Abnormal behavior monitoring method, device, equipment and storage medium for slow door

Abstract

The present invention relates to the field of digital medicine, and provides a method, device, equipment and storage medium for monitoring abnormal behavior of chronic diseases, wherein the method comprises: obtaining chronic disease information of a designated user, and drug information of each reimbursement drug for the designated user based on the chronic disease, inputting into a chronic disease correlation calculation model for calculation, obtaining the correlation between each drug information and the chronic disease information, and identifying the abnormal behavior of the designated user based on the correlation. Thus, automatic identification of abnormal behavior is achieved, saving the time of relevant personnel to check the drug information of each patient one by one, thereby reducing labor costs.

Description

Technical Field

[0001] The present invention relates to the field of digital medicine, and in particular to a method, device, equipment and storage medium for monitoring abnormal behavior of a patient. Background Art

[0002] Outpatient chronic disease is the abbreviation of outpatient chronic disease. Its main characteristics are long course of disease, complex etiology, and long-term medication. When the disease develops to a certain stage, it usually causes damage to multiple organs such as the heart, brain, and kidneys, causing disability or even death. Patients with outpatient chronic disease qualifications have long-term preferential policies for obtaining chronic disease drugs and related treatment projects at low prices from hospitals. At the same time, these preferential policies may also be exploited by patients to earn the difference by reselling drugs and then make profits from it, so it is necessary to supervise this abnormal behavior.

[0003] The traditional supervision method is mainly to monitor the medication use of each patient by experienced relevant personnel. However, experienced relevant personnel need to undergo a lot of training, resulting in huge manpower costs. Summary of the invention

[0004] The main purpose of the present invention is to provide a method, device, equipment and storage medium for monitoring abnormal behavior of a patient, aiming to solve the problem that monitoring the medication of each patient by experienced relevant personnel will cause huge manpower costs.

[0005] The present invention provides a method for monitoring abnormal behavior of a slow door, comprising:

[0006] Obtaining chronic disease information of a designated user, and drug information of each reimbursement drug for the designated user based on the chronic disease;

[0007] The chronic disease information and each of the drug information are sequentially combined and input into a chronic disease correlation calculation model for calculation to obtain the correlation between each drug information and the chronic disease information; wherein the chronic disease correlation calculation model is trained based on various chronic disease information, drug information of reimbursed drugs corresponding to various chronic disease information, and the correlation between each of the drug information and the chronic disease information;

[0008] Determining whether the drug information includes drug information whose relevance is less than a first preset relevance;

[0009] If there is drug information with the correlation less than the first preset correlation, it is determined that the designated user has abnormal behavior.

[0010] Furthermore, the chronic disease correlation calculation model includes a plurality of sub-models, each of which is used to calculate the sub-correlation of each drug information with the chronic disease information, and the chronic disease information and each drug information are sequentially combined and input into the chronic disease correlation calculation model for calculation to obtain the correlation between each drug information and the chronic disease information, including:

[0011] Combining the chronic disease information and each of the drug information in sequence to obtain a plurality of combinations to be calculated;

[0012] Inputting each of the combinations to be calculated into each of the sub-models respectively, and obtaining a plurality of sub-correlations of each of the combinations to be calculated respectively;

[0013] The correlation between the corresponding drug information and the chronic disease information is calculated based on the multiple sub-correlations of each combination to be calculated.

[0014] Furthermore, the calculation of the correlation between the corresponding drug information and the chronic disease information based on the multiple sub-correlations of each combination to be calculated includes:

[0015] Obtaining a test set for detecting the loss value of each of the sub-models;

[0016] Input the test set into each of the sub-models respectively, and obtain the prediction results corresponding to each of the sub-models respectively;

[0017] Based on the actual results in the test set and each prediction result, the loss value of each sub-model is obtained by using a loss value calculation formula;

[0018] According to the formula Calculate the correlation; where cor is the correlation, Y is the actual result, and y i is the prediction result of the i-th sub-model, l(y i , Y) represents the loss value of the i-th sub-model, w i (l(y i , Y)) represents the weight of the i-th sub-model, x i represents the sub-correlation corresponding to the i-th sub-model, Ω(l(y i , Y)) represents the penalty term corresponding to the i-th sub-model.

