Traditional Chinese medicine syndrome type auxiliary distinguishing method, system and equipment and storage medium

Abstract

The invention discloses a traditional Chinese medicine syndrome type auxiliary distinguishing method, system and device and a storage medium, and the method comprises the steps: obtaining a symptom set of a to-be-distinguished medical case, and the symptom set comprises a plurality of candidate symptoms; obtaining a weight value from each candidate symptom to a syndrome element from a preset symptom and syndrome element weight value library, and accumulating the weight values to obtain an accumulated weight value of each syndrome element; performing optimization processing on the cumulative weighted value of each syndrome element according to a preset optimization rule base to obtain an optimization weight value of each syndrome element, and generating a weight value optimization vector; and inputting the weight optimization vector into a pre-trained certificate name identification model to obtain the certificate name of the medical case to be identified. According to the invention, the universality and robustness of syndrome element weight calculation are improved, and the accuracy and efficiency of syndrome type discrimination are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of syndrome differentiation, and in particular to a method, system, device and storage medium for auxiliary differentiation of traditional Chinese medicine syndrome types. Background Art

[0002] TCM diagnosis and treatment can be divided into three processes: syndrome differentiation, formulation and prescription. Syndrome differentiation is the process in which doctors comprehensively judge the patient's current state based on the four diagnostic information and obtain the patient's TCM syndrome type. It is a key step in determining the entire TCM diagnosis and treatment. The traditional TCM syndrome differentiation and identification method first presets the decision weight of each symptom for each syndrome type in advance, then calculates the score of each syndrome type, and takes the syndrome type with a higher score as the output of the system, but the accuracy of the results obtained by this method is not high.

[0003] However, the current syndrome differentiation mainly relies on expert knowledge to calculate syndrome weights, and the syndrome weights are fixed. It does not take into account the adjustment of syndrome weights due to changes in different diagnostic scenarios, which will lead to deviations in the inferred TCM syndrome types.

[0004] Based on this, how to improve the accuracy of TCM syndrome identification is a technical problem that urgently needs to be solved. Summary of the invention

[0005] The main purpose of the present invention is to provide a method, system, device and storage medium for auxiliary identification of TCM syndrome types, aiming to solve the above-mentioned technical problems in the background technology.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a method for auxiliary identification of TCM syndrome types, comprising:

[0007] Acquire a symptom set of a medical case to be identified, wherein the symptom set includes a plurality of candidate symptoms;

[0008] Obtaining the weight of each candidate symptom to each syndrome from a preset symptom and syndrome weight library, and obtaining the accumulated weight of each syndrome by accumulating the weights;

[0009] Optimizing the accumulated weight of each of the certificate elements according to a preset optimization rule base to obtain an optimized weight of each of the certificate elements and generate a weight optimization vector;

[0010] The weight optimization vector is input into a pre-trained syndrome name recognition model to obtain the syndrome name of the medical case to be identified.

[0011] In a second aspect, the present invention further provides a TCM syndrome type auxiliary identification system, comprising:

[0012] A symptom acquisition module, used to acquire a symptom set of the medical case to be identified, wherein the symptom set includes a plurality of symptoms;

[0013] A syndrome factor weight calculation module, used to obtain the weight of each symptom to syndrome factor from a preset symptom and syndrome factor weight library, and obtain the accumulated weight of each syndrome factor by accumulating the weights;

[0014] The certificate element weight optimization module is used to optimize the accumulated weight of each certificate element according to a preset optimization rule base, obtain the optimized weight of each certificate element, and generate a certificate element weight vector;

[0015] The certificate name identification module inputs the weight optimization vector into a pre-trained certificate name identification model to obtain the certificate name of the medical case to be identified.

[0016] In a third aspect, the present invention further provides an electronic device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the method for auxiliary identification of TCM syndrome types as described in the first aspect above is implemented.

[0017] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for auxiliary identification of TCM syndrome types as described in the first aspect above is implemented.

[0018] The above-mentioned TCM syndrome-aided identification method, system, device and storage medium, for the medical case to be identified, first obtains the weight of each candidate symptom to each syndrome in the medical case to be identified from the symptom and syndrome weight library, and accumulates the weights according to the syndrome, and then optimizes the syndrome weights using the optimization rule library, which can avoid the impact of the imbalance of the number of symptoms on the weights and improve the universality and robustness of the syndrome weight calculation; further, the weight optimization vector is processed using a pre-trained syndrome name recognition model to obtain the syndrome name of the medical case to be identified, thereby improving the accuracy and efficiency of syndrome type identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a diagram of the application environment of the TCM syndrome-assisted identification method according to an embodiment of the present invention;

[0020] Figure 2 It is a schematic diagram of the flow chart of an embodiment of a method for auxiliary identification of TCM syndrome types of the present invention;

[0021] Figure 3 It is a flow chart of another embodiment of the TCM syndrome-aided identification method of the present invention;

[0022] Figure 4 It is a schematic diagram of the flow chart of an embodiment of a TCM syndrome type auxiliary identification system of the present invention;

[0023] Figure 5 It is a flow chart of another embodiment of the TCM syndrome type auxiliary identification system of the present invention;

[0024] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment.

