Medical health data automatic grading method and device and storage medium

Abstract

The invention provides a medical health data automatic grading method and device and a storage medium, and the method comprises the steps: obtaining a to-be-graded medical data set, and carrying out the preprocessing of the to-be-graded medical data set, and obtaining a first-stage grading element feature matrix; compressing the first-level grading element feature matrix to obtain a second-level grading element feature matrix having a mapping relation with the first-level grading element feature matrix; the second-level grading element feature matrix is input into an intelligent grading model, a grading result of the second-level grading element feature matrix is obtained, and the intelligent grading model is a grading model obtained by training historical medical data with grading labels; and mapping grading labels of the graded second-level grading element feature matrix back to the first-level grading element feature matrix according to a mapping relation during compression, and completing grading of the first-level grading element feature matrix. According to the automatic data grading method provided by the invention, the data grading efficiency and the data grading reliability can be improved.

Description

Technical Field

[0001] The present invention relates to the field of medical data security technology, and in particular to a method, device and storage medium for automatic classification of medical health data. Background Art

[0002] Data plays an increasingly important role in economic and social development. Once tampered with, destroyed, leaked, or illegally obtained or used, it will cause varying degrees of harm to national security, public interests, or the legitimate rights and interests of individuals and organizations. Therefore, the classification and grading of data protection is related to national security, the lifeline of the national economy, important people's livelihood and major public interests.

[0003] In recent years, massive data in fields such as medical care has grown rapidly. However, the current data classification work is still mainly done manually, especially data classification involves high-dimensional classification element information (including data fields, groups, regions, accuracy, scale, depth, coverage, importance, security risks, etc.), and data classification also requires timely dynamic adjustment, etc., which will consume a lot of manpower, low efficiency, and prone to classification errors.

[0004] Therefore, how to improve the efficiency and reliability of data classification in the medical field has become a technical problem that needs to be urgently solved by technical personnel in this field. Summary of the invention

[0005] The present invention provides a method, device and storage medium for automatic classification of medical health data, which solves the problem of low data classification efficiency existing in the related art.

[0006] As a first aspect of the present invention, a method for automatically grading medical and health data is provided, comprising:

[0007] Obtaining a medical data set to be classified;

[0008] Preprocessing the medical data set to be classified to obtain a first-level classification element feature matrix of the medical data set;

[0009] Compressing the first-level hierarchical element feature matrix of the medical data set according to the association or characteristics of data distribution to obtain a second-level hierarchical element feature matrix having a mapping relationship with the first-level hierarchical element feature matrix;

[0010] Inputting the second-level grading factor feature matrix into an intelligent grading model to obtain a grading result of the second-level grading factor feature matrix, wherein the intelligent grading model is a grading model trained based on historical medical data with grading labels;

[0011] The classification labels of the classified second-level classification feature matrix are mapped back to the first-level classification feature matrix according to the mapping relationship during compression, thereby completing the classification of the first-level classification feature matrix.

[0012] Furthermore, it also includes the classification of new medical data sets, including:

[0013] When collecting a new medical data set, the newly added medical data set is processed according to the hierarchical elements respectively, and the processed data is discretized to obtain the first-level hierarchical element feature matrix of the newly added medical data set; the first-level hierarchical element feature matrix of the newly added medical data set obtained after preprocessing is compressed to obtain the second-level hierarchical element feature matrix of the newly added medical data set having a mapping relationship with the first-level hierarchical element feature matrix;

[0014] Input the data of the second-level hierarchical element feature matrix of the newly added medical data set into the intelligent grading model for grading according to the similarity. If the similarity is higher than or equal to the threshold, the data in the second-level hierarchical element feature matrix adopts the grading label of the intelligent grading model, and the grading label of the second-level hierarchical element feature matrix after the grading of the newly added medical data set is mapped back to the first-level hierarchical element feature matrix according to the mapping relationship during compression;

[0015] The data with similarity below the threshold is manually graded and assigned with graded labels, then input into the intelligent grading model, stored as a new intelligent grading label dataset, and the intelligent grading model is updated.