[0019] Furthermore, the sub-model is any one of a Pearson calculation sub-model, an xgboost sub-model and a lightgbm sub-model.

[0020] Furthermore, if there is drug information with a correlation less than a first preset correlation, after determining that the designated user has abnormal behavior, the method further includes:

[0021] Count the monthly reimbursement amounts of various types of drugs in the designated area;

[0022] According to the formula Calculate the outlier value for each reimbursement amount; where Z ij represents the abnormal value corresponding to the reimbursement amount of the j-th drug in the i-th month, μ represents the average value of each month in the j-th drug, and σ represents the standard deviation of drug class j, and x ij represents the reimbursement amount of the jth category of drugs in the i-th month;

[0023] |Z ij |>Z thr The corresponding reimbursement amount is extracted and used as the abnormal behavior of reimbursement of this type of drug, where Z thr is the set parameter value.

[0024] Furthermore, after determining whether the drug information includes drug information whose relevance is less than the first preset relevance, the method further includes:

[0025] Determining whether each of the drug information contains suspicious drug information whose correlation is greater than a first preset correlation and less than a second preset correlation;

[0026] If the suspicious drug information is available, the suspicious drug information and the chronic disease information are extracted and sent to relevant personnel for identification.

[0027] Furthermore, after obtaining the chronic disease information of the designated user and the drug information of each reimbursement drug for the designated user based on the chronic disease, the method further includes:

[0028] Acquiring target drug information that can be reimbursed for the designated user based on the chronic disease information;

[0029] Determining whether the drug information of each of the reimbursed drugs contains drug information that is inconsistent with the target drug information;

[0030] If there is drug information that does not match the target drug information, it is determined that the designated user has abnormal behavior.

[0031] The present invention also provides an abnormal behavior monitoring device based on door slowness, comprising:

[0032] An acquisition module, used to acquire chronic disease information of a designated user, and drug information of each reimbursement drug for the designated user based on the chronic disease;

[0033] A combination module, used to combine the chronic disease information and each of the drug information in sequence, and input them into a chronic disease correlation calculation model for calculation, so as to obtain the correlation between each drug information and the chronic disease information; wherein the chronic disease correlation calculation model is trained based on various chronic disease information, drug information of reimbursed drugs corresponding to various chronic disease information, and the correlation between each of the drug information and the chronic disease information;

[0034] A judgment module, used for judging whether the drug information contains drug information whose correlation is less than a first preset correlation;

[0035] The determination module is used to determine that the designated user has abnormal behavior if the designated user has drug information with a correlation less than a first preset correlation.

[0036] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0037] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0038] The beneficial effects of the present invention are as follows: by obtaining the chronic disease information of the designated user and the drug information of each reimbursement drug for the designated user based on the chronic disease, the information is input into the chronic disease correlation calculation model for calculation, and the correlation between each drug information and the chronic disease information is obtained, and the abnormal behavior of the designated user is identified based on the correlation. Thus, the abnormal behavior is automatically identified, which saves the time of relevant personnel to check the drug information of each patient one by one, thereby reducing the labor cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of a method for monitoring abnormal behavior of a slow door according to an embodiment of the present invention;

[0040] Figure 2 It is a schematic block diagram of the structure of a device for monitoring abnormal behavior of a slow door according to an embodiment of the present invention;

[0041] Figure 3 A schematic block diagram of the structure of a computer device according to an embodiment of the present application.

[0042] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0043] 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.

[0044] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly. The connection can be a direct connection or an indirect connection.

[0045] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and B can represent three situations: A exists alone, A and B exist at the same time, and B exists alone.