[0025] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0027] The TCM syndrome-aided identification method provided by the embodiment of the present invention can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process, such as medical record data and certificate element weights. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, and tablet computers. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0028] Those skilled in the art will understand that Figure 1 The terminal structure shown in the figure does not constitute a limitation on the terminal, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0029] Reference Figure 2 , provides a method for auxiliary identification of TCM syndromes, which can be applied to Figure 1 Taking the server in the example as an example, the TCM syndrome type auxiliary identification method specifically includes the following steps:

[0030] Step S10, obtaining a symptom set of the medical case to be identified, wherein the symptom set includes a number of candidate symptoms.

[0031] In this embodiment, the medical case to be identified refers to the medical case that needs to be identified by syndrome name; the syndrome name refers to the syndrome type name given by experts based on comprehensive symptoms and syndrome elements.

[0032] Specifically, a medical case to be identified is obtained, and candidate symptoms in the medical case to be identified are identified through natural language processing technology to form a symptom set of the medical case to be identified.

[0033] It should be noted that in this embodiment, extracting the target object in a text by using natural language processing technology is a common method in the art and will not be described in detail here.

[0034] Step S20, obtaining the weight of each candidate symptom to each syndrome from a preset symptom and syndrome weight library, and obtaining the accumulated weight of each syndrome by accumulating the weights.

[0035] In this embodiment, the symptom and syndrome weight library refers to a database storing the weights of standard symptoms to syndromes. The standard symptoms adopt 637 symptoms in "Syndrome Differentiation", and the syndrome adopts 53 syndromes in "Syndrome Differentiation", including 20 syndromes of disease location (mind (or brain), heart, lung, spleen, liver, kidney, stomach, gallbladder, small intestine, large intestine, bladder, uterus, sperm chamber, chest diaphragm, lower abdomen, surface, half surface and half interior, skin, meridians, tendons and bones, etc.) and 33 syndromes of disease nature (external wind, cold, cold blood, heat, hot blood, summer heat, dryness, dampness, phlegm, drink, water retention, qi stagnation, qi stagnation, blood stasis, pus, poison, worm accumulation, food accumulation, yang hyperactivity, qi deficiency, qi collapse, qi instability, loss, blood deficiency, yang deficiency, yang loss, yang floating, yin deficiency, fluid deficiency, yin loss, essence deficiency, and wind movement, blood movement, etc. 33).

[0036] Specifically, for each candidate symptom in the symptom set, firstly, a standard symptom matching the candidate symptom is retrieved from the preset symptom and syndrome weight library to be determined as the target symptom, and then the weight of the target symptom to each syndrome is extracted, and the weights are accumulated according to the syndromes, so as to obtain all syndromes of the medical case to be identified and the corresponding accumulated weights. Furthermore, the syndrome weight vector of the medical case to be identified can be generated according to all syndromes of the medical case to be identified and the corresponding accumulated weights, which is specifically expressed as Among them, x i is the weight of the i-th prime, i∈[1,…,K], K is the prime number, and its value is 53.

[0037] It is understandable that for a candidate symptom in the symptom set, if the weight of the candidate symptom to a syndrome is not extracted from the symptom and syndrome weight library, the weight of the candidate symptom to the syndrome is determined to be 0 to achieve the purpose of completing the syndrome weight.

[0038] Step S30, optimizing the accumulated weight of each of the certificate elements according to a preset optimization rule base, obtaining the optimized weight of each of the certificate elements, and generating a weight optimization vector.

[0039] In this embodiment, the preset optimization rule base refers to a database for storing the optimization values ​​of the certificate elements; the optimization values ​​of the certificate elements are set manually.

[0040] More specifically, firstly, all syndrome elements of the medical case to be identified are marked according to the preset disturbance syndrome elements, and then the syndrome elements with the mark are down-weighted or up-weighted according to the preset optimization rule base, and the original weights of the syndrome elements without the mark are retained, so as to obtain the optimized weights of all syndrome elements of the medical case to be identified, and finally, the weight optimization vector of the medical case to be identified is generated according to the optimized weights of all syndrome elements. Among them, the disturbance syndrome element refers to the syndrome element that brings great interference to the weight determination of the symptom-to-symptom element.

