Medical insurance fixed-point management supervision system based on big data

Abstract

The invention relates to a medical insurance fixed-point management supervision system based on big data. The system comprises a data acquisition module which collects identity card information and finger vein features of insured persons, constructs an initial identity file, and extracts identity feature data from the initial identity file to form dispatching data; the task planning module generates a task detail message covering a checking position and a biological characteristic acquisition requirement according to the dispatch data, and generates a field data set with a space-time mark. The result processing module performs feature comparison on the field data set and the initial identity file, and if the matching degree is higher than a preset threshold value, an audit passing instruction is generated; and if the matching degree is lower than a preset threshold value, formulating an exception handling scheme according to the difference feature type. By adopting the system, precision and intelligence of medical insurance fixed-point management and supervision are realized, efficiency and accuracy of medical insurance fixed-point management and supervision are improved, medical insurance violation behaviors can be found and processed in time, and healthy and stable operation of a medical insurance system is promoted.

Description

Technical Field

[0001] This invention belongs to the field of medical security information technology, and in particular relates to a medical insurance designated management and supervision system based on big data. Background Technology

[0002] In the field of medical security, the coverage of the medical insurance system is constantly expanding, the number of insured persons is increasing dramatically, and the number of designated medical insurance institutions is also rising steadily. This has led to an ever-growing scale of medical insurance business and an explosive growth in the amount of data generated. At the same time, the security of the medical insurance fund faces severe challenges. Traditional methods of managing and supervising designated medical insurance institutions mainly rely on manual review and irregular spot checks. When faced with massive amounts of data and complex methods of violation, these methods are inadequate and fail to achieve comprehensive, efficient, and accurate supervision. Summary of the Invention

[0003] Therefore, it is necessary to provide a big data-based medical insurance designated point management and supervision system that can achieve precise and intelligent management and supervision of the aforementioned technical problems.

[0004] Firstly, this application provides a big data-based medical insurance designated provider management and supervision system, including:

[0005] The data acquisition module is used to obtain the ID card information and finger vein characteristics of insured persons and generate an initial identity file; it is also used to extract identity feature data from the initial identity file, generate audit task sheets according to preset priority rules, and sort them to obtain the assigned data.

[0006] The task planning module is used to generate task detail messages containing verification locations and biometric collection requirements based on the assignment data; it is also used to collect on-site ID card information and finger vein features at the target location based on the task detail messages, and generate an on-site dataset containing spatiotemporal markers.

[0007] The results processing module is used to compare the features of the on-site dataset with the initial identity file. If the matching degree exceeds the preset threshold, an audit pass instruction is generated. If the matching degree is lower than the preset threshold, an anomaly handling plan is generated based on the difference feature type.

[0008] In one embodiment, identity feature data is extracted from the initial identity file, and an audit task sheet is generated and sorted according to a preset priority rule to obtain assignment data, including:

[0009] The rule matching results are obtained by matching multiple identity feature fields in the initial identity file according to the preset priority rules; the rule matching results include field type identifiers and weight coefficients.

[0010] Input the rule matching results into the task generation model, and generate an audit task sheet based on the weight coefficients and task type.

[0011] The audit task sheets are sorted using a dynamic priority algorithm based on a time decay factor, and the output is assignment data carrying the task number and the set of target objects.

[0012] In one embodiment, a task details message containing the verification location and biometric collection requirements is generated based on the assignment data, including:

[0013] The location coordinates are extracted from the preset geographic information database based on the task number in the assigned data.

[0014] Biometric type identifiers are obtained based on the target object identifiers in the assigned data; biometric type identifiers include fingerprint or iris classification codes.

[0015] The acquisition device parameter configuration instructions are generated based on the biometric type identifier; the acquisition device parameter configuration instructions include the resolution threshold and the light intensity range.

[0016] The location coordinates and data acquisition device parameter configuration instructions are encapsulated to generate an encrypted task details message.

[0017] In one embodiment, based on the task details message, on-site ID card information and finger vein features are collected at the target location to generate an on-site dataset containing spatiotemporal markers, including:

[0018] Based on the task details message, obtain the geographic coordinates and collection time of the target location, and generate a spatiotemporal identifier.

[0019] Obtain on-site ID card information and finger vein images based on spatiotemporal identifiers.

[0020] A feature extraction algorithm is used to process finger vein images to obtain vein feature vectors.

[0021] The ID card information, vein feature vector, and spatiotemporal identifiers are linked to form a structured data record.