[0016] Furthermore, an intelligent classification model is obtained by training based on historical medical data with classification labels, including:

[0017] Obtain historical medical data with hierarchical labels;

[0018] Construct a sample hierarchical element feature matrix based on historical medical data with hierarchical labels;

[0019] Initialize the artificial immune network;

[0020] The sample classification element feature matrix is ​​trained according to an artificial immune network to obtain an intelligent classification model.

[0021] Furthermore, the sample classification element feature matrix is ​​trained according to an artificial immune network to obtain an intelligent classification model, including:

[0022] Taking N data classification element vectors in the classification element feature matrix of the sample data set as antigens, where N is a natural number greater than 1;

[0023] retaining antibodies with high fitness for the antigen as memory network data;

[0024] Iteratively search for memory network data that meets fitness conditions based on the artificial immune network algorithm;

[0025] The memory network data satisfying the fitness condition is used as an intelligent graded label data set, and the intelligent graded label data set is stored.

[0026] Furthermore, the classification element feature matrix of the medical data set is input into the intelligent classification model to obtain the classification result of the medical data set, including:

[0027] Performing fitness calculation on the hierarchical element feature matrix of the medical data set and the intelligent hierarchical label data set;

[0028] Determining whether the fitness is greater than a preset threshold;

[0029] If it is greater than a preset threshold, the grading result of the medical data set is determined according to the intelligent grading label data set.

[0030] Further, if it is not greater than a preset threshold, manually performing classification processing on the classification element feature matrix of the medical data set, and obtaining a classification result of the medical data set;

[0031] The intelligent grading label dataset is updated according to the grading results of the medical dataset obtained by manual grading.

[0032] Furthermore, the medical data set to be classified is preprocessed to obtain a classification element feature matrix of the medical data set, including:

[0033] Performing data preprocessing on the medical data set to be classified to obtain a medical processing data set;

[0034] The medical treatment data set is processed according to the hierarchical elements respectively, and the processed data is discretized to obtain the hierarchical element feature matrix of the medical data set.

[0035] The grading elements include qualitative grading elements and quantitative grading elements.

[0036] Furthermore, the medical treatment data set is processed according to the classification elements respectively, and the processed data is discretized, including:

[0037] Qualitatively describing the medical treatment data set according to qualitative grading elements to obtain qualitative grading discrete values;

[0038] Calculating the medical treatment data set according to the quantitative grading elements to obtain data distribution results of different quantitative grading elements;

[0039] Discretize the data distribution results of different quantitative grading elements to obtain quantitative grading discrete values;

[0040] The qualitative grading discrete values ​​and the quantitative grading discrete values ​​are constructed as a grading element feature matrix of the medical data set.

[0041] The present invention also provides an automatic classification device for medical and health data, comprising: a processor and a memory, the processor and the memory are communicatively connected, the memory stores at least one program instruction, and the processor is used to load and execute at least one program instruction to implement the automatic data classification method described in any one of claims 1 to 8.

[0042] The present invention also provides a computer storage medium, wherein the computer storage medium stores at least one program instruction, and the at least one program instruction is loaded and executed by a processor to implement the data automatic classification method described in any one of claims 1 to 8.

[0043] The automatic data classification method provided by the present invention preprocesses the original medical data set in a matrix-constructing manner according to a classification element matrix, and further compresses the preprocessed classification element matrix, and then intelligently classifies the compressed matrix-constructed data through an intelligent classification model, thereby greatly reducing the workload of grading and labeling data. The second-level classification element matrix after classification maps the classification labels back to the first-level classification element matrix through a mapping relationship with the first-level classification element matrix, which can ensure the efficiency and accuracy of data classification. In addition, the classification method can automatically perform data classification processing online, thereby realizing automatic data classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present invention, but do not constitute a limitation of the present invention.

[0045] Figure 1 This is a flow chart of the method for automatic classification of medical and health data provided by the present invention.

[0046] Figure 2 A flowchart for preprocessing the medical data set provided by the present invention.

[0047] Figure 3 This is a flow chart of the intelligent grading model training process provided by the present invention.

[0048] Figure 4 A flow chart of the intelligent grading model training for the artificial immune network provided by the present invention.

[0049] Figure 5 This is a structural block diagram of the medical and health data automatic classification device provided by the present invention. DETAILED DESCRIPTION

[0050] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0051] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present invention.

[0052] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or apparatus.