[0046] In addition, in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0047] Reference Figure 1 The present invention proposes a method for monitoring abnormal behavior of a slow door, comprising:

[0048] S1: Obtaining chronic disease information of a designated user, and drug information of each reimbursement drug for the designated user based on the chronic disease;

[0049] S2: combining the chronic disease information and each of the drug information in sequence, and inputting them into a chronic disease correlation calculation model for calculation, to obtain the correlation between each drug information and the chronic disease information; wherein the chronic disease correlation calculation model is trained based on various chronic disease information, drug information of reimbursed drugs corresponding to various chronic disease information, and the correlation between each of the drug information and the chronic disease information;

[0050] S3: Determine whether the drug information has drug information whose relevance is less than the first preset relevance;

[0051] S4: If there is drug information with the correlation less than the first preset correlation, it is determined that the designated user has abnormal behavior.

[0052] As described in the above step S1, the chronic disease information of the designated user and the drug information of each reimbursement drug based on the chronic disease of the designated user are obtained. In this embodiment, each designated user is tested, and the designated user is any patient who suffers from a chronic disease and is reimbursed, and the patient's identity information, chronic disease information, and reimbursement information are all stored in the system. Therefore, the chronic disease information of the designated user and the drug information of the reimbursement drug based on the chronic disease can be directly obtained from the database of the system. It should be noted that the drug information of each reimbursed drug is the sum of the information of each reimbursed drug, such as raw materials, which are the raw materials of all reimbursed drugs, wherein the chronic disease information may include "age range for taking", "treatable diseases", "prohibited diseases", "patient gender", "whether it can be taken during pregnancy", "raw materials", "auxiliary materials", "ingredients and content", "applicable population", "unsuitable population", "functional efficacy", "consumption amount", "specifications", "consumption method" and other characteristics. Chronic disease information may include "symptoms", "onset population", "onset age range", "therapeutic drugs", "foods that can be eaten", "foods that are not suitable for eating", etc.

[0053] As described in the above step S2, the chronic disease information and each of the drug information are combined in sequence and input into the chronic disease correlation calculation model for calculation to obtain the correlation between each drug information and the chronic disease information. Among them, a drug information and chronic disease information are combined and input into the chronic disease correlation calculation model, wherein the calculation model is trained based on various chronic disease information and drug information corresponding to various chronic disease information. That is, only the drug information of the chronic disease information and the drug information of the drug can be input to obtain the correlation between the drug and the chronic disease information. The chronic disease correlation calculation model is a model calculated by multi-dimensional information, that is, it is obtained based on the multi-dimensional information of the drug information and the multi-dimensional information of the chronic disease information. The specific calculation method of the chronic disease correlation calculation model will be described in detail later and will not be repeated here.

[0054] As described in the above step S3, it is determined whether there is drug information whose correlation is less than the first preset correlation in each of the drug information. Since the correlation corresponding to each drug information is obtained through the chronic disease correlation calculation model, it can be compared with the first preset correlation, wherein the first preset correlation is a pre-set correlation, and if it is greater than the first preset correlation, the drug information is considered normal, and if it is less than the first preset correlation, the drug information is considered abnormal, that is, the reimbursement drug corresponding to the drug information is also abnormal.

[0055] As described in the above step S4, if there is drug information with the relevance less than the first preset relevance, it is determined that the designated user has abnormal behavior. That is, there is drug information with a relevance less than the first preset relevance. This situation may include that a certain drug information exceeds the standard, for example, drug A can also treat the chronic disease, and drug B can also treat the chronic disease. In fact, the total amount of drug A and drug B prescribed to treat the chronic disease is equivalent to the amount used by two chronic disease patients, which can be regarded as abnormal behavior. It can also be that the drug information is completely unrelated to the chronic disease information corresponding to the patient, that is, the drug information has no therapeutic effect on the chronic disease corresponding to the chronic disease information, that is, it contains reimbursed drugs that are not used to treat the chronic disease, which is also regarded as abnormal behavior. Of course, the specific determination of abnormal behavior requires the corresponding personnel to re-identify. This solution only detects abnormal behavior, saving the time of relevant personnel to check the drug information of each patient one by one, thereby reducing labor costs.