[0022] The structured data records are stored in a preset format to obtain the field dataset.

[0023] In one embodiment, after obtaining the field dataset, the process further includes:

[0024] Determine whether the spatiotemporal identifiers in the field dataset contain valid geographic location coordinates. If there is a coordinate offset, call the correction algorithm to correct it and obtain the corrected spatiotemporal identifiers.

[0025] The structured data record is updated based on the corrected spatiotemporal identifier, and the updated data record is matched with the preset encryption strategy to generate an encrypted data packet.

[0026] The encrypted data packet is transmitted to the target storage node, and a data storage log is generated.

[0027] Extract transmission timestamps and node addresses from data storage logs to generate data traceability identifiers.

[0028] The data traceability identifier is bound to the spatiotemporal identifier and written into the blockchain evidence repository.

[0029] In one embodiment, feature comparison is performed between the on-site dataset and the initial identity profile, including:

[0030] Obtain the feature vector set of the on-site dataset and the initial identity file.

[0031] The dynamic weight parameters extracted from the feature vector set are input into the matching degree calculation model, and the output is a multidimensional similarity matrix containing the difference feature types.

[0032] The multidimensional similarity matrix can be constructed using the following formula:

[0033]

[0034] Among them, S ij Let S represent the elements in the multidimensional similarity matrix S, n represent the number of feature vectors in the on-site dataset and the initial identity profile, m represent the dimension of the feature vectors, and ω represent the number of feature vectors. k Let x represent the weight of the k-th feature dimension. ik y represents the value of the i-th feature vector in the k-th dimension of the dataset. jk This represents the value of the j-th feature vector in the k-th dimension of the initial identity file.

[0035] A heatmap of difference distribution is generated based on a multidimensional similarity matrix; each region in the heatmap of difference distribution corresponds to a specific feature dimension.

[0036] If the number of regions exceeding a preset threshold in the differential distribution heatmap reaches a critical value, an anomaly handling scheme is generated. The anomaly handling scheme calls the corresponding set of compensation algorithms based on the type distribution of low-matching features in the multidimensional similarity matrix.

[0037] The set of compensation algorithms is executed to generate a repaired subset of feature vectors; the subset of feature vectors replaces the abnormal feature regions in the on-site dataset.

[0038] The repaired feature vector subset is input into the matching degree calculation model for secondary verification, and the updated multidimensional similarity matrix is ​​output.

[0039] If the number of regions below the preset threshold in the updated multidimensional similarity matrix becomes zero, an audit pass instruction is generated.

[0040] In one embodiment, the result processing module further includes:

[0041] Obtain the set of spatiotemporal coordinates of the on-site dataset; the set of spatiotemporal coordinates includes timestamps and geographic location information.

[0042] Parse the administrative division code corresponding to the registered address in the initial identity file to generate a set of address codes.

[0043] Geographic grid mapping is performed based on the spatiotemporal coordinate set and the address code set to obtain the grid correlation degree between each spatiotemporal coordinate and the corresponding address code.

[0044] The weight coefficients of grid correlation are calculated using the geographic distance decay algorithm and input into the dynamic threshold model in the priority rules to obtain the evaluation results; the dynamic threshold model includes a preset correlation score range.

[0045] The parameters of the dynamic threshold model are adjusted according to the correlation score interval in which the weight coefficients are located, and the address matching threshold in the priority rule is updated to obtain the updated evaluation results; the address matching threshold is used to filter the correlation targets of the next round of spatiotemporal labeling.

[0046] Secondly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the system as described above.

[0047] Thirdly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the system as described above.

[0048] The aforementioned big data-based medical insurance designated provider management and supervision system, computer equipment, and storage media mainly include a data acquisition module, a task planning module, and a result processing module. The data acquisition module is responsible for collecting the ID card information and finger vein characteristics of insured individuals, constructing an initial identity file, extracting identity feature data, generating audit task orders according to preset priority rules, and sorting them to form assignment data. The task planning module generates a task detail message covering the verification location and biometric collection requirements based on the assignment data. Simultaneously, based on this message, it collects on-site ID card information and finger vein characteristics at the target location, generating an on-site dataset with spatiotemporal markers. The result processing module compares the on-site dataset with the initial identity file. If the matching degree is higher than a preset threshold, an audit pass instruction is generated; if the matching degree is lower than the preset threshold, an anomaly handling plan is formulated based on the difference feature type. This system achieves precision and intelligence in medical insurance designated provider management and supervision. It improves the efficiency and accuracy of medical insurance designated provider management and supervision, enables timely detection and handling of medical insurance violations, effectively safeguards the security of the medical insurance fund, protects the legitimate rights and interests of insured individuals, and promotes the healthy and stable operation of the medical insurance system. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A structural block diagram of a big data-based medical insurance designated management and supervision system provided for embodiments of the present invention;

[0051] Figure 2 The flowchart provided in this embodiment of the invention is as follows: extracting identity feature data from an initial identity file, generating an audit task sheet according to a preset priority rule, and sorting the data to obtain the assignment data.