[0053] In this embodiment, a method for automatically grading medical and health data is provided. Figure 1 is a flowchart of a method for automatically grading medical and health data according to an embodiment of the present invention. Figure 1 As shown, including:

[0054] S100, obtaining a medical data set to be classified;

[0055] In the embodiment of the present invention, the medical data set to be classified may specifically be medical equipment data, medical detection data, and the like.

[0056] S200, preprocessing the medical data set to be classified to obtain a first-level classification element feature matrix of the medical data set;

[0057] Before identifying the medical data set, the medical data set needs to be preprocessed. In the embodiment of the present invention, Figure 2 As shown, specifically including:

[0058] S210, performing data preprocessing on the medical data set to be classified to obtain a medical processing data set;

[0059] In the embodiment of the present invention, since the obtained medical data sets may be of different types or some data may be missing, it is necessary to preprocess the data in the medical data sets in order to obtain complete medical data and facilitate subsequent hierarchical processing of the medical data.

[0060] In the embodiment of the present invention, preprocessing the medical data set to obtain medical processing data may specifically include:

[0061] Performing type conversion on the medical data set to obtain data of a uniform type;

[0062] Here, the data in the medical data set is converted into a type, which may specifically include converting different types of data into the same type.

[0063] The missing values ​​in the medical data set are supplemented to obtain complete medical treatment data.

[0064] Since the medical dataset includes a large amount of data, it is inevitable that there will be some missing data. In order to facilitate subsequent data processing, it is necessary to supplement the missing values ​​in the medical dataset. For example, mean filling, random filling or model-based filling can be used. The specific method of supplementing missing values ​​is well known to technicians in this field and will not be described here.

[0065] S220, processing the medical treatment data set according to the hierarchical elements respectively, and discretizing the processed data to obtain a first-level hierarchical element feature matrix of the medical data set;

[0066] The grading elements include qualitative grading elements and quantitative grading elements.

[0067] In order to reduce the amount of data for subsequent hierarchical processing and at the same time ensure the accuracy of data grading results, the medical treatment data needs to be compressed. Before compression, a grading element matrix needs to be constructed according to the grading elements so that the medical treatment data can be processed differently according to different grading elements, and a grading element feature matrix constructed according to the grading elements can be obtained. This ensures that the accuracy of the data grading results will not be affected after data compression.

[0068] It should be noted that the qualitative grading factors include fields, groups, regions, importance and security risks, and the quantitative grading factors include accuracy, scale, coverage and derived data.

[0069] In an embodiment of the present invention, the medical treatment data set is processed according to the classification elements respectively, and the processed data is discretized, including:

[0070] Qualitatively describing the medical treatment data set according to qualitative grading elements to obtain qualitative grading discrete values;

[0071] Calculating the medical treatment data set according to the quantitative grading elements to obtain data distribution results of different quantitative grading elements;

[0072] Discretize the data distribution results of different quantitative grading elements to obtain quantitative grading discrete values;

[0073] The qualitative grading discrete values ​​and the quantitative grading discrete values ​​are constructed as a grading element feature matrix of the medical data set.

[0074] It should be understood that the qualitative classification elements of the fields, groups, regions, importance and safety risks are qualitatively described. For example, the importance can be qualitatively described as three levels, with the first level being unimportant, the second level being important, and the third level being very important. After the qualitative description of the importance, it is processed into discrete values ​​and represented by two dimensions in a binary matrix, for example, 00 is used to represent the first level, 01 is used to represent the second level, and 10 is used to represent the third level. In addition, the safety risk can also be qualitatively described as an X-based level. If X is less than or equal to 4, it can still be represented by two dimensions in a two-level matrix to form a two-level matrix representing a safety fan. Similarly, for groups, for example, they can be divided into five groups: children (less than 12 years old), teenagers (13 to 17 years old), youth (18 to 40 years old), middle-aged (41 to 65 years old), and elderly (greater than 65 years old). These five groups can be represented by 3-bit codes in a binary matrix, 000, 001, 010, 011, 100. For example, the regions can be divided into five regions: North China, Central China, South China, West China and East China, which can also be represented by a two-level 3-digit code.

[0075] It can be seen that by qualitatively describing the above-mentioned qualitative grading elements and converting them into the encoding of the two-level matrix, qualitative grading discrete values ​​are formed, and these qualitative grading discrete values ​​can constitute the grading element feature matrix of the sample data set.