[0056] In one embodiment, the chronic disease correlation calculation model includes a plurality of sub-models, each of which is used to calculate the sub-correlation of each drug information with the chronic disease information, and the step S2 of combining the chronic disease information and each drug information in sequence and inputting them into the chronic disease correlation calculation model for calculation to obtain the correlation of each drug information with the chronic disease information includes:

[0057] S201: combining the chronic disease information and each drug information in sequence to obtain a plurality of combinations to be calculated;

[0058] S202: Inputting each of the combinations to be calculated into each of the sub-models to obtain a plurality of sub-correlations of each of the combinations to be calculated;

[0059] S203: Calculate the correlation between the corresponding drug information and the chronic disease information based on the multiple sub-correlations of each combination to be calculated.

[0060] As described in the above steps S201-S203, a comprehensive calculation of the correlation is implemented to make the obtained correlation more reasonable to reduce the error of the correlation. Specifically, the chronic disease correlation calculation model includes multiple sub-models, and each combination to be calculated is input into each sub-model respectively to obtain the sub-correlation output by each sub-model. The subsequent method of calculating the correlation of each combination to be calculated based on the sub-correlation can be to find the average value, or to assign different weights to each sub-model, and then obtain the final correlation based on the weight and sub-correlation. Among them, the method of assigning different weights for calculation can be to set the corresponding weights for each sub-model in advance, or to set them based on the loss value of each sub-model, or to take the smallest sub-correlation among all sub-correlations as the correlation. In a preferred embodiment, since each sub-model outputs sub-correlations of different dimensions, when the correlation output by any sub-correlation is low, it can be considered that the drugs reimbursed by the chronic disease patient are abnormal and have abnormal behavior. Therefore, it is preferred to regard the value of each sub-correlation as a dimension of the correlation, and the obtained correlation is a vector composed of different dimensions of each sub-correlation, so that relevant personnel can conduct another check on the patient according to the dimension of the patient's abnormal behavior. By calculating the loss value of each sub-model, automatic detection of abnormal behavior of patients in multiple dimensions is achieved, making the detection range wider and improving the comprehensiveness of reimbursement information detection of designated users of patients.

[0061] In one embodiment, the step S203 of calculating the correlation between the corresponding drug information and the chronic disease information based on the multiple sub-correlations of each combination to be calculated includes:

[0062] S2031: Obtain a test set for detecting the loss value of each sub-model;

[0063] S2032: Inputting the test set into each of the sub-models respectively, and obtaining prediction results corresponding to each of the sub-models respectively;

[0064] S2033: Based on the actual results in the test set and each prediction result, a loss value of each sub-model is obtained by using a loss value calculation formula;

[0065] S2034: According to the formula Calculate the correlation; where cor is the correlation, Y is the actual result, and y i is the prediction result of the i-th sub-model, l(y i , Y) represents the loss value of the i-th sub-model, w i (l(y i , Y)) represents the weight of the i-th sub-model, x irepresents the sub-correlation corresponding to the i-th sub-model, Ω(l(y i , Y)) represents the penalty term corresponding to the i-th sub-model.

[0066] As described in the above steps S2031-S2034, relevant weights are set for each sub-model based on the loss value to make the obtained correlation more reasonable. Specifically, a test set is obtained, wherein the test set can be composed of a training set, which is a set of training chronic disease correlation calculation models. Since the training set is used as the test set, the loss values ​​calculated by each model will have some errors. Therefore, in order to avoid obtaining loss values ​​close to the true values, it is best to use brand new data for the test set. It should be understood that the test set is also based on various chronic disease information and the drug information of the reimbursement drugs corresponding to the various chronic disease information and the corresponding correlation. According to the formula The correlation is calculated, wherein the calculation of the correlation incorporates the penalty term of each sub-model, so that the loss value obtained by each sub-model is closer to the actual value. Thus, weights are assigned to each sub-model based on the loss value and the final correlation value is calculated, making the calculated correlation value more reasonable.