[0052] Figure 3 This is a flowchart illustrating how, based on task detail messages, on-site ID card information and finger vein features are collected at a target location to generate a field dataset containing spatiotemporal markers, as provided in an embodiment of the present invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0054] In one embodiment, such as Figure 1 As shown, this application provides a big data-based medical insurance designated provider management and supervision system, which may include:

[0055] The data acquisition module 101 is used to acquire the ID card information and finger vein characteristics of the insured persons and generate an initial identity file; it is also used to extract identity feature data from the initial identity file, generate an audit task sheet according to the preset priority rules, and sort and obtain the assignment data.

[0056] Specifically, this module utilizes advanced identity information collection equipment and high-precision finger vein recognition technology to accurately obtain the insured person's ID card information and finger vein characteristics. After collection, the system integrates this information to generate a structured and standardized initial identity file, providing raw data support for subsequent medical insurance management. Simultaneously, to achieve accurate and efficient audit task allocation, the data collection module extracts key identity feature data from the initial identity file, such as the insured person's age, insurance type, and frequency of medical visits. This data is analyzed according to preset priority rules to generate audit task sheets. Then, a professional sorting algorithm sorts the audit task sheets, ultimately yielding allocation data containing task priority, task number, and target object set.

[0057] The task planning module 102 is used to generate a task detail message containing the verification location and biometric collection requirements based on the assignment data; it is also used to collect on-site ID card information and finger vein features at the target location based on the task detail message, and generate an on-site dataset containing spatiotemporal markers.

[0058] Upon receiving the assigned data, the module first quickly extracts the corresponding verification location coordinates from its built-in geographic information database based on the task number. Simultaneously, it combines this with the target object identifier to obtain the corresponding biometric data collection type identifier, such as fingerprint, iris, or finger vein. Based on this information, the task planning module generates detailed configuration instructions for the data collection equipment, specifying technical parameters such as the collection resolution threshold and light intensity range to ensure the accuracy and validity of the collected data. Subsequently, the verification location coordinates and the data collection equipment parameter configuration instructions are integrated and encrypted to generate a task details message. After completing task planning, the task planning module, based on the task details message, uses a mobile smart device to collect on-site ID card information and finger vein features at the target location. During the collection process, an integrated positioning system and time synchronization technology are used to add precise spatiotemporal markers to the collected data, ultimately generating a field dataset containing spatiotemporal markers.

[0059] The result processing module 103 is used to compare the features of the on-site dataset with the initial identity file. If the matching degree exceeds the preset threshold, an audit pass instruction is generated. If the matching degree is lower than the preset threshold, an anomaly handling scheme is generated according to the difference feature type.

[0060] Specifically, this module first extracts features from the on-site dataset and the initial identity files, obtaining their respective feature vector sets. Next, dynamic weight parameters are extracted from the feature vector sets and input into the matching degree calculation model. Through complex algorithms, a multi-dimensional similarity matrix containing the types of differing features is output. To more intuitively analyze the comparison results, the system generates a difference distribution heatmap based on the multi-dimensional similarity matrix. Different regions in the heatmap correspond to different feature dimensions, making the differences readily apparent. If the number of regions exceeding a preset threshold in the difference distribution heatmap reaches a critical value, the result processing module will determine that an anomaly exists. Based on the type distribution of low-matching features in the multi-dimensional similarity matrix, it calls the corresponding set of compensation algorithms to generate a repaired feature vector subset, replacing the abnormal feature regions in the on-site dataset. Afterward, the repaired feature vector subset is input again into the matching degree calculation model for secondary verification, outputting an updated multi-dimensional similarity matrix. If the number of regions below the preset threshold in the updated multidimensional similarity matrix becomes zero, it indicates that the identity matching is successful, and the system will generate an audit approval instruction; otherwise, the exception handling process will continue to be executed, and a corresponding exception handling plan will be generated according to the difference feature type, such as initiating further investigation procedures or notifying relevant departments, to ensure the rigor and accuracy of medical insurance designated management and supervision.