[0076] For the quantitative grading factors such as accuracy, scale, coverage and derivative data, accurate calculations can be performed based on the actual data to obtain the data distribution, and then the data distribution can be graded, so as to discretize these accurately calculated quantitative results. The discretized data is also qualitatively described in the same way as the qualitative grading factors mentioned above to form quantitative grading discrete values, so as to construct the grading factor feature matrix of the sample data set.

[0077] S300, compressing the first-level hierarchical element feature matrix obtained after preprocessing according to the association or characteristics of data distribution to obtain a second-level hierarchical element feature matrix having a mapping relationship with the first-level hierarchical element feature matrix;

[0078] In an embodiment of the present invention, the association or characteristics of data distribution can be specifically expressed by the similarity of data features. Data with a similarity higher than a threshold value is compressed. The second-level grading element feature matrix obtained after compression has a mapping relationship with the first-level grading element feature matrix, and the amount of data to be graded is greatly reduced.

[0079] S400. Input the second-level grading factor feature matrix into an intelligent grading model to obtain a grading result of the second-level grading factor feature matrix, wherein the intelligent grading model is a grading model trained based on historical medical data with grading labels.

[0080] In an embodiment of the present invention, the second-level grading element feature matrix of the medical data set obtained after the previous processing is input into the intelligent grading model, and the intelligent grading model performs automatic intelligent grading according to similarity to obtain the grading result of the second-level grading element feature matrix.

[0081] It should be understood that the intelligent grading model in the embodiment of the present invention may specifically be a grading model trained based on historical medical data with grading labels.

[0082] Specifically, Figure 3 As shown in the figure, the intelligent classification model is obtained by training the historical medical data with classification labels, including:

[0083] S410, obtaining historical medical data with classification labels;

[0084] Historical medical data can be manually classified in advance to set classification labels, and the classification elements can be used as a reference when setting specific classification labels.

[0085] S420, constructing a sample grading element feature matrix based on historical medical data with grading labels;

[0086] After preprocessing the historical medical data with classification labels, the sample qualitative classification discrete values ​​and the sample quantitative classification discrete values ​​can be obtained according to the classification elements. The sample qualitative classification discrete values ​​and the sample quantitative classification discrete values ​​are constructed as a sample classification element feature matrix.

[0087] S430, initializing the artificial immune network;

[0088] Specifically, the training is performed in an artificial immune network manner, where the artificial immune network is initialized to generate an initial antibody population and calculate its fitness value.

[0089] S440: training the sample classification factor feature matrix according to an artificial immune network to obtain an intelligent classification model.

[0090] In the embodiment of the present invention, Figure 4 As shown, the sample classification element feature matrix is ​​trained according to the artificial immune network to obtain an intelligent classification model, including:

[0091] S441, using N data classification element vectors in the classification element feature matrix of the sample data set as antigens, where N is a natural number greater than 1;

[0092] Specifically, antigen is represented by Ag, which corresponds to the N data classification element vectors x in the data set. i (i=1,2,…,N), available x i ∈S L Description, L represents the dimension of the data; when each dimension in the hierarchical feature matrix is ​​represented by a one-dimensional vector, L = 9.

[0093] S442, retaining antibodies with high fitness for the antigen as memory network data;

[0094] It should be noted that the artificial immune network describes the stress competition process exhibited by antibodies during antigen attack. Antibodies with high fitness will be retained as memory cells and continue to evolve until a dynamic equilibrium is reached. The embodiment of the present invention describes the antibodies in the artificial immune network as a set of typical sample data V1,...,V K . Antibodies Indicates that LK = L × K is the product of the data dimension and the number of solutions.

[0095] S443, iteratively searching for memory network data that meets the fitness condition according to the artificial immune network algorithm;

[0096] It should be noted that since fitness is an indicator of the quality of antibodies in artificial immune networks, antibodies with high fitness will be evolved and antibodies with low fitness will be eliminated. The smaller the objective function value is, the better the result is. In the algorithm based on artificial immune network, antibody Ab i The fitness calculation is shown in formula (1), J(Ab i , Ag) is antibody Ab i The corresponding objective function value. The smaller J is, the greater the fitness value of the antibody is, indicating that the solution encoded by the antibody is better. The value range of f is between [0, 1], then solving the minimum value of the objective function is transformed into a process in which the artificial immune network continuously evolves according to the fitness to find memory cells with high fitness.