[0067] In one embodiment, the submodel is any one of the Pearson calculation submodel, the xgboost submodel and the lightgbm submodel. Among them, the Pearson calculation submodel, the xgboost submodel and the lightgbm submodel are all submodels for calculating correlation. Among them, the xgboost submodel has the highest calculation accuracy, calculates each dimension, and has the widest scope of application, but its operation time is long, and it is suitable for calculating whether the reimbursement amount is abnormal. The lightgbm submodel only calculates the dimension with the largest amount of data, so its calculation accuracy is low and it is easy to produce overfitting, so it is suitable for calculating whether there is an excess of a certain drug in the drug information. The Pearson calculation submodel is only suitable for calculations where the two have a linear relationship, for example, it is suitable for some dimensions that require refined calculations, such as the correlation between the patient's chronic disease information and its corresponding drug information.

[0068] In one embodiment, after the step S4 of determining that the designated user has abnormal behavior if there is drug information with the correlation less than the first preset correlation, the method further includes:

[0069] S501: Count the monthly reimbursement amounts of various types of drugs in a designated area;

[0070] S502: According to the formula Calculate the outlier value for each reimbursement amount; where Z ijrepresents the abnormal value corresponding to the reimbursement amount of the j-th drug in the i-th month, μ represents the average value of each month in the j-th drug, and σ represents the standard deviation of drug class j, and x ij represents the reimbursement amount of the jth category of drugs in the i-th month;

[0071] S503: |Z ij |>Z thr The corresponding reimbursement amount is extracted and used as the abnormal behavior of reimbursement of this type of drug, where Z thr is the set parameter value.

[0072] As described in the above steps S501-S503, the abnormal situation is analyzed from the overall reimbursement amount. After determining that the designated user has abnormal behavior, it is necessary to further analyze the abnormal behavior of the user to determine which abnormal situation it belongs to. In this embodiment, the analysis is performed from the overall reimbursement amount. That is, the monthly reimbursement amount of each type of medicine in the designated area is counted. The amount has a relevant statistical value in the system, so it can be directly obtained. It is necessary to calculate the abnormal value of the reimbursement amount of each month obtained. It should be noted that each data has a certain abnormal value. If the amount of the reimbursement amount of a certain month is completely correct, it is considered that its abnormal value is 0. When the abnormal value exceeds the set parameter value, it can be considered that the performance indicator data is abnormal. Of course, it is not ruled out that the reimbursement amount of a certain month may have an abnormal value. For example, a certain chronic disease is more frequent in a certain month. It is necessary to verify and process this data and compare the data with the actual situation. This embodiment only extracts the abnormal situation, and the subsequent analysis still needs to be processed manually.

[0073] In one embodiment, after the step S3 of determining whether the drug information has drug information with a relevance less than the first preset relevance, the method further includes:

[0074] S401: Determine whether there is any suspicious drug information in each of the drug information whose correlation is greater than a first preset correlation and less than a second preset correlation;

[0075] S402: If the suspicious drug information is available, extract the suspicious drug information and the chronic disease information and send them to relevant personnel for identification.

[0076] As described in the above steps S401-S402, the detection of suspicious drug information is realized, that is, the actual situation is that the designated user has over-reported drugs, but the over-reported amount is not large, or the designated user's condition is more serious and needs to increase the dosage. These two situations cannot be judged, so it is necessary to extract this part of drug information and conduct a second identification. The second preset correlation is a pre-set threshold, and drug information between the first preset correlation and the second preset correlation is regarded as suspicious drug information and needs to be re-identified. In this way, the situation where the over-reported amount of reimbursement drugs is not large is detected.