[0061] The aforementioned big data-based medical insurance designated provider management and supervision system, computer equipment, and storage media mainly include a data acquisition module, a task planning module, and a result processing module. The data acquisition module is responsible for collecting the ID card information and finger vein characteristics of insured individuals, constructing an initial identity file, extracting identity feature data, generating audit task orders according to preset priority rules, and sorting them to form assignment data. The task planning module generates a task detail message covering the verification location and biometric collection requirements based on the assignment data. Simultaneously, based on this message, it collects on-site ID card information and finger vein characteristics at the target location, generating an on-site dataset with spatiotemporal markers. The result processing module compares the on-site dataset with the initial identity file. If the matching degree is higher than a preset threshold, an audit pass instruction is generated; if the matching degree is lower than the preset threshold, an anomaly handling plan is formulated based on the difference feature type. This system achieves precision and intelligence in medical insurance designated provider management and supervision. It improves the efficiency and accuracy of medical insurance designated provider management and supervision, enables timely detection and handling of medical insurance violations, effectively safeguards the security of the medical insurance fund, protects the legitimate rights and interests of insured individuals, and promotes the healthy and stable operation of the medical insurance system.

[0062] In one embodiment, such as Figure 2 As shown, extracting identity feature data from the initial identity file, generating audit task sheets according to preset priority rules, and sorting them to obtain assignment data may include the following steps:

[0063] Step S201: Match the rule types of multiple identity feature fields in the initial identity file according to the preset priority rules to obtain the rule matching result; the rule matching result includes field type identifier and weight coefficient.

[0064] Step S202: Input the rule matching results into the task generation model, and generate an audit task sheet based on the weight coefficients and task type.

[0065] Step S203: The audit task sheets are sorted using a dynamic priority algorithm based on time decay factor, and the assignment data carrying the task number and the target object set is output.

[0066] Preferably, the rule types of multiple identity feature fields in the initial identity file are first matched according to preset priority rules to obtain rule matching results containing field type identifiers and weight coefficients. Then, this rule matching result is input into the task generation model, and an audit task sheet is generated by combining the weight coefficients and task type. Finally, a dynamic priority algorithm based on a time decay factor is used to sort the audit task sheets, ultimately outputting assignment data carrying task numbers and target object sets.

[0067] This embodiment matches identity feature fields using preset rules, accurately determining the importance of different features, i.e., weight coefficients, providing a crucial reference for task generation. The task generation model combines weight coefficients with task type to generate audit task sheets, making tasks more targeted. Furthermore, the dynamic priority algorithm based on a time decay factor considers the impact of time on task urgency, prioritizing tasks that require immediate attention. The generated assignment data clearly identifies the task number and target object set, allowing staff to clearly understand the task content and execution targets, effectively improving the efficiency and accuracy of medical insurance auditing, helping to promptly detect medical insurance violations, and safeguarding the security of the medical insurance fund.

[0068] In one embodiment, generating a task details message containing verification locations and biometric collection requirements based on the assignment data may include the following steps:

[0069] Step S301: Extract the verification location coordinates from the preset geographic information database according to the task number in the assignment data.

[0070] Step S302: Obtain biometric type identifier based on the target object identifier in the assignment data; biometric type identifier includes fingerprint or iris classification code.

[0071] Step S303: Generate a parameter configuration instruction for the acquisition device based on the biometric type identifier; the parameter configuration instruction for the acquisition device includes a resolution threshold and a light intensity range.

[0072] Step S304: Encapsulate the verification location coordinates and the data acquisition device parameter configuration instructions to generate an encrypted task details message.

[0073] First, starting with the task number in the assignment data, the corresponding verification location coordinates are accurately extracted from a pre-defined geographic information database. Simultaneously, based on the target object identifier in the assignment data, the corresponding biometric type identifier is obtained, encompassing classification codes such as fingerprints or iris scans. Then, based on the obtained biometric type identifier, a data acquisition device parameter configuration instruction containing information such as resolution thresholds and illumination intensity ranges is generated. Finally, the extracted verification location coordinates and the generated data acquisition device parameter configuration instruction are integrated and encapsulated, and then encrypted to generate an encrypted task details message.