[0041] Step S40, inputting the weight optimization vector into a pre-trained syndrome name recognition model to obtain the syndrome name of the medical case to be identified.

[0042] In this embodiment, the certificate name recognition model adopts a fully connected neural network model, which takes the weight of each certificate element in the weight optimization vector as input and outputs the certificate name.

[0043] That is, the weight optimization vector obtained after the optimization processing in step S30 is input into the pre-trained certificate name recognition model for processing, so as to obtain the certificate name of the medical case to be identified.

[0044] In summary, the TCM syndrome-assisted identification method of this embodiment, for the medical case to be identified, first obtains the weight of each candidate symptom to syndrome in the medical case to be identified from the symptom and syndrome weight library, and accumulates the weights according to the syndrome, and then optimizes the syndrome weights using the optimization rule library, which can avoid the impact of the imbalance of the number of symptoms on the weights, improve the universality and robustness of the syndrome weight calculation, and provide more accurate data for subsequent syndrome identification; further, the weight optimization vector is processed using a pre-trained syndrome name recognition model to obtain the syndrome name of the medical case to be identified, thereby improving the accuracy and efficiency of syndrome identification.

[0045] Further, see Figure 3 Another embodiment of the present invention provides a method for auxiliary identification of TCM syndromes. Figure 2 In the embodiment shown, the TCM syndrome-aided identification method further includes the following steps:

[0046] Step S50, obtaining a medical case data set including a plurality of medical cases, each of the medical cases including a plurality of symptoms, a plurality of syndrome elements and at least one expert syndrome name.

[0047] In this embodiment, 5125 medical records are obtained from the TCM medical record database to form a medical record data set, which includes medical records of TCM internal medicine, surgery, gynecology, and pediatrics, which can be expressed as C = {c1, c2, ..., c M}, M is the size of the medical record data set. Further, all symptoms can be extracted from the medical record data set to generate a total set of symptoms, which can be expressed as S = {s1, s2, …, s N}, N is the size of the total symptom set, N = 637, and all syndrome elements are extracted from the medical record data set to generate the total syndrome element set, which can be expressed as E = {e1, e2, …, e K}, K is the size of the total set of evidence elements, K=53.

[0048] For any medical case c in the medical case data set, it includes multiple symptoms, multiple syndrome elements and at least one expert certificate name, which can be expressed as c = (sc, ec, dc), sc is the symptom set of the medical case c, which can be expressed as sc = {s1, s2, ..., s i}, ec is the syndrome element set of the medical case c, which can be expressed as ec = {e1, e2, ..., e k}, dc is the set of certificate names of the medical case c, which can be expressed as dc = {d1, d2, ..., d j Among them, the TCM medical case database is cited from the TCM medical case database disclosed by Yan Junfeng in the article "Research on Syndrome Differentiation Method Based on Data Mining Technology".

[0049] Step S60, generating a symptom and syndrome network according to the relationship between symptoms and syndrome elements in the medical case data set; the relationship between symptoms and syndrome elements refers to whether the symptoms and syndrome elements appear in the same medical case.

[0050] In this embodiment, the symptom and syndrome element network refers to a network with symptoms as symptom type nodes, syndrome elements as syndrome element type nodes, and the relationship between symptoms and syndrome elements as edges. Preferably, step S60 specifically includes the following steps:

[0051] Step S601, obtaining the first medical case in the medical case data set;

[0052] Step S602, determining each symptom in the first medical case as a first node in the network to be constructed, and determining each syndrome element in the first medical case as a second node in the network to be constructed;

[0053] Step S603, generating an edge between the first node and the second node to obtain an initial symptom and syndrome factor network;

[0054] Step S604, obtaining the next medical case in the medical case data set, and performing node and edge updates on the initial symptom and syndrome network according to the symptoms and syndromes in the next medical case, until the edge updates are completed according to all the medical cases, thereby obtaining the final symptom and syndrome network.

[0055] That is, first, each symptom and each syndrome element in a medical case is used as a symptom type node (i.e., the first node) and syndrome element type node (i.e., the second node) in the network, and the symptom type nodes and syndrome element type nodes are connected in pairs to establish edges. Secondly, all medical cases in the medical case data set C are constructed with edges between symptom type nodes and syndrome element type nodes in the above manner, and finally a symptom and syndrome element network is obtained, which can be expressed as:

[0056] Formula (1)

[0057] In formula (1), SE graph is a symptom and syndrome network, c = (sc, ec, sc) is a medical case in the medical case set C, if (s i ,e j ) appear in the same medical record c, then its degree is 1, indicating that they are connected by an edge; otherwise, it is 0, indicating that they are not connected by an edge.