[0077] In one embodiment, after the step S1 of acquiring the chronic disease information of the designated user and the drug information of each reimbursement drug for the designated user based on the chronic disease, the step further includes:

[0078] S211: Acquire target drug information that can be reimbursed for the designated user based on the chronic disease information;

[0079] S212: Determine whether the drug information of each reimbursable drug contains drug information that is inconsistent with the target drug information;

[0080] S213: If there is drug information that does not match the target drug information, the drug corresponding to the drug information is regarded as an abnormal behavior of the designated user.

[0081] As described in the above steps S211-S213, direct detection of abnormal information that does not conform to drug information is achieved, that is, by obtaining the target drug information that can be reimbursed by the designated user based on the chronic disease information, wherein the drugs for treating each chronic disease are existing and can be directly obtained, and the corresponding therapeutic drugs can be obtained according to the chronic disease information of the designated user, and then the drug information of each therapeutic drug can be obtained, and then it is determined whether the drug information of each reimbursed drug has drug information that does not conform to the target drug information, that is, the inconsistent drug information contains the situation that it is not for treating the chronic disease of the designated user, so it is necessary to extract and analyze this part of the content. If there is drug information that does not conform to the target drug information, the drug corresponding to the drug information is regarded as abnormal behavior of the designated user. The abnormal behavior will be further identified later. Thereby, direct identification of drug information is achieved.

[0082] Reference Figure 2 The present invention also provides an abnormal behavior monitoring device based on door slowness, comprising:

[0083] The acquisition module 10 is used to acquire the chronic disease information of the designated user and the drug information of each reimbursement drug based on the chronic disease of the designated user;

[0084] The combination module 20 is used to sequentially combine the chronic disease information and each of the drug information, and input them into a chronic disease correlation calculation model for calculation to obtain the correlation between each drug information and the chronic disease information; wherein the chronic disease correlation calculation model is trained based on various chronic disease information, drug information of reimbursed drugs corresponding to various chronic disease information, and the correlation between each of the drug information and the chronic disease information;

[0085] A judgment module 30, used for judging whether the drug information has drug information whose correlation is less than a first preset correlation;

[0086] The determination module 40 is configured to determine that the designated user has abnormal behavior if the designated user has drug information with the correlation less than a first preset correlation.

[0087] In one embodiment, the chronic disease correlation calculation model includes multiple sub-models, each of which is used to calculate the sub-correlation between each of the drug information and the chronic disease information, and the combination module 20 includes:

[0088] A drug information combination submodule, used to sequentially combine the chronic disease information and each drug information to obtain a plurality of combinations to be calculated;

[0089] A submodule for inputting combinations to be calculated, used for inputting each of the combinations to be calculated into each of the sub-models, respectively, to obtain a plurality of sub-correlations of each of the combinations to be calculated;

[0090] The correlation calculation submodule is used to calculate the correlation between the corresponding drug information and the chronic disease information based on the multiple sub-correlations of each combination to be calculated.

[0091] In one embodiment, the relevance calculation submodule includes:

[0092] A test set acquisition unit, used to acquire a test set for detecting the loss value of each sub-model;

[0093] A test set input unit, used to input the test set into each of the sub-models to obtain prediction results corresponding to each of the sub-models;

[0094] A loss value calculation unit, used to obtain the loss value of each sub-model through a loss value calculation formula based on the actual results in the test set and each prediction result;

[0095] Correlation calculation unit, used to calculate the correlation according to the formula Calculate the correlation; where cor is the correlation, Y is the actual result, and y i is the prediction result of the i-th sub-model, l(yi , Y) represents the loss value of the i-th sub-model, w i (l(y i , Y)) represents the weight of the i-th sub-model, x i represents the sub-correlation corresponding to the i-th sub-model, Ω(l(y i , Y)) represents the penalty term corresponding to the i-th sub-model.