[0074] This embodiment extracts the location coordinates for verification by task number, enabling precise location of the specific location requiring medical insurance audit, thus making the audit work more targeted. Based on the target object identifier, it obtains the biometric type identifier and generates parameter configuration instructions for the data collection device, ensuring that the data collection device can be specifically configured according to different biometric types, guaranteeing the reliability, accuracy, and effectiveness of the collected data. The location coordinates and device parameter configuration instructions are encapsulated and encrypted to generate a task details message, achieving both information integration and standardized transmission, while also ensuring the security and integrity of the data during transmission. This optimizes the medical insurance audit task planning process, improving the efficiency and accuracy of the audit work.

[0075] In one embodiment, such as Figure 3 As shown, collecting on-site ID card information and finger vein features at the target location based on the task details message to generate an on-site dataset containing spatiotemporal markers may include the following steps:

[0076] Step S401: Obtain the geographic coordinates and collection time of the target location based on the task details message, and generate a spatiotemporal identifier.

[0077] Step S402: Obtain on-site ID card information and finger vein image based on spatiotemporal identifier.

[0078] Step S403: The finger vein image is processed using a feature extraction algorithm to obtain the vein feature vector.

[0079] Preferably, the specific steps are as follows:

[0080] Step S4031: The acquired finger vein image is converted to grayscale to reduce the amount of data.

[0081] Step S4032: Use Gaussian filtering to remove noise interference in the image and enhance the image clarity.

[0082] Step S4033: Use an edge detection algorithm to determine the approximate outline of the finger veins and extract the region of interest containing finger vein features.

[0083] Step S4034: Use algorithms such as SIFT to detect feature points in the region of interest. The feature points have rotation, scale and illumination invariance.

[0084] Step S4035: Calculate the gradient direction and magnitude of the detected feature points to generate corresponding feature vectors.

[0085] Step S404: Associate the ID card information, vein feature vector, and spatiotemporal identifier to form a structured data record.

[0086] Step S405: Store the structured data records in a preset format to obtain the field dataset.

[0087] Furthermore, the system first obtains the geographic coordinates of the target location and the collection time based on the task details message, and combines these two to generate a spatiotemporal identifier. Using this identifier, the system acquires the ID card information and finger vein images of the insured individuals on-site. Then, a specific feature extraction algorithm is used to process the finger vein images to obtain vein feature vectors. Next, the system integrates the acquired ID card information, the generated vein feature vectors, and the spatiotemporal identifier to form structured data records. Finally, these structured data records are stored according to a preset format, thus obtaining the on-site dataset.

[0088] This embodiment generates a spatiotemporal identifier through task detail messages, ensuring that the collected data has accurate time and geographical location identification. Based on the spatiotemporal identifier, ID card information and finger vein images are collected, and feature extraction is performed on the finger vein images, guaranteeing the accuracy and validity of the collected data and facilitating precise identification of insured individuals. Various data types are correlated to form structured data records and stored in a preset format as a field dataset, making data management more standardized and orderly. This facilitates subsequent comparison and analysis with initial identity files, improving the efficiency and accuracy of medical insurance designated management and supervision, and effectively safeguarding the security of the medical insurance fund and the legitimate rights and interests of insured individuals.

[0089] In one embodiment, after obtaining the field dataset, the process may further include:

[0090] Step S501: Determine whether the spatiotemporal identifier in the field dataset contains valid geographic location coordinates. If there is a coordinate offset, call the correction algorithm to correct it and obtain the corrected spatiotemporal identifier.

[0091] Step S502: Update the structured data record according to the corrected spatiotemporal identifier, match the updated data record with the preset encryption strategy, and generate an encrypted data packet.

[0092] Step S503: Transmit the encrypted data packet to the target storage node and generate a data storage log.

[0093] Step S504: Extract the transmission timestamp and node address from the data storage log to generate a data traceability identifier.

[0094] Step S505: Bind the data traceability identifier with the spatiotemporal identifier and write it into the blockchain evidence repository.

[0095] Specifically, the system first checks the validity of spatiotemporal identifiers in the on-site dataset, focusing on the validity of their geographical coordinates. If coordinate offsets are found, the system automatically invokes a correction algorithm to obtain accurate spatiotemporal identifiers. Based on this correction, the system further updates the structured data records to ensure data accuracy. The updated data records are then encrypted according to a preset encryption strategy, generating encrypted data packets. Subsequently, the encrypted data packets are transmitted to the target storage node for storage, and a detailed data storage log is generated simultaneously. The system extracts key information such as transmission timestamps and node addresses from the data storage log to generate a data traceability identifier. Finally, the data traceability identifier is bound to the spatiotemporal identifier and written into a blockchain repository for permanent storage.