[0058] It should be noted that the Symptom and Syndrome Network SE graph All edges in are undirected edges.

[0059] It can be understood that, in this embodiment, through the above steps S601 to S604, the purpose of quickly generating a symptom and syndrome factor network can be achieved.

[0060] Step S70, according to the symptom and syndrome factor network, obtain the dynamic weights of the symptoms to the syndrome factors through multi-layer normalization processing, and generate a symptom and syndrome factor weight library.

[0061] In this embodiment, multi-layer normalization processing refers to multi-level calculation of weights, including global weights, syndrome-factor normalization weights, and symptom normalization weights. Preferably, step S70 specifically includes the following steps:

[0062] Step S701, calculating the global weight of the symptom to the syndrome element according to the degree of the edge between the symptom and the syndrome element in the symptom and syndrome element network.

[0063] First, determine the symptom s from the symptom and syndrome network i E j The edge (s i ,e j ) and use the edge degree as the symptom s i E j The initial weights between, then the symptoms s i E j The initial weights between , calculate the symptoms s in the network i To the certificate element j The global weight of , the calculation formula of the global weight can be expressed as:

[0064] Formula (2-1)

[0065] Formula (2-2)

[0066] In formula (2), |sc k | represents the prime number of the kth medical case, M is the size of the medical case dataset, Indicates symptom type node s i With the certificate type node e j If the degree between the current medical case c k The degree is 1 if there is an edge, otherwise it is 0.

[0067] It can be understood that this embodiment can reduce the impact of the imbalance in the number of symptoms on the weight by calculating the global weight.

[0068] Step S702, according to the global weight from the symptom to the syndrome factor, respectively calculate the syndrome factor normalized weight and the symptom normalized weight of the symptom and the syndrome factor.

[0069] First, based on the global weight, the normalized weights of the certificate elements in their respective categories are calculated according to different certificate element categories to reduce the negative impact of the direct accumulation of global weights. The calculation formula for the normalized weights of the certificate elements is:

[0070] Formula (3-1)

[0071] Formula (3-1)

[0072] In formula (3), Indicates symptoms i To the certificate element j The normalized weight of the certificate element is Symptoms i To the certificate element j The global weight of Indicates and evidence element j All relevant edges (s i ,e j ) is the maximum value among the global weights.

[0073] Then, the global weight is converted and calculated from the symptom perspective to obtain the symptom normalized weight. The calculation formula of the symptom normalized weight is:

[0074] Formula (4-1)

[0075] Formula (4-2)

[0076] In formula (4), Indicates symptoms i To the certificate element j The symptom normalized weights, Symptoms i To the certificate element j The global weight of Signs and symptoms i All relevant edges (s i ,e j ) is the maximum value among the global weights.

[0077] It is understandable that this embodiment can reduce the excessive influence of the global weight part through the syndrome-factor normalization weight calculation, and can avoid the limitations of syndrome-factor normalization weight calculation through the symptom normalization weight calculation. Therefore, this embodiment can effectively suppress the influence of the imbalance of medical case samples and the imbalance of the number of symptoms on the weight determination through multi-layer normalization processing.

[0078] Step S703, performing weighted averaging processing on the syndrome normalized weight and the symptom normalized weight to obtain a comprehensive weight from the symptom to the syndrome.

[0079] In step S703, the calculation formula of the comprehensive weight is:

[0080]

[0081] In formula (5), is the comprehensive weight of the symptom to the syndrome factor; f1 is the weight of the syndrome-factor normalized weight of the symptom and the syndrome factor; f2 is the weight of the symptom and the syndrome-factor normalized weight of the syndrome; f1 and f2 are set according to demand, and optionally, the values ​​of f1 and f2 are both 100%.

[0082] It can be understood that, in this embodiment, after weighted averaging processing, the weights of the symptoms to the syndrome factors are all within a reasonable range.

[0083] Step S704, generating a symptom and syndrome weight library according to the comprehensive weights from the symptoms to the syndrome elements.

[0084] It is understandable that in other embodiments, the syndrome-element normalized weights and symptom normalized weights of the symptom to syndrome element are obtained in step S702, and a symptom and syndrome element weight library can be generated based on the syndrome-element normalized weights and symptom normalized weights. When the weights of the candidate symptoms to each syndrome element are subsequently obtained from the symptom and syndