Medical insurance abnormal use identification method and device, equipment and storage medium

Through the preset large language model and Gaussian mixed model to process medical insurance consumption data, the problems of low efficiency and low accuracy of abnormal use recognition of medical insurance are solved, and rapid and accurate abnormal identification is achieved, which improves the safety of medical insurance funds and institutional fairness. .

CN119089352APending Publication Date: 2024-12-06XIAN BEILIN PEKING BIRD INTERNET HOSPITAL CO LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202411145248.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing technology is inefficient and low in accuracy in identifying abnormal use of medical insurance, mainly due to the inconsistent reliance on manual review and professional knowledge, which has affected the safety and institutional fairness of medical insurance funds.

Method used

The preset large language model is used to process medical insurance consumption data, extract target data relationships, and calculate the clustering center through the Gaussian hybrid model to determine the abnormal usage results of medical insurance users, achieving fast and accurate abnormal recognition.

Benefits of technology

Through the combination of large language model and Gaussian hybrid model, abnormal usage of medical insurance users can be quickly and accurately identified, improving identification efficiency and accuracy, and helping relevant departments to discover and detect non-compliant medical insurance use in advance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089352A_ABST
    Figure CN119089352A_ABST
Patent Text Reader

Abstract

The invention discloses a medical insurance abnormal use identification method and device, equipment and a storage medium, and the method comprises the steps: processing a plurality of pieces of target medical insurance consumption data through a preset big language model, can rapidly and accurately extract a target data relation, carrying out the clustering processing of each average cost and the number of times of remote medical treatment through a preset Gaussian mixture model, and carrying out the recognition of the abnormal use of medical insurance. Finally, the abnormal use result of the target medical insurance consumption data of each medical insurance user can be accurately determined based on the clustering center, so that the real-time data of the medical insurance user using the medical insurance can be calculated and judged, and whether the situation of abnormal medical insurance use exists or not can be quickly and accurately judged. And related departments are helped to find and detect medical insurance non-compliance use in advance.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to methods, devices, equipment and storage media for identifying abnormal use of medical insurance. Background Art

[0002] With the popularization of the medical insurance system, abnormal use of medical insurance has become increasingly serious, affecting the security of medical insurance funds and the fairness of the system. In order to effectively identify and warn of these behaviors, an advanced technical means is urgently needed.

[0003] At present, the identification of abnormal use of medical insurance is mainly carried out through manual review. However, due to the limited number of reviewers, the efficiency of identifying abnormal use of medical insurance is low when the number of medical insurance policies is large. In addition, it is necessary to rely on the professional knowledge of reviewers and there are problems with inconsistent review standards, which leads to low accuracy in identifying abnormal use of medical insurance. Summary of the invention

[0004] The present application aims to at least solve the technical problems existing in the prior art. To this end, the first aspect of the present application proposes a method for identifying abnormal use of medical insurance, the method comprising:

[0005] Obtain target medical insurance consumption data of each medical insurance user; wherein each target medical insurance consumption data includes medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content;

[0006] The target medical insurance consumption data is processed by a preset large language model to obtain the target data relationship; wherein the target data relationship is the correspondence between various types of diseases, various medical insurance users, average expenditure and the number of out-of-town medical treatments;

[0007] For each average cost and each number of out-of-town medical treatment visits, the preset Gaussian mixture model is used to calculate the target cluster center corresponding to the average cost and the number of out-of-town medical treatment visits;

[0008] Based on each target cluster center, determine the abnormal usage results of the target medical insurance consumption data of each medical insurance user.

[0009] In a possible implementation, the medical insurance recipient information includes the medical insurance recipient's identity information and the medical insurance payment location information. The target medical insurance consumption data is processed by a preset large language model to obtain a target data relationship, including:

[0010] For each diagnosis content, a preset large language model is used to classify the diagnosis content to obtain multiple types of diseases;

[0011] Obtain detailed disease classification data of each medical insurance user when using medical insurance; wherein, detailed disease classification data includes the corresponding relationship between medical insurance user identity information, disease type, medical insurance expenditure information, medical insurance payment location information and medical insurance consumption location information;

[0012] The detailed disease classification data of each medical insurance user when using medical insurance is statistically processed to obtain the target data relationship.

[0013] In a possible implementation, for each diagnosis content, a preset large language model is used to classify the diagnosis content to obtain multiple types of diseases, including:

[0014] For each diagnostic content, the preset large language model uses a preset input template to classify the diagnostic content to obtain multiple types of diseases; wherein the preset input template includes language instructions for analyzing the diagnostic content, a list of all types of diseases and an answer format.

[0015] In a possible implementation, for each average cost and each number of times of out-of-town medical treatment, a preset Gaussian mixture model is used to calculate the target cluster center corresponding to the average cost and the number of times of out-of-town medical treatment, including:

[0016] Get the number and initial parameters of Gaussian distribution corresponding to the preset Gaussian mixture model;

[0017] For each average cost and number of out-of-town medical treatment, calculate the posterior probability that the average cost and number of out-of-town medical treatment belong to each Gaussian distribution;

[0018] Based on each posterior probability, the parameters of the Gaussian distribution are updated to obtain the updated parameters of the Gaussian distribution; wherein the parameters of the Gaussian distribution include a mixing coefficient, a mean, and a covariance matrix;

[0019] The initial parameters are updated based on the updated parameters of the Gaussian distribution to obtain updated initial parameters;

[0020] If the updated initial parameters meet the preset conditions, the target cluster center is determined based on the parameters of the updated Gaussian distribution; if the updated initial parameters do not meet the preset conditions, the step of updating the initial parameters is repeated until the updated initial parameters meet the preset conditions.

[0021] In a possible implementation, based on each target cluster center, determining the abnormal usage result of the target medical insurance consumption data of each medical insurance user includes:

[0022] Calculate the Euclidean distance between the average cost, the number of out-of-town medical treatments and the corresponding target cluster center to obtain the Euclidean distance value;

[0023] Based on the Euclidean distance value, the target clusters of average expenditure and number of out-of-town medical treatment are determined;

[0024] Calculate the cosine similarity between the average cost, the number of out-of-town medical treatments, and the target cluster center of the target cluster to obtain the cosine similarity value;

[0025] Based on the cosine similarity value, the abnormal usage results of the target medical insurance consumption data of each medical insurance user are determined.