[0096] This embodiment ensures the accuracy of spatiotemporal information in the collected data and improves data quality by checking and correcting the coordinate validity of spatiotemporal identifiers. Data encryption and storage guarantee data security and integrity, preventing data tampering during storage and transmission. The generation of data storage logs and the creation of data traceability identifiers provide a clear traceability path for the source and flow of data. Furthermore, writing relevant information into a blockchain repository leverages the immutability of blockchain to further enhance the credibility and authority of the data. Overall, this process effectively supports the standardized operation of the medical insurance designated provider management and supervision system, provides a reliable data foundation for the safe supervision of medical insurance funds, helps safeguard the legitimate rights and interests of insured individuals, and enhances the credibility of medical insurance management.

[0097] In one embodiment, comparing the on-site dataset with the initial identity profile may include the following steps:

[0098] Step S601: Obtain the feature vector set of the on-site dataset and the initial identity file.

[0099] Step S602: Extract the dynamic weight parameters from the feature vector set and input them into the matching degree calculation model to output a multidimensional similarity matrix containing the difference feature types.

[0100] Preferably, the multidimensional similarity matrix can be constructed using the following formula:

[0101]

[0102] Among them, S ij Let x represent the element in the multidimensional similarity matrix S, which represents the i-th feature vector x of the field dataset. i The j-th feature vector y of the initial identity file j The similarity is given by ω, where n represents the number of feature vectors in the on-site dataset and the initial identity profile, m represents the dimension of the feature vectors, and ω represents the similarity. k ω represents the weight of the k-th feature dimension. Different feature dimensions have different importance in similarity calculation. k To reflect this difference, for example, in a medical insurance identity verification scenario, the weight of finger vein features may be higher, while the weight of other features such as age may be relatively lower, ω k It will be set and adjusted according to the actual situation, x ik y represents the value of the i-th feature vector in the k-th dimension of the dataset. jk This represents the value of the j-th feature vector in the k-th dimension of the initial identity file.

[0103] Step S603: Generate a difference distribution heatmap based on the multidimensional similarity matrix; each region in the difference distribution heatmap corresponds to a specific feature dimension.

[0104] If the number of regions exceeding a preset threshold in the differential distribution heatmap reaches a critical value, an anomaly handling scheme is generated. The anomaly handling scheme calls the corresponding set of compensation algorithms based on the type distribution of low-matching features in the multidimensional similarity matrix.

[0105] Step S604: Execute the compensation algorithm set to generate a repaired feature vector subset; the feature vector subset replaces the abnormal feature regions in the on-site dataset.

[0106] Step S605: Input the repaired feature vector subset into the matching degree calculation model for secondary verification, and output the updated multidimensional similarity matrix.

[0107] If the number of regions below the preset threshold in the updated multidimensional similarity matrix becomes zero, an audit pass instruction is generated.

[0108] Preferably, the system first acquires feature vector sets from the on-site dataset and the initial identity file, respectively. Next, dynamic weight parameters are extracted from these feature vector sets and input into the matching degree calculation model, resulting in a multidimensional similarity matrix containing the types of differing features. Then, a difference distribution heatmap is generated based on the multidimensional similarity matrix, where different regions correspond to specific feature dimensions. When the number of regions exceeding a preset threshold in the difference distribution heatmap reaches a critical value, the system calls the corresponding compensation algorithm set based on the type distribution of low-matching features in the multidimensional similarity matrix to generate a repaired subset of feature vectors, which is used to replace the abnormal feature regions in the on-site dataset. Afterward, the repaired subset of feature vectors is input again into the matching degree calculation model for secondary verification, resulting in an updated multidimensional similarity matrix. If the number of regions below the preset threshold in the updated multidimensional similarity matrix becomes zero, the system determines that the audit has passed and generates an audit pass command.

[0109] This embodiment quantifies the differences between on-site data and initial files by acquiring a set of feature vectors and calculating a multidimensional similarity matrix, clearly presenting the types of differences. A heatmap of the difference distribution visually displays the discrepancies, facilitating quick identification of potential problems by staff. When anomalies occur, a compensation algorithm is invoked based on low-matching features for repair, and secondary verification ensures accuracy, avoiding misjudgments. This improves the accuracy of medical insurance auditing, effectively identifies and handles medical insurance violations, safeguards the medical insurance fund, and maintains the normal operation of the medical insurance system.