[0097]

[0098] In order to prevent the network from retaining a large amount of identical or very similar data and to improve the execution speed of the algorithm, similar antibodies need to be deleted. The similarity calculation method between antibodies is shown in formula (2), where j , Ab k The Euclidean distance between .

[0099]

[0100] S444. Use the memory network data that meets the fitness condition as an intelligent graded label data set, and store the intelligent graded label data set.

[0101] It should be noted that the evolution of antibodies is achieved through the process of clonal mutation, and the number of clones of each antibody is N. c The dynamic regulation of the network follows the principle of immune network, including the increase of a new generation of memory cells, network compression between memory cells and dynamic renewal of network cells. % of high fitness antibodies are added to the artificial immune network as memory cells. In order to control the scale of the network, network compression is required. By setting the network compression threshold ts, the low fitness memory cells in the two memory cells with similarity less than the threshold are cleared to reduce the number of memory cells stored in the network and improve the execution speed of the algorithm. The dynamic update process of network cells uses randomly generated new network memory cells to replace d% of low fitness memory cells to increase the population diversity of the network. After continuous evolution, the memory cells in the artificial immune network are the retained intelligent hierarchical label data set.

[0102] Therefore, in an embodiment of the present invention, the second-level grading element feature matrix of the medical data set is input into the intelligent grading model to obtain the grading result of the second-level grading element feature matrix, including:

[0103] Performing fitness calculation on the second-level hierarchical element feature matrix of the medical data set and the intelligent hierarchical label data set;

[0104] Determining whether the fitness is greater than a preset threshold;

[0105] If it is greater than a preset threshold, the grading result is determined according to the intelligent grading label data set.

[0106] It should be understood that by performing fitness calculation on the second-level grading element feature matrix of the medical data set constructed above and the intelligent grading label data set, the grading result can be determined according to the fitness calculation result.

[0107] Specifically, when the fitness calculation result is greater than a preset threshold, the matching result in the intelligent grading label data set can be output as a grading result.

[0108] If it is not greater than a preset threshold, manually performing classification processing on the classification element feature matrix of the medical data set and obtaining a classification result;

[0109] The intelligent grading label data set is updated according to the grading results obtained by manual grading.

[0110] When the fitness is not greater than the preset threshold, manual grading is performed to obtain a grading result, and the grading result is updated to the intelligent grading label data set.

[0111] It should be understood that when the calculated fitness is not greater than the preset threshold, it is considered that there is a lack of matching results in the intelligent grading label data set. After the grading is completed by manual grading, the above intelligent grading label data set is updated to improve the intelligent grading label data set. This update method can gradually improve the intelligent grading label data set and provide efficient and accurate grading results for the subsequent intelligent grading of medical data.

[0112] S500, mapping the hierarchical labels of the hierarchical second-level hierarchical element feature matrix back to the first-level hierarchical element feature matrix according to the mapping relationship during compression, that is, mapping the hierarchical labels of the data in the second-level hierarchical element feature matrix to the data of the first-level hierarchical element feature matrix with a mutual mapping relationship, completing the grading of the first-level hierarchical element feature matrix, thereby greatly compressing the grading workload of the first-level hierarchical element feature matrix.

[0113] S600, added classification of medical data sets, including:

[0114] When a new medical data set is collected, the new medical data set is processed according to steps S200 and S300, that is, the newly added medical data set is processed according to the hierarchical elements respectively, and the processed data is discretized to obtain a first-level hierarchical element feature matrix of the newly added medical data set; then the first-level hierarchical element feature matrix of the newly added medical data set obtained after preprocessing is compressed to obtain a second-level hierarchical element feature matrix of the newly added medical data set having a mapping relationship with the first-level hierarchical element feature matrix;

[0115] Input the data of the second-level hierarchical element feature matrix of the newly added medical data set into the intelligent grading model for grading according to the similarity. If the similarity is higher than or equal to the threshold, the data in the second-level hierarchical element feature matrix adopts the grading label of the intelligent grading model, and the grading label of the second-level hierarchical element feature matrix after the grading of the newly added medical data set is mapped back to the first-level hierarchical element feature matrix according to the mapping relationship during compression;

[0116] The data with similarity below the threshold is manually graded and assigned with graded labels, then input into the intelligent grading model, stored as a new intelligent grading label dataset, and the intelligent grading model is updated.

[0117] In this step, for the classification process of the newly added medical data set, on the one hand, the data is compressed and then classified in the intelligent classification model to reduce the amount of data processing. On the other hand, the update of the old intelligent classification model is combined with manual classification update to make the update of the intelligent classification model more in line with the actual classification requirements and standards, and the classification is more accurate.

[0118] In summary, the automatic data classification method provided by the present invention preprocesses the original medical data set by constructing a matrix according to a classification factor matrix, and further compresses the preprocessed classification factor matrix, and then intelligently classifies the data constructed by the compressed matrix through an intelligent classification model, which greatly reduces the workload of grading and labeling data. The second-level classification factor matrix after classification maps the classification labels back to the first-level classification factor matrix through the mapping relationship with the first-level classification factor matrix, which can ensure the efficiency and accuracy of data classification, and the classification method can automatically perform data classification processing online, thereby realizing automatic data classification.

[0119] As another embodiment of the present invention, there is provided an automatic data classification device, which includes: a processor and a memory, the processor and the memory are communicatively connected, the memory stores at least one program instruction, and the processor is used to load and execute at least one program instruction to implement the automatic data classification method described above.

[0120] like Figure 5 As shown, the data automatic classification device 50 may include: at least one processor 51, such as a CPU (Central Processing Unit), at least one communication interface 53, a memory 54, and at least one communication bus 52. The communication bus 52 is used to realize the connection and communication between these components. The communication interface 53 may include a display screen (Display) and a keyboard (Keyboard), and the optional communication interface 53 may also include a standard wired interface and a wireless interface. The memory 54 may be a high-speed RAM memory (Random Access Memory, volatile random access memory) or a non-volatile memory (non-volatile memory), such as at least one disk storage. The memory 54 may also be at least one storage device located away from the aforementioned processor 51. The memory 54 stores an application program, and the processor 51 calls the program code stored in the memory 54 to execute any of the above method steps.

[0121] The communication bus 52 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The communication bus 52 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0122] Among them, the memory 54 may include a volatile memory (English: volatile memory), such as a random access memory (English: random-access memory, abbreviated: RAM); the memory may also include a non-volatile memory (English: non-volatile memory), such as a flash memory (English: flash memory), a hard disk drive (English: hard disk drive, abbreviated: HDD) or a solid-state drive (English: solid-state drive, abbreviated: SSD); the memory 54 may also include a combination of the above types of memory.

[0123] The processor 51 may be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and a NP.

[0124] The processor 51 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0125] Optionally, the memory 54 is also used to store program instructions. The processor 51 can call the program instructions to implement the present invention. Figure 1 The automatic data classification method shown in the embodiment.

[0126] As another embodiment of the present invention, a computer storage medium is provided, wherein the computer storage medium stores at least one program instruction, and the at least one program instruction is loaded and executed by a processor to implement the automatic data classification method described above.

[0127] In an embodiment of the present invention, a non-transitory computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions can execute the data automatic classification method in any of the above method embodiments. The storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk (HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above types of memories.

[0128] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for automatically grading medical and health data, characterized in that: include: Obtaining a medical data set to be classified; Preprocessing the medical data set to be classified to obtain a first-level classification element feature matrix of the medical data set; Compressing the first-level hierarchical element feature matrix of the medical data set according to the association or characteristics of data distribution to obtain a second-level hierarchical element feature matrix having a mapping relationship with the first-level hierarchical element feature matrix; Inputting the second-level grading factor feature matrix into an intelligent grading model to obtain a grading result of the second-level grading factor feature matrix, wherein the intelligent grading model is a grading model trained based on historical medical data with grading labels; The classification labels of the classified second-level classification feature matrix are mapped back to the first-level classification feature matrix according to the mapping relationship during compression, thereby completing the classification of the first-level classification feature matrix.