[0026] In a possible implementation, obtaining target medical insurance consumption data of each medical insurance user includes:

[0027] Obtain the current medical insurance consumption data of each medical insurance user from the preset database;

[0028] Performing data cleaning on the current medical insurance consumption data to obtain cleaned medical insurance consumption data; wherein the data cleaning process includes using preset data verification rules to verify the data range and data type of the current medical insurance consumption data;

[0029] The cleaned medical insurance consumption data is subjected to data standardization processing to obtain the target medical insurance consumption data of each medical insurance user.

[0030] In a possible implementation, the method further includes:

[0031] After the target user's identity authentication is passed, the target interface is called through the access token to collect the original medical insurance consumption data of each medical insurance user in real time; wherein, the original medical insurance consumption data is obtained after being encrypted using a preset encryption protection protocol;

[0032] The original medical insurance consumption data is processed to obtain the current medical insurance consumption data; the data processing operation includes data cleaning, data standardization, data deduplication and merging.

[0033] The second aspect of the present application provides a device for identifying abnormal use of medical insurance, the device comprising:

[0034] The acquisition module is used to acquire the target medical insurance consumption data of each medical insurance user; wherein each target medical insurance consumption data includes medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content;

[0035] A processing module is used to process the target medical insurance consumption data through a preset large language model to obtain a target data relationship; wherein the target data relationship is a correspondence between various types of diseases, various medical insurance users, average expenditures and the number of times of medical treatment in other places;

[0036] A calculation module is used to calculate the target cluster center corresponding to the average cost and the number of out-of-town medical treatments using a preset Gaussian mixture model for each average cost and each number of out-of-town medical treatments;

[0037] The determination module is used to determine the abnormal usage results of the target medical insurance consumption data of each medical insurance user based on each target cluster center.

[0038] In a possible implementation, the medical insurance recipient information includes the medical insurance recipient's identity information and the medical insurance payment location information, and the above processing module is specifically used to:

[0039] For each diagnosis content, a preset large language model is used to classify the diagnosis content to obtain multiple types of diseases;

[0040] Obtain detailed disease classification data of each medical insurance user when using medical insurance; wherein, detailed disease classification data includes the corresponding relationship between medical insurance user identity information, disease type, medical insurance expenditure information, medical insurance payment location information and medical insurance consumption location information;

[0041] The detailed disease classification data of each medical insurance user when using medical insurance is statistically processed to obtain the target data relationship.

[0042] In a possible implementation manner, the processing module is further used to:

[0043] For each diagnostic content, the preset large language model uses a preset input template to classify the diagnostic content to obtain multiple types of diseases; wherein the preset input template includes language instructions for analyzing the diagnostic content, a list of all types of diseases and an answer format.

[0044] In a possible implementation manner, the above-mentioned calculation module is specifically used for:

[0045] Get the number and initial parameters of Gaussian distribution corresponding to the preset Gaussian mixture model;

[0046] For each average cost and number of out-of-town medical treatment, calculate the posterior probability that the average cost and number of out-of-town medical treatment belong to each Gaussian distribution;

[0047] Based on each posterior probability, the parameters of the Gaussian distribution are updated to obtain the updated parameters of the Gaussian distribution; wherein the parameters of the Gaussian distribution include a mixing coefficient, a mean, and a covariance matrix;

[0048] The initial parameters are updated based on the updated parameters of the Gaussian distribution to obtain updated initial parameters;

[0049] If the updated initial parameters meet the preset conditions, the target cluster center is determined based on the parameters of the updated Gaussian distribution; if the updated initial parameters do not meet the preset conditions, the step of updating the initial parameters is repeated until the updated initial parameters meet the preset conditions.

[0050] In a possible implementation manner, the above-mentioned determination module is specifically used to:

[0051] Calculate the Euclidean distance between the average cost, the number of out-of-town medical treatments and the corresponding target cluster center to obtain the Euclidean distance value;

[0052] Based on the Euclidean distance value, the target clusters of average expenditure and number of out-of-town medical treatment are determined;

[0053] Calculate the cosine similarity between the average cost, the number of out-of-town medical treatments, and the target cluster center of the target cluster to obtain the cosine similarity value;

[0054] Based on the cosine similarity value, the abnormal usage results of the target medical insurance consumption data of each medical insurance user are determined.

[0055] In a possible implementation manner, the acquisition module is specifically used to:

[0056] Obtain the current medical insurance consumption data of each medical insurance user from the preset database;

[0057] Performing data cleaning on the current medical insurance consumption data to obtain cleaned medical insurance consumption data; wherein the data cleaning process includes using preset data verification rules to verify the data range and data type of the current medical insurance consumption data;

[0058] The cleaned medical insurance consumption data is subjected to data standardization processing to obtain the target medical insurance consumption data of each medical insurance user.

[0059] In a possible implementation manner, the acquisition module is further used to:

[0060] After the target user's identity authentication is passed, the target interface is called through the access token to collect the original medical insurance consumption data of each medical insurance user in real time; wherein, the original medical insurance consumption data is obtained after being encrypted using a preset encryption protection protocol;

[0061] The original medical insurance consumption data is processed to obtain the current medical insurance consumption data; the data processing operation includes data cleaning, data standardization, data deduplication and merging.

[0062] The third aspect of the present application proposes an electronic device, which includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the method for identifying abnormal use of medical insurance as described in the first aspect.

[0063] The fourth aspect of the present application proposes a computer-readable storage medium, which stores at least one instruction, at least one program, a code set or an instruction set. The at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by a processor to implement the method for identifying abnormal use of medical insurance as described in the first aspect.