[0110] In one embodiment, the result processing module may further include:

[0111] Step S701: Obtain the spatiotemporal coordinate set of the on-site dataset; the spatiotemporal coordinate set includes timestamps and geographic location information.

[0112] Step S702: Parse the administrative division code corresponding to the registration address of the initial identity file to generate an address code set.

[0113] Step S703: Perform geographic grid mapping based on the spatiotemporal coordinate set and the address code set to obtain the grid correlation degree between each spatiotemporal coordinate and the corresponding address code.

[0114] Step S704: The weight coefficient of the grid correlation degree is calculated using the geographical distance decay algorithm and input into the dynamic threshold model in the priority rule to obtain the evaluation result; the dynamic threshold model includes a preset correlation degree score range.

[0115] Step S705: Adjust the parameters of the dynamic threshold model according to the correlation score interval in which the weight coefficient is located, update the address matching threshold in the priority rule, and obtain the updated evaluation result; the address matching threshold is used to filter the associated targets of the next round of spatiotemporal labeling.

[0116] In the data analysis phase of the medical insurance designated provider management and supervision system, the spatiotemporal coordinate set of the on-site dataset is first acquired, encompassing timestamps and geographic location information. Simultaneously, the registration address of the initial identity file is parsed to generate a corresponding set of administrative division codes. Then, the spatiotemporal coordinate set and the address code set are geographically mapped to a grid to obtain the grid correlation degree between each spatiotemporal coordinate and its corresponding address code. Next, a geographic distance decay algorithm is used to calculate the weight coefficient of the grid correlation degree, which is then input into a dynamic threshold model in the priority rules to obtain the evaluation result. This model presets a correlation degree scoring interval. Finally, based on the position of the weight coefficient within the correlation degree scoring interval, the parameters of the dynamic threshold model are adjusted to update the address matching threshold in the priority rules, thus obtaining an updated evaluation result. The updated address matching threshold is used to filter the correlation targets for the next round of spatiotemporal marking.

[0117] This embodiment provides scientific and dynamic data analysis support for the medical insurance designated hospital management and supervision system. Through geographic grid mapping and correlation calculation, it can effectively analyze the relationship between insured individuals' medical treatment behavior and their registered addresses. The application of a geographic distance decay algorithm makes the weighting coefficients more consistent with reality, and the adjustment of dynamic threshold models and address matching thresholds allows the system to flexibly optimize priority rules based on actual data feedback. This not only improves the accuracy of medical insurance supervision, enabling more accurate screening of insured individuals and medical treatment behaviors with potential risks, but also helps to rationally allocate supervisory resources, improve supervisory efficiency, ensure the safety of the medical insurance fund, and maintain the stable operation of the medical insurance system.

[0118] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0119] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of a big data-based medical insurance designated management and supervision system as described above.

[0120] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0121] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0122] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A big data-based medical insurance designated provider management and supervision system, characterized in that, The system includes: The data acquisition module is used to acquire the ID card information and finger vein characteristics of insured persons and generate an initial identity file; it is also used to extract identity feature data from the initial identity file, generate an audit task sheet according to a preset priority rule, and sort the data to obtain the assignment data. The task planning module is used to generate a task detail message containing the verification location and biometric collection requirements based on the assignment data; it is also used to collect on-site ID card information and finger vein features at the target location based on the task detail message, and generate an on-site dataset containing spatiotemporal markers. The result processing module is used to compare the on-site dataset with the initial identity file by feature. If the matching degree exceeds a preset threshold, an audit pass instruction is generated. If the matching degree is lower than the preset threshold, an anomaly handling scheme is generated according to the difference feature type.

2. The system according to claim 1, characterized in that, The process of extracting identity feature data from the initial identity file, generating audit task sheets according to preset priority rules, and sorting them to obtain assignment data includes: The rule matching results are obtained by matching multiple identity feature fields in the initial identity file according to the preset priority rules; the rule matching results include field type identifiers and weight coefficients. The rule matching results are input into the task generation model, and an audit task sheet is generated according to the weight coefficients and task type. The audit task sheets are sorted using a dynamic priority algorithm based on a time decay factor, and the output is assignment data carrying task number and target object set.