[0096] In one embodiment, the sub-model is any one of a Pearson calculation sub-model, an xgboost sub-model and a lightgbm sub-model.

[0097] In one embodiment, the abnormal behavior monitoring device based on the slow door also includes:

[0098] The reimbursement amount statistics module is used to count the monthly reimbursement amounts of various types of drugs in a specified area;

[0099] The outlier calculation module is used to calculate the outlier value according to the formula Calculate the outlier value for each reimbursement amount; where Z ij represents the abnormal value corresponding to the reimbursement amount of the j-th drug in the i-th month, μ represents the average value of each month in the j-th drug, and σ represents the standard deviation of drug class j, and x ij represents the reimbursement amount of the jth category of drugs in the i-th month;

[0100] Reimbursement amount extraction module, used to convert |Z ij |>Z thr The corresponding reimbursement amount is extracted and used as the abnormal behavior of reimbursement of this type of drug, where Z thr is the set parameter value.

[0101] In one embodiment, the abnormal behavior monitoring device based on the slow door also includes:

[0102] A suspicious drug information judging module, used to judge whether there is any suspicious drug information in the drug information whose correlation is greater than a first preset correlation and less than a second preset correlation;

[0103] The suspicious drug information sending module is used to extract the suspicious drug information and the chronic disease information and send them to relevant personnel for identification if the suspicious drug information is available.

[0104] In one embodiment, the abnormal behavior monitoring device based on the slow door also includes:

[0105] A target drug information acquisition module, used to acquire target drug information that can be reimbursed for the designated user based on the chronic disease information;

[0106] A target drug information determination module is used to determine whether the drug information of each reimbursable drug contains drug information that is inconsistent with the target drug information;

[0107] The abnormal behavior judgment module is used to determine that the designated user has abnormal behavior if there is drug information that does not match the target drug information.

[0108] The beneficial effects of the present invention are as follows: by obtaining the chronic disease information of the designated user and the drug information of each reimbursement drug for the designated user based on the chronic disease, the information is input into the chronic disease correlation calculation model for calculation, and the correlation between each drug information and the chronic disease information is obtained, and the abnormal behavior of the designated user is identified based on the correlation. Thus, the abnormal behavior is automatically identified, which saves the time of relevant personnel to check the drug information of each patient one by one, thereby reducing the labor cost.

[0109] Reference Figure 3 In the embodiment of the present application, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store various chronic disease information, etc. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the abnormal behavior monitoring method of the slow door described in any of the above embodiments can be implemented.

[0110] Those skilled in the art will understand that Figure 3 The structure shown in is merely a block diagram of a portion of the structure related to the present application solution and does not constitute a limitation on the computer device to which the present application solution is applied.

[0111] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0112] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0113] An embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the abnormal behavior monitoring method of the door slowness described in any of the above embodiments can be implemented.

[0114] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0115] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method 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, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, device, article or method including the element.

[0116] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the scope of the claims of the present invention.