[0064] The embodiments of the present application have the following beneficial effects:

[0065] The embodiment of the present application provides a method for identifying abnormal use of medical insurance, the method comprising: obtaining target medical insurance consumption data of each medical insurance user, wherein each target medical insurance consumption data comprises medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content, processing the target medical insurance consumption data through a preset large language model to obtain a target data relationship, wherein the target data relationship is a correspondence between various types of diseases, each medical insurance user, average expenditure and the number of out-of-town medical treatments, for each average expenditure and each out-of-town medical treatment number, using a preset Gaussian mixture model to calculate the target cluster center corresponding to the average expenditure and the number of out-of-town medical treatments, and based on each target cluster center, determining the abnormal use result of the target medical insurance consumption data of each medical insurance user. This solution processes multiple target medical insurance consumption data through a preset large language model, which can quickly and accurately extract the target data relationship, and then cluster the average expenditure and the number of out-of-town medical treatment through a preset Gaussian mixture model. Finally, based on the cluster center, the abnormal usage results of the target medical insurance consumption data of each medical insurance user can be accurately determined. It can not only calculate and judge the real-time data of medical insurance users' use of medical insurance, and quickly and accurately determine whether there is an abnormal use of medical insurance, but also help relevant departments to discover and detect non-compliant use of medical insurance in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A block diagram of a computer device provided in an embodiment of the present application;

[0067] Figure 2 A flowchart of a method for identifying abnormal use of medical insurance provided in an embodiment of the present application;

[0068] Figure 3 A flowchart of the steps for obtaining target medical insurance consumption data provided in an embodiment of the present application;

[0069] Figure 4 A flowchart of the steps for obtaining current medical insurance consumption data provided in an embodiment of the present application;

[0070] Figure 5 A flowchart of steps for obtaining a target data relationship provided in an embodiment of the present application;

[0071] Figure 6A flowchart of the steps for calculating the target cluster center provided in an embodiment of the present application;

[0072] Figure 7 A flowchart of the steps for determining abnormal usage results provided in an embodiment of the present application;

[0073] Figure 8 A structural block diagram of a device for identifying abnormal use of medical insurance provided in an embodiment of the present application. DETAILED DESCRIPTION

[0074] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0075] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present disclosure, unless otherwise specified, "multiple" means two or more. In addition, the use of "based on" or "according to" means openness and inclusiveness, because the process, steps, calculations or other actions "based on" or "according to" one or more of the conditions or values ​​may be based on additional conditions or values ​​beyond the described values ​​in practice.

[0076] The abnormal medical insurance usage identification method provided in the present application can be applied to computer devices (electronic devices). The computer device can be a server or a terminal. The server can be a single server or a server cluster composed of multiple servers. The embodiments of the present application do not make specific limitations on this. The terminal can be but is not limited to various personal computers, laptops, smart phones, tablet computers and portable wearable devices.

[0077] Take the computer device as a server as an example. Figure 1 A block diagram of a server is shown, such as Figure 1 As shown, the server may include a processor and a memory connected via a system bus. The processor of the server is used to provide computing and control capabilities. The memory of the server 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 internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. When the computer program is executed by the processor, a method for identifying abnormal use of medical insurance is implemented.

[0078] Those skilled in the art will understand that Figure 1 The structure shown in the figure is only a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the server to which the solution of the present application is applied. Optionally, the server may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0079] It should be noted that the execution subject of the embodiments of the present application may be a computer device or a device for identifying abnormal use of medical insurance. The following method embodiments will be described using a computer device as the execution subject.

[0080] Figure 2 A flowchart of a method for identifying abnormal use of medical insurance provided in an embodiment of the present application. Figure 2 As shown, the method comprises the following steps:

[0081] Step 202: Obtain target medical insurance consumption data of each medical insurance user.

[0082] Among them, each target medical insurance consumption data may include medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content, and may also include other types of data, which is not specifically limited in the embodiments of the present application.

[0083] In some optional embodiments, when obtaining the target medical insurance consumption data of each medical insurance user, such as Figure 3 As shown, Figure 3 A flowchart of steps for obtaining target medical insurance consumption data provided in an embodiment of the present application includes:

[0084] Step 302: Obtain the current medical insurance consumption data of each medical insurance user from a preset database.

[0085] Among them, when obtaining the current medical insurance consumption data of each medical insurance user, such as Figure 4 As shown, Figure 4 A flowchart of steps for obtaining current medical insurance consumption data provided in an embodiment of the present application includes:

[0086] Step 402: After the target user's identity authentication is passed, the target interface is called through the access token to collect the original medical insurance consumption data of each medical insurance user in real time.

[0087] Step 404: Perform data processing operations on the original medical insurance consumption data to obtain current medical insurance consumption data.

[0088] When obtaining original medical insurance data, it is possible to connect to provincial medical insurance management systems to obtain users' medical insurance consumption data, including registration information, diagnosis information, medicines and treatment costs, etc. At the same time, it is also possible to obtain global data from the central medical insurance system to ensure the comprehensiveness and consistency of the data.

[0089] Optionally, a target interface can be pre-developed to connect to the provincial medical insurance systems. The target interface can be designed using a RESTful architecture to ensure the simplicity, scalability, and maintainability of the interface. In addition, the target interface can pre-define multiple functions such as data acquisition, data submission, and identity authentication.

[0090] Next, the OAuth 2.0 protocol can be used to authenticate the target user to ensure the legitimacy and security of data collection. After the target user's identity is authenticated, the authorization code is used to request an access token from the authentication server, and then the target interface is called through the access token to collect the original medical insurance consumption data of each medical insurance user in real time.

[0091] When collecting the original medical insurance consumption data of each medical insurance user, it can be collected in batches and in real time to meet different business needs. In addition, the original medical insurance consumption data is obtained after being encrypted using a preset encryption protection protocol, and the preset encryption protection protocol can be an SSL / TLS protocol. In other words, during data transmission, storage, etc., the SSL / TLS protocol can be used for encryption to prevent the data from being stolen or tampered with. In some optional embodiments, when data is transmitted, the data transmission can be in JSON format to ensure the readability and ease of parsing of the data.

[0092] After obtaining the original medical insurance consumption data, the original medical insurance consumption data can be stored in a preset database. The preset database can be a Hadoop HDFS database or a NoSQL database, so that massive data can be stored to ensure high availability and scalability of the data.

[0093] Next, the original medical insurance consumption data can be processed to obtain the current medical insurance consumption data. Among them, data processing operations can include data cleaning, data standardization, data deduplication and merging. Data cleaning can include missing value filling, outlier processing and data format conversion to ensure data quality and consistency. Data standardization can standardize data from different sources, unify data formats and field naming, and facilitate subsequent analysis and modeling. Data deduplication and merging can deduplicate duplicate data and merge multi-source data to ensure data integrity and accuracy.

[0094] Step 304: Perform data cleaning on the current medical insurance consumption data to obtain cleaned medical insurance consumption data.

[0095] After obtaining the current medical insurance consumption data, the current medical insurance consumption data may be cleaned. Optionally, the data cleaning process may include using preset data validation rules to validate the data range and data type of the current medical insurance consumption data. For illegal or inconsistent data, the data may be repaired or deleted to maintain data consistency and reliability.

[0096] Step 306: Perform data standardization processing on the cleaned medical insurance consumption data to obtain target medical insurance consumption data for each medical insurance user.

[0097] Among them, data standardization processing may include normalization processing, which is a preprocessing step for adjusting the data scale to ensure that different features have the same importance in model training. Therefore, all cleaned medical insurance consumption data can be normalized. For example, the minimum-maximum normalization algorithm can be used for normalization processing, and the calculation formula is shown in formula (1).

[0098]

[0099] Among them, x represents the cleaned medical insurance consumption data; min(x) represents the minimum value of all cleaned medical insurance consumption data; max(x) represents the maximum value of all cleaned medical insurance consumption data; x′ represents the normalized data, that is, the target medical insurance consumption data.

[0100] Step 204: Process the target medical insurance consumption data through a preset large language model to obtain a target data relationship.

[0101] The target data relationship is the correspondence between various types of diseases, various medical insurance users, average expenses and the number of out-of-town medical treatments. Based on this target data relationship, it can be determined whether the medical insurance user has committed insurance fraud or other behaviors. For example, if the same medical insurance user has repeatedly visited out-of-town doctors for the same type of disease, there is a high probability that insurance fraud has occurred.

[0102] Since the target medical insurance consumption data includes the diagnosis content field, which is non-standardized content and is also the specific diagnosis made by the physician based on the patient's condition, it is necessary to first standardize and map this field through a preset large language model so that the data can be classified in detail during subsequent modeling.

[0103] Exemplarily, as shown in Table 1, Table 1 is a partial example of diagnosis content in the target medical insurance consumption data. The ID number is used for the identity information of the medical insurance user and can be obtained from the medical insurance person information.

[0104] Table 1

[0105] ID number Diagnosis content xxxx Male, 25 years old, insomnia... xxxx Female, 49 years old, back pain... xxxx Male, 67 years old, dizziness and headache... xxxx Male, 37 years old, insomnia and dreaminess...... xxxx Female, 59 years old, thigh fracture... xxxx Male, 9 years old, diarrhea...... ……

[0106] In some optional embodiments, Figure 5 As shown, Figure 5 A flowchart of steps for obtaining a target data relationship provided in an embodiment of the present application includes:

[0107] Step 502: For each diagnosis content, a preset large language model is used to classify the diagnosis content to obtain multiple types of diseases.

[0108] Step 504: Obtain detailed disease classification data of each medical insurance user when using medical insurance.

[0109] Step 506: Statistically process the detailed disease classification data of each medical insurance user when using the medical insurance to obtain the target data relationship.

[0110] Among them, the preset large language model can be the Phi-3 SLM open source large language model, so that the model can be used to classify the diagnosis content and obtain multiple types of diseases. Due to the wide variety of diseases, in order to classify the diagnosis content, it is necessary to pre-set the classification of diseases. Exemplarily, diseases can be divided into t categories and M can be used as the classification of diseases, that is, a list of all types of diseases, so that M = {m1, m2, ..., m t},m t Indicates the classification name of each type of disease.

[0111] Optionally, for each diagnosis content, a large language model can be preset to classify the diagnosis content using a preset input template to obtain multiple types of diseases. The preset input template includes language instructions for analyzing the diagnosis content, a list of all types of diseases, and an answer format.

[0112] The analysis language instructions for the diagnosis content and the list of all types of diseases can constitute the prompts input into the model. For example, the prompt can be: "You are now a diagnostic physician. I now need you to analyze the diagnosis content of other physicians and give the disease classification corresponding to the diagnosis content. I need you to divide the physician's diagnosis content into the following categories. The diseases included in this category are {disease_list}. Please answer me the content of the diseases in this category. The answer format is as follows: "Among them, {disease_list} is a reserved placeholder. In actual operation, its content needs to be replaced with a pre-established disease classification M. The answer format can be: "Disease content: XXX".

[0113] Next, it is also necessary to obtain detailed disease classification data of each medical insurance user when using medical insurance. The detailed disease classification data may include the corresponding relationship between the medical insurance person's identity information, disease type, medical insurance expenditure information, medical insurance payment location information and medical insurance consumption location information.

[0114] For example, as shown in Table 2, Table 2 is a partial example of detailed disease classification data. It should be noted that the ID number corresponds to the medical insurance person's identity information, the disease corresponds to the disease type, the cost corresponds to the medical insurance cost information, the medical insurance payment city corresponds to the medical insurance payment location information, and the medical insurance consumption city corresponds to the medical insurance consumption location information. In addition, the cost field is mapped to a decimal between 0 and 1 after normalization.

[0115] Table 2

[0116] ID number disease spend Medical insurance payment city Medical insurance consumption cities xxxx heart disease 0.8 Beijing Beijing xxxx heart disease 0.5 Shanghai Shanghai xxxx diabetes 0.13 Guangzhou Huizhou xxxx hypertension 0.34 Shenzhen Shenzhen xxxx hypertension 0.83 Wuhan Jingzhou xxxx heart disease 0.93 Zhengzhou Luoyang …… …… …… …… ……

[0117] Thus, the detailed disease classification data of each medical insurance user when using medical insurance can be statistically processed to obtain the target data relationship. Optionally, the "detailed disease classification data" in Table 2 above can be summarized and calculated to obtain the target data relationship. For example, as shown in Table 3, Table 3 is a partial example of the target data relationship.

[0118] Table 3

[0119] ID number disease Average cost Number of out-of-town medical treatments xxxx heart disease 0.8 4 xxxx heart disease 0.5 0 xxxx diabetes 0.13 1 xxxx hypertension 0.34 4 xxxx hypertension 0.83 1 xxxx heart disease 0.93 0 …… …… …… ……

[0120] Thus, according to the classification of diseases M, a set of data S corresponding to the “average cost” - “number of times of seeking medical treatment in other places” for each disease can be obtained, S = {s m |m∈M},s m =(u m , v m ). Among them, s m is the two-dimensional coordinate, u m and v m They represent the average cost and the number of times of seeking medical treatment in other places. Finally, we get a set of data points about the "average cost" - "number of times of seeking medical treatment in other places" for several patients of each type of disease.

[0121] Step 206: For each average cost and each number of times of out-of-town medical treatment, a preset Gaussian mixture model is used to calculate the target cluster center corresponding to the average cost and the number of times of out-of-town medical treatment.

[0122] Among them, Figure 6 As shown, Figure 6 A flowchart of the steps of calculating the target cluster center provided in an embodiment of the present application includes:

[0123] Step 602: Obtain the number and initial parameters of Gaussian distributions corresponding to the preset Gaussian mixture model.

[0124] Step 604: For each average cost and number of out-of-town medical treatment visits, calculate the posterior probability that the average cost and number of out-of-town medical treatment visits belong to each Gaussian distribution.

[0125] Step 606: Based on each posterior probability, update the parameters of the Gaussian distribution to obtain updated parameters of the Gaussian distribution.

[0126] Step 608: Update the initial parameters based on the updated parameters of the Gaussian distribution to obtain updated initial parameters.

[0127] Step 610: If the updated initial parameters meet the preset conditions, the target cluster center is determined based on the parameters of the updated Gaussian distribution; if the updated initial parameters do not meet the preset conditions, the step of updating the initial parameters is repeated until the updated initial parameters meet the preset conditions.

[0128] The preset Gaussian mixture model regards a data set consisting of multiple data points as a mixture model consisting of multiple Gaussian distributions. Each Gaussian distribution represents a classification cluster. Its goal is to cluster each data point by estimating the parameters of each Gaussian distribution and the probability that each data point belongs to each Gaussian distribution.

[0129] The number K of Gaussian distributions can be preset, and the initial parameter θ can be expressed in the form of formula (2).

[0130]

[0131] Among them, μ k represents the mean of the kth Gaussian distribution; ∑ k represents the covariance matrix of the kth Gaussian distribution; π k It represents the mixing coefficient of the k-th Gaussian distribution, that is, the probability that the data point belongs to the k-th Gaussian distribution.

[0132] A set of average expenses and the corresponding number of out-of-town medical visits is taken as a data point x i , thus, for each data point, the posterior probability γ of the average cost and number of out-of-town medical treatments belonging to each Gaussian distribution can be calculated using formula (3). ik .

[0133]

[0134] in, and Both represent the probability density function of multivariate Gaussian distribution; the value range of k is 1-K; the value range of j is 1-K.

[0135] Thus, the parameters of the Gaussian distribution can be updated based on each posterior probability to obtain updated parameters of the Gaussian distribution, and then the initial parameters can be updated based on the updated parameters of the Gaussian distribution to obtain updated initial parameters.

[0136] The parameters of the Gaussian distribution include a mixing coefficient, a mean, and a covariance matrix. The mixing coefficient of each Gaussian distribution can be updated using formula (4), the mean of each Gaussian distribution can be updated using formula (5), and the covariance matrix of each Gaussian distribution can be updated using formula (6). The updated parameters of the Gaussian distribution can be re-substituted into the above formula (2) to update the initial parameters.

[0137]

[0138] Here, N represents the total number of data points.

[0139]

[0140] If the updated initial parameters meet the preset conditions, the target cluster center is determined based on the parameters of the updated Gaussian distribution. If the updated initial parameters do not meet the preset conditions, the step of updating the initial parameters is repeated until the updated initial parameters meet the preset conditions, and the target cluster center is determined based on the parameters of the updated Gaussian distribution that finally meet the preset conditions. The preset conditions may be that the updated initial parameters are less than a preset threshold or reach a preset number of iterations.

[0141] Step 208: Based on each target cluster center, determine the abnormal usage results of the target medical insurance consumption data of each medical insurance user.

[0142] Among them, Figure 7 As shown, Figure 7 A flowchart of steps for determining abnormal usage results provided in an embodiment of the present application includes:

[0143] Step 702: Calculate the Euclidean distance between the average cost, the number of out-of-town medical treatments and the corresponding target cluster center to obtain the Euclidean distance value.

[0144] Step 704: Based on the Euclidean distance value, determine the target cluster of average cost and number of times of out-of-town medical treatment.

[0145] Step 706: Calculate the cosine similarity between the average cost, the number of out-of-town medical treatments, and the target cluster center of the target cluster to obtain a cosine similarity value.

[0146] Step 708: Based on the cosine similarity value, determine the abnormal usage results of the target medical insurance consumption data of each medical insurance user.

[0147] Among them, if the data point formed by the current average cost and the corresponding number of out-of-town medical treatment is recorded as s y ,s y =(u y , v y ), the determined target cluster centers are the updated mean μ k , so the Euclidean distance between the average cost, the number of out-of-town medical treatment and the corresponding target cluster center can be calculated by formula (7), and the Euclidean distance value distance (s y , μ k ).

[0148]

[0149] After calculating the Euclidean distance between the data point and each target cluster center, multiple Euclidean distance values ​​are obtained, so that the target clustering cluster of average cost and number of out-of-town medical treatment can be determined according to the target cluster center corresponding to the smallest Euclidean distance value.

[0150] Therefore, the data point s can be calculated by formula (8) y The target cluster center μ corresponding to the target cluster k The cosine similarity between them is obtained by Cosine Similarity(s y , μ k ).

[0151]

[0152] Among them, s y ·μ k Represents data point s y and the target cluster center μ k The inner product of ||s y || represents data point s y The norm of ||μ k || represents the target cluster center μ k The norm of .

[0153] Finally, the abnormal usage results of the target medical insurance consumption data of each medical insurance user can be determined based on the cosine similarity value. Optionally, the range of the cosine similarity value is [0,1]. When the cosine similarity value is closer to 1, the better the data point conforms to the target cluster. Conversely, when the cosine similarity value is closer to 0, the data point conforms less to the target cluster. Thus, this indicator can be used to judge the degree of abnormality of the input data point. Optionally, when the calculated cosine similarity value is less than 0.3, the data point is judged to be abnormal data, otherwise it is normal data. Finally, the abnormal usage results of the target medical insurance consumption data of each medical insurance user can be obtained.

[0154] In an embodiment of the present application, the method includes: obtaining target medical insurance consumption data of each medical insurance user, wherein each target medical insurance consumption data includes medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content, processing the target medical insurance consumption data through a preset large language model to obtain a target data relationship, wherein the target data relationship is a correspondence between various types of diseases, each medical insurance user, average expenditure and the number of medical treatments in other places, and for each average expenditure and each number of medical treatments in other places, a preset Gaussian mixture model is used to calculate the target cluster center corresponding to the average expenditure and the number of medical treatments in other places, and based on each target cluster center, the abnormal usage result of the target medical insurance consumption data of each medical insurance user is determined. This solution processes multiple target medical insurance consumption data through a preset large language model, which can quickly and accurately extract the target data relationship, and then cluster the average expenditure and the number of out-of-town medical treatment through a preset Gaussian mixture model. Finally, based on the cluster center, the abnormal usage results of the target medical insurance consumption data of each medical insurance user can be accurately determined. It can not only calculate and judge the real-time data of medical insurance users' use of medical insurance, and quickly and accurately determine whether there is an abnormal use of medical insurance, but also help relevant departments to discover and detect non-compliant use of medical insurance in advance.

[0155] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0156] Figure 8 A structural block diagram of a device for identifying abnormal use of medical insurance provided in an embodiment of the present application.

[0157] like Figure 8 As shown, the medical insurance abnormal use identification device 800 includes:

[0158] The acquisition module 802 is used to acquire the target medical insurance consumption data of each medical insurance user; wherein each target medical insurance consumption data includes medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content.

[0159] The processing module 804 is used to process the target medical insurance consumption data through a preset large language model to obtain a target data relationship; wherein the target data relationship is the correspondence between various types of diseases, various medical insurance users, average expenditures and the number of times of medical treatment in other places.

[0160] The calculation module 806 is used to calculate the target cluster center corresponding to the average cost and the number of times of medical treatment in other places by using a preset Gaussian mixture model for each average cost and each number of times of medical treatment in other places.

[0161] The determination module 808 is used to determine the abnormal usage results of the target medical insurance consumption data of each medical insurance user based on each target cluster center.

[0162] Regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment of the method, and will not be elaborated here. Each module in the construction method of the above-mentioned abnormal medical insurance use identification method can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations of the above modules.

[0163] In one embodiment of the present application, a computer device is provided, the computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the following steps when executing the computer program:

[0164] Obtain target medical insurance consumption data of each medical insurance user; wherein each target medical insurance consumption data includes medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content;

[0165] The target medical insurance consumption data is processed by a preset large language model to obtain the target data relationship; wherein the target data relationship is the correspondence between various types of diseases, various medical insurance users, average expenditure and the number of out-of-town medical treatments;

[0166] For each average cost and each number of out-of-town medical treatment visits, the preset Gaussian mixture model is used to calculate the target cluster center corresponding to the average cost and the number of out-of-town medical treatment visits;

[0167] Based on each target cluster center, determine the abnormal usage results of the target medical insurance consumption data of each medical insurance user.

[0168] In one embodiment of the present application, the medical insurance recipient information includes the medical insurance recipient's identity information and the medical insurance payment location information. When the processor executes the computer program, the following steps are also implemented:

[0169] For each diagnosis content, a preset large language model is used to classify the diagnosis content to obtain multiple types of diseases;

[0170] Obtain detailed disease classification data of each medical insurance user when using medical insurance; wherein, detailed disease classification data includes the corresponding relationship between medical insurance user identity information, disease type, medical insurance expenditure information, medical insurance payment location information and medical insurance consumption location information;

[0171] The detailed disease classification data of each medical insurance user when using medical insurance is statistically processed to obtain the target data relationship.

[0172] In one embodiment of the present application, when the processor executes the computer program, the processor further implements the following steps:

[0173] For each diagnostic content, the preset large language model uses a preset input template to classify the diagnostic content to obtain multiple types of diseases; wherein the preset input template includes language instructions for analyzing the diagnostic content, a list of all types of diseases and an answer format.

[0174] In one embodiment of the present application, when the processor executes the computer program, the processor further implements the following steps:

[0175] Get the number and initial parameters of Gaussian distribution corresponding to the preset Gaussian mixture model;

[0176] For each average cost and number of out-of-town medical treatment, calculate the posterior probability that the average cost and number of out-of-town medical treatment belong to each Gaussian distribution;

[0177] Based on each posterior probability, the parameters of the Gaussian distribution are updated to obtain the updated parameters of the Gaussian distribution; wherein the parameters of the Gaussian distribution include a mixing coefficient, a mean, and a covariance matrix;

[0178] The initial parameters are updated based on the updated parameters of the Gaussian distribution to obtain updated initial parameters;

[0179] If the updated initial parameters meet the preset conditions, the target cluster center is determined based on the parameters of the updated Gaussian distribution; if the updated initial parameters do not meet the preset conditions, the step of updating the initial parameters is repeated until the updated initial parameters meet the preset conditions.

[0180] In one embodiment of the present application, when the processor executes the computer program, the processor further implements the following steps:

[0181] Calculate the Euclidean distance between the average cost, the number of out-of-town medical treatments and the corresponding target cluster center to obtain the Euclidean distance value;

[0182] Based on the Euclidean distance value, the target clusters of average expenditure and number of out-of-town medical treatment are determined;

[0183] Calculate the cosine similarity between the average cost, the number of out-of-town medical treatments, and the target cluster center of the target cluster to obtain the cosine similarity value;

[0184] Based on the cosine similarity value, the abnormal usage results of the target medical insurance consumption data of each medical insurance user are determined.

[0185] In one embodiment of the present application, when the processor executes the computer program, the processor further implements the following steps:

[0186] Obtain the current medical insurance consumption data of each medical insurance user from the preset database;

[0187] Performing data cleaning on the current medical insurance consumption data to obtain cleaned medical insurance consumption data; wherein the data cleaning process includes using preset data verification rules to verify the data range and data type of the current medical insurance consumption data;

[0188] The cleaned medical insurance consumption data is subjected to data standardization processing to obtain the target medical insurance consumption data of each medical insurance user.

[0189] In one embodiment of the present application, when the processor executes the computer program, the processor further implements the following steps:

[0190] After the target user's identity authentication is passed, the target interface is called through the access token to collect the original medical insurance consumption data of each medical insurance user in real time; wherein, the original medical insurance consumption data is obtained after being encrypted using a preset encryption protection protocol;

[0191] The original medical insurance consumption data is processed to obtain the current medical insurance consumption data; the data processing operation includes data cleaning, data standardization, data deduplication and merging.

[0192] The computer device provided in the embodiment of the present application has similar implementation principles and technical effects to those of the above-mentioned method embodiment, and will not be described in detail here.

[0193] In one embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0194] Obtain target medical insurance consumption data of each medical insurance user; wherein each target medical insurance consumption data includes medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content;

[0195] The target medical insurance consumption data is processed by a preset large language model to obtain the target data relationship; wherein the target data relationship is the correspondence between various types of diseases, various medical insurance users, average expenditure and the number of out-of-town medical treatments;

[0196] For each average cost and each number of out-of-town medical treatment visits, the preset Gaussian mixture model is used to calculate the target cluster center corresponding to the average cost and the number of out-of-town medical treatment visits;

[0197] Based on each target cluster center, determine the abnormal usage results of the target medical insurance consumption data of each medical insurance user.

[0198] In one embodiment of the present application, the medical insurance recipient information includes the medical insurance recipient's identity information and the medical insurance payment location information. When the computer program is executed by the processor, the following steps are also implemented:

[0199] For each diagnosis content, a preset large language model is used to classify the diagnosis content to obtain multiple types of diseases;

[0200] Obtain detailed disease classification data of each medical insurance user when using medical insurance; wherein, detailed disease classification data includes the corresponding relationship between medical insurance user identity information, disease type, medical insurance expenditure information, medical insurance payment location information and medical insurance consumption location information;

[0201] The detailed disease classification data of each medical insurance user when using medical insurance is statistically processed to obtain the target data relationship.

[0202] In one embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0203] For each diagnostic content, the preset large language model uses a preset input template to classify the diagnostic content to obtain multiple types of diseases; wherein the preset input template includes language instructions for analyzing the diagnostic content, a list of all types of diseases and an answer format.

[0204] In one embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0205] Get the number and initial parameters of Gaussian distribution corresponding to the preset Gaussian mixture model;

[0206] For each average cost and number of out-of-town medical treatment, calculate the posterior probability that the average cost and number of out-of-town medical treatment belong to each Gaussian distribution;

[0207] Based on each posterior probability, the parameters of the Gaussian distribution are updated to obtain the updated parameters of the Gaussian distribution; wherein the parameters of the Gaussian distribution include a mixing coefficient, a mean, and a covariance matrix;

[0208] The initial parameters are updated based on the updated parameters of the Gaussian distribution to obtain updated initial parameters;

[0209] If the updated initial parameters meet the preset conditions, the target cluster center is determined based on the parameters of the updated Gaussian distribution; if the updated initial parameters do not meet the preset conditions, the step of updating the initial parameters is repeated until the updated initial parameters meet the preset conditions.

[0210] In one embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0211] Calculate the Euclidean distance between the average cost, the number of out-of-town medical treatments and the corresponding target cluster center to obtain the Euclidean distance value;

[0212] Based on the Euclidean distance value, the target clusters of average expenditure and number of out-of-town medical treatment are determined;

[0213] Calculate the cosine similarity between the average cost, the number of out-of-town medical treatments, and the target cluster center of the target cluster to obtain the cosine similarity value;

[0214] Based on the cosine similarity value, the abnormal usage results of the target medical insurance consumption data of each medical insurance user are determined.

[0215] In one embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0216] Obtain the current medical insurance consumption data of each medical insurance user from the preset database;

[0217] Performing data cleaning on the current medical insurance consumption data to obtain cleaned medical insurance consumption data; wherein the data cleaning process includes using preset data verification rules to verify the data range and data type of the current medical insurance consumption data;

[0218] The cleaned medical insurance consumption data is subjected to data standardization processing to obtain the target medical insurance consumption data of each medical insurance user.

[0219] In one embodiment of the present application, when the computer program is executed by a processor, the following steps are further implemented:

[0220] After the target user's identity authentication is passed, the target interface is called through the access token to collect the original medical insurance consumption data of each medical insurance user in real time; wherein, the original medical insurance consumption data is obtained after being encrypted using a preset encryption protection protocol;

[0221] The original medical insurance consumption data is processed to obtain the current medical insurance consumption data; the data processing operation includes data cleaning, data standardization, data deduplication and merging.