3. The system according to claim 1, characterized in that, The step of generating a task details message containing the verification location and biometric collection requirements based on the assignment data includes: The location coordinates are extracted from a preset geographic information database based on the task number in the assignment data. A biometric type identifier is obtained based on the target object identifier in the allocation data; the biometric type identifier includes fingerprint or iris classification code; A parameter configuration instruction for the acquisition device is generated based on the biometric type identifier; the parameter configuration instruction for the acquisition device includes a resolution threshold and a light intensity range. The verification location coordinates and the parameter configuration instructions of the acquisition device are encapsulated to generate an encrypted task details message.

4. The system according to claim 1, characterized in that, The step of collecting on-site ID card information and finger vein features at the target location based on the task details message, and generating an on-site dataset containing spatiotemporal markers, includes: Based on the task details message, obtain the geographic coordinates and collection time of the target location, and generate a spatiotemporal identifier; Obtain on-site ID card information and finger vein images based on the spatiotemporal identifier; The finger vein image is processed using a feature extraction algorithm to obtain a vein feature vector; The ID card information, vein feature vector, and spatiotemporal identifier are associated to form a structured data record; The structured data records are stored in a preset format to obtain the field dataset.

5. The system according to claim 4, characterized in that, After obtaining the field dataset, the following is also included: Determine whether the spatiotemporal identifier in the field dataset contains valid geographic location coordinates. If there is a coordinate offset, call the correction algorithm to correct it and obtain the corrected spatiotemporal identifier. The structured data record is updated according to the corrected spatiotemporal identifier, and the updated data record is matched with a preset encryption strategy to generate an encrypted data packet; The encrypted data packet is transmitted to the target storage node, and a data storage log is generated; Extract the transmission timestamp and node address from the data storage log to generate a data traceability identifier; The data traceability identifier is bound to the spatiotemporal identifier and written into the blockchain evidence repository.

6. The system according to claim 1, characterized in that, The step of comparing the on-site dataset with the initial identity file by feature includes: Obtain the feature vector set of the on-site dataset and the initial identity file; The dynamic weight parameters extracted from the feature vector set are input into the matching degree calculation model, and the output is a multidimensional similarity matrix containing the difference feature types; The multidimensional similarity matrix is ​​constructed using the following formula: Among them, S ij Let S represent the elements in the multidimensional similarity matrix S, n represent the number of feature vectors in the on-site dataset and the initial identity profile, m represent the dimension of the feature vectors, and ω represent the number of feature vectors. k Let x represent the weight of the k-th feature dimension. ik y represents the value of the i-th feature vector in the k-th dimension of the dataset. jk This represents the value of the j-th feature vector in the k-th dimension of the initial identity file. A difference distribution heatmap is generated based on the multidimensional similarity matrix; each region in the difference distribution heatmap corresponds to a specific feature dimension. If the number of regions exceeding the preset threshold in the differential distribution heatmap reaches a critical value, an anomaly handling scheme is generated; the anomaly handling scheme calls the corresponding set of compensation algorithms based on the type distribution of low matching features in the multidimensional similarity matrix. The set of compensation algorithms is executed to generate a repaired subset of feature vectors; the subset of feature vectors replaces the abnormal feature regions in the on-site dataset. The repaired feature vector subset is input into the matching degree calculation model for secondary verification, and the updated multidimensional similarity matrix is ​​output. If the number of regions below the preset threshold in the updated multidimensional similarity matrix becomes zero, an audit pass instruction is generated.

7. The system according to claim 1, characterized in that, The result processing module further includes: Obtain the spatiotemporal coordinate set of the on-site dataset; the spatiotemporal coordinate set includes timestamps and geographic location information; Parse the administrative division code corresponding to the registration address of the initial identity file to generate an address code set; Geographic grid mapping is performed based on the spatiotemporal coordinate set and the address code set to obtain the grid correlation degree between each spatiotemporal coordinate and the corresponding address code; The weight coefficients of the grid correlation degree are calculated using the geographic distance decay algorithm and input into the dynamic threshold model in the priority rule to obtain the evaluation result; the dynamic threshold model includes a preset correlation degree scoring range. The parameters of the dynamic threshold model are adjusted according to the correlation score interval in which the weight coefficient is located, and the address matching threshold in the priority rule is updated to obtain the updated evaluation result; the address matching threshold is used to filter the associated targets of the next round of spatiotemporal labeling.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the system according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the system according to any one of claims 1 to 7.

Images