multiple historical monitoring sample data to obtain multiple first categories, and count the ratio of the number of abnormalities in each of the first categories to the total number of monitoring times as the probability that the sample data in the first category will be abnormal within a future preset time period; S323: Take the category information, basic information, number of data anomalies within a preset time interval, and the first degree of the anomaly of the monitored organization in each sample data as input, and take the probability of the anomaly occurring within a future preset time period as output, train a convolutional neural network, and obtain the anomaly probability prediction model.

6. A comprehensive supervision method for health management according to claim 5, characterized in that: The S4 comprises the following sub-steps: S41: acquiring an abnormal probability and an information acquisition frequency corresponding to each historical monitoring sample data according to the plurality of historical monitoring sample data; S42: fitting a first relationship between the abnormal probability and the information acquisition frequency according to the plurality of abnormal probabilities and information acquisition frequencies: F=A*Y, where F is the information acquisition frequency, Y is the abnormal probability, and A is a fitting coefficient; S43: Substitute the first abnormal probability into the first relationship to obtain the first information acquisition frequency.

7. A comprehensive supervision method for health management according to claim 6, characterized in that: The S5 comprises the following sub-steps: S51: Determine the information acquisition frequency determination model according to the category of the organization to be monitored; S52: inputting the first historical monitoring abnormality information, the average monitoring abnormality information of the same category, and the first information acquisition frequency into the information acquisition frequency determination model to obtain a first information acquisition frequency adjustment value; S53: Obtain the second information acquisition frequency according to the first information acquisition frequency and the first information acquisition frequency adjustment value.

8. A comprehensive supervision method for health management according to claim 7, characterized in that: The abnormal data analysis model includes multiple abnormal data frequent item sets of similar monitored institutions. The second information is semantically identified through the abnormal data analysis model, so that one or more of the first abnormal data frequent item sets contained in the second information can be identified.

9. A comprehensive supervision system for health management, used to implement a comprehensive supervision method for health management as described in any one of claims 1-8 above.

Images