2. The method for automatic classification of medical and health data according to claim 1, characterized in that: Also included are the classifications of newly added medical datasets, including: When collecting a new medical data set, the newly added medical data set is processed according to the hierarchical elements respectively, and the processed data is discretized to obtain the first-level hierarchical element feature matrix of the newly added medical data set; the first-level hierarchical element feature matrix of the newly added medical data set obtained after preprocessing is compressed to obtain the second-level hierarchical element feature matrix of the newly added medical data set having a mapping relationship with the first-level hierarchical element feature matrix; Input the data of the second-level hierarchical element feature matrix of the newly added medical data set into the intelligent grading model for grading according to the similarity. If the similarity is higher than or equal to the threshold, the data in the second-level hierarchical element feature matrix adopts the grading label of the intelligent grading model, and the grading label of the second-level hierarchical element feature matrix after the grading of the newly added medical data set is mapped back to the first-level hierarchical element feature matrix according to the mapping relationship during compression; The data with similarity below the threshold is manually graded and assigned with graded labels, then input into the intelligent grading model, stored as a new intelligent grading label dataset, and the intelligent grading model is updated.

3. The method for automatic classification of medical and health data according to claim 1, characterized in that: The intelligent classification model is trained based on historical medical data with classification labels, including: Obtain historical medical data with hierarchical labels; Construct a sample hierarchical element feature matrix based on historical medical data with hierarchical labels; Initialize the artificial immune network; The sample classification element feature matrix is ​​trained according to an artificial immune network to obtain an intelligent classification model.

4. The automatic classification method for medical and health data according to claim 3, characterized in that: The sample classification element feature matrix is ​​trained according to an artificial immune network to obtain an intelligent classification model, including: Taking N data classification element vectors in the classification element feature matrix of the sample data set as antigens, where N is a natural number greater than 1; retaining antibodies with high fitness for the antigen as memory network data; Iteratively search for memory network data that meets fitness conditions based on the artificial immune network algorithm; The memory network data satisfying the fitness condition is used as an intelligent graded label data set, and the intelligent graded label data set is stored.

5. The method for automatic classification of medical and health data according to claim 2, characterized in that: Inputting the classification element feature matrix of the medical data set into the intelligent classification model to obtain the classification result of the medical data set, including: Performing fitness calculation on the hierarchical element feature matrix of the medical data set and the intelligent hierarchical label data set; Determining whether the fitness is greater than a preset threshold; If it is greater than a preset threshold, the grading result of the medical data set is determined according to the intelligent grading label data set.

6. The method for automatic classification of medical and health data according to claim 3, characterized in that: If it is not greater than a preset threshold, manually performing classification processing on the classification element feature matrix of the medical data set and obtaining a classification result of the medical data set; The intelligent grading label dataset is updated according to the grading results of the medical dataset obtained by manual grading.

7. The method for automatic classification of medical and health data according to any one of claims 1 to 6, characterized in that: The medical data set to be classified is preprocessed to obtain the classification element feature matrix of the medical data set, including: Performing data preprocessing on the medical data set to be classified to obtain a medical processing data set; The medical treatment data set is processed according to the hierarchical elements respectively, and the processed data is discretized to obtain the hierarchical element feature matrix of the medical data set. The grading elements include qualitative grading elements and quantitative grading elements.

8. The method for automatic classification of medical and health data according to claim 5, characterized in that: The medical treatment data set is processed according to the classification elements respectively, and the processed data is discretized, including: Qualitatively describing the medical treatment data set according to qualitative grading elements to obtain qualitative grading discrete values; Calculating the medical treatment data set according to the quantitative grading elements to obtain data distribution results of different quantitative grading elements; Discretize the data distribution results of different quantitative grading elements to obtain quantitative grading discrete values; The qualitative grading discrete values ​​and the quantitative grading discrete values ​​are constructed as a grading element feature matrix of the medical data set.

9. A medical health data automatic classification device, characterized in that: include: A processor and a memory, wherein the processor and the memory are communicatively connected, wherein at least one program instruction is stored in the memory, and wherein the processor is used to load and execute at least one program instruction to implement the automatic data classification method according to any one of claims 1 to 8.

10. A computer storage medium, characterized in that: The computer storage medium stores at least one program instruction, which is loaded and executed by the processor to implement the automatic data classification method described in any one of claims 1 to 8.

Images