Claims

1. A method for monitoring abnormal behavior of slow doors, It is characterized in that include: Obtaining chronic disease information of a designated user, and drug information of each reimbursement drug for the designated user based on the chronic disease; The chronic disease information and each of the drug information are sequentially combined and input into a chronic disease correlation calculation model for calculation to obtain the correlation between each drug information and the chronic disease information; wherein the chronic disease correlation calculation model is trained based on various chronic disease information, drug information of reimbursed drugs corresponding to various chronic disease information, and the correlation between each of the drug information and the chronic disease information; Determining whether the drug information includes drug information whose relevance is less than a first preset relevance; If there is drug information with the relevance less than the first preset relevance, it is determined that the designated user has abnormal behavior; After determining whether the drug information has drug information whose correlation is less than the first preset correlation, the method further includes: Determining whether each of the drug information contains suspicious drug information whose correlation is greater than a first preset correlation and less than a second preset correlation; If the suspicious drug information is available, extract the suspicious drug information and the chronic disease information and send them to relevant personnel for identification; The second preset correlation is a preset threshold, and drug information between the first preset correlation and the second preset correlation is regarded as suspicious drug information; The chronic disease correlation calculation model includes a plurality of sub-models, each of which is used to calculate the sub-correlation of each drug information with the chronic disease information. The chronic disease information and each drug information are sequentially combined and input into the chronic disease correlation calculation model for calculation, and the correlation between each drug information and the chronic disease information is obtained, including: Combining the chronic disease information and each of the drug information in sequence to obtain a plurality of combinations to be calculated; Inputting each of the combinations to be calculated into each of the sub-models respectively, and obtaining a plurality of sub-correlations of each of the combinations to be calculated respectively; Based on the multiple sub-correlations of each combination to be calculated, calculating the correlation between the corresponding drug information and the chronic disease information; The calculating the correlation between the corresponding drug information and the chronic disease information based on the plurality of sub-correlations of each combination to be calculated comprises: Obtaining a test set for detecting the loss value of each of the sub-models; Input the test set into each of the sub-models respectively, and obtain the prediction results corresponding to each of the sub-models respectively; Based on the actual results in the test set and each prediction result, the loss value of each sub-model is obtained by using a loss value calculation formula; According to the formula Calculate the correlation; where cor is the correlation, Y is the actual result, and y i is the prediction result of the i-th sub-model, l(y i , Y) represents the loss value of the i-th sub-model, w i (l(y i , Y)) represents the weight of the i-th sub-model, x i represents the sub-correlation corresponding to the i-th sub-model, Ω(l(y i , Y)) represents the penalty term corresponding to the i-th sub-model.

2. The abnormal behavior monitoring method of the door slowness as claimed in claim 1, It is characterized in that The sub-model is any one of the Pearson calculation sub-model, the xgboost sub-model and the lightgbm sub-model.

3. The abnormal behavior monitoring method of the door slowness according to claim 1, It is characterized in that If the drug information has a correlation less than the first preset correlation, after determining that the designated user has abnormal behavior, the method further includes: Count the monthly reimbursement amounts of various types of drugs in the designated area; According to the formula Calculate the outlier value for each reimbursement amount; where Z ij represents the abnormal value corresponding to the reimbursement amount of the j-th drug in the i-th month, μ represents the average value of each month in the j-th drug, and σ represents the standard deviation of drug class j, and x ij represents the reimbursement amount of the jth category of drugs in the i-th month; |Z ij |>Z thr The corresponding reimbursement amount is extracted and used as the abnormal behavior of reimbursement of this type of drug, where Z thr is the set parameter value.

4. An abnormal behavior monitoring device based on door slowness, It is characterized in that The abnormal behavior monitoring method for implementing any one of claims 1 to 3 comprises: An acquisition module, used to acquire chronic disease information of a designated user, and drug information of each reimbursement drug for the designated user based on the chronic disease; A combination module, used to combine the chronic disease information and each of the drug information in sequence, and input them into a chronic disease correlation calculation model for calculation, so as to obtain the correlation between each drug information and the chronic disease information; wherein the chronic disease correlation calculation model is trained based on various chronic disease information, drug information of reimbursed drugs corresponding to various chronic disease information, and the correlation between each of the drug information and the chronic disease information; A judgment module, used for judging whether the drug information contains drug information whose correlation is less than a first preset correlation; A determination module, configured to determine that the designated user has abnormal behavior if the designated user has drug information with the correlation less than a first preset correlation; A suspicious drug information judging module, used to judge whether there is any suspicious drug information in the drug information whose correlation is greater than a first preset correlation and less than a second preset correlation; A suspicious drug information sending module is used to extract and send the suspicious drug information and the chronic disease information to relevant personnel for identification if the suspicious drug information is available; The second preset correlation is a preset threshold, and drug information between the first preset correlation and the second preset correlation is regarded as suspicious drug information.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, It is characterized in that When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images