[0222] The computer-readable storage medium provided in this embodiment has similar implementation principles and technical effects to those of the above method embodiments, and will not be described in detail here.

[0223] 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 used in the embodiments provided in this application 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 a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), 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.

[0224] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0225] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for identifying abnormal use of medical insurance, characterized in that: The method comprises: Obtaining target medical insurance consumption data of each medical insurance user; wherein each of the target medical insurance consumption data includes medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content; The target medical insurance consumption data is processed by a preset large language model to obtain a target data relationship; wherein the target data relationship is a correspondence between various types of diseases, various medical insurance users, average expenditures and the number of times of medical treatment in other places; For each of the average costs and the number of out-of-town medical treatments, a preset Gaussian mixture model is used to calculate the target cluster center corresponding to the average costs and the number of out-of-town medical treatments; Based on each of the target cluster centers, abnormal usage results of the target medical insurance consumption data of each of the medical insurance users are determined.

2. The method according to claim 1, characterized in that The medical insurance person information includes the medical insurance person's identity information and the medical insurance payment location information. The target medical insurance consumption data is processed by a preset large language model to obtain a target data relationship, including: For each of the diagnosis contents, the preset large language model is used to classify the diagnosis contents to obtain multiple types of diseases; Obtain detailed disease classification data of each medical insurance user when using medical insurance; wherein the detailed disease classification data includes the corresponding relationship between the medical insurance user's identity information, disease type, medical insurance expense information, medical insurance payment location information and medical insurance consumption location information; The target data relationship is obtained by statistically processing the detailed disease classification data of each medical insurance user when using the medical insurance.

3. The method according to claim 2, characterized in that For each of the diagnosis contents, the preset large language model is used to classify the diagnosis contents to obtain multiple types of diseases, including: For each of the diagnostic contents, the preset large language model uses a preset input template to classify the diagnostic contents to obtain multiple types of diseases; wherein the preset input template includes language instructions for analyzing the diagnostic contents, a list of all types of diseases and an answer format.

4. The method according to any one of claims 1 to 3, characterized in that: For each of the average costs and the number of out-of-town medical treatments, a preset Gaussian mixture model is used to calculate the target cluster center corresponding to the average costs and the number of out-of-town medical treatments, including: Get the number and initial parameters of Gaussian distribution corresponding to the preset Gaussian mixture model; For each of the average expenses and the number of out-of-town medical treatments, calculate the posterior probability that the average expenses and the number of out-of-town medical treatments belong to each of the Gaussian distributions; Based on each of the posterior probabilities, the parameters of the Gaussian distribution are updated to obtain updated parameters of the Gaussian distribution; wherein the parameters of the Gaussian distribution include a mixing coefficient, a mean, and a covariance matrix; Updating the initial parameters based on the updated Gaussian distribution parameters to obtain updated initial parameters; If the updated initial parameters meet the preset conditions, the target cluster center is determined based on the parameters of the updated Gaussian distribution; if the updated initial parameters do not meet the preset conditions, the step of updating the initial parameters is repeated until the updated initial parameters meet the preset conditions.

5. The method according to any one of claims 1 to 3, characterized in that: Determining the abnormal usage results of the target medical insurance consumption data of each medical insurance user based on each target cluster center includes: Calculate the Euclidean distance between the average cost, the number of out-of-town medical treatments and the corresponding target cluster center to obtain a Euclidean distance value; Based on the Euclidean distance value, determine a target cluster of the average cost and the number of times of out-of-town medical treatment; Calculate the cosine similarity between the average cost, the number of out-of-town medical treatments, and the target cluster center of the target cluster to obtain a cosine similarity value; Based on the cosine similarity value, abnormal usage results of the target medical insurance consumption data of each of the medical insurance users are determined.

6. The method according to any one of claims 1 to 3, characterized in that: The obtaining of target medical insurance consumption data of each medical insurance user includes: Obtain the current medical insurance consumption data of each medical insurance user from the preset database; Performing data cleaning processing on the current medical insurance consumption data to obtain cleaned medical insurance consumption data; wherein the data cleaning processing includes verifying the data range and data type of the current medical insurance consumption data using preset data verification rules; The cleaned medical insurance consumption data is subjected to data standardization processing to obtain target medical insurance consumption data of each medical insurance user.

7. The method according to claim 6, characterized in that The method further comprises: After the identity authentication of the target user is passed, the target interface is called through the access token to collect the original medical insurance consumption data of each medical insurance user in real time; wherein, the original medical insurance consumption data is obtained after being encrypted by a preset encryption protection protocol; The original medical insurance consumption data is subjected to data processing operations to obtain the current medical insurance consumption data; wherein the data processing operations include data cleaning processing, data standardization processing, data deduplication and merging processing.

8. A device for identifying abnormal use of medical insurance, characterized in that: The device comprises: An acquisition module is used to acquire target medical insurance consumption data of each medical insurance user; wherein each target medical insurance consumption data includes medical insurance person information, medical insurance consumption amount information, medical insurance consumption location information and diagnosis content; A processing module, used to process the target medical insurance consumption data through a preset large language model to obtain a target data relationship; wherein the target data relationship is a correspondence between various types of diseases, various medical insurance users, average expenditures and the number of times of medical treatment in other places; A calculation module, for calculating the target cluster center corresponding to each of the average costs and the number of out-of-town medical treatments using a preset Gaussian mixture model; A determination module is used to determine the abnormal usage results of the target medical insurance consumption data of each of the medical insurance users based on each of the target cluster centers.

9. An electronic device, characterized in that: The electronic device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the steps of the method described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The storage medium stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the steps of the method described in any one of claims 1 to 7.

Citation Information

Patent Citations

  • A method and device for detecting the medical insurance claim cost, a compute device and a storage medium

    CN109377387A

  • Data security management system for medical insurance main data storage

    CN116185722A

  • Medical insurance auditing method, apparatus and device based on data analysis, and storage medium

    WO2020119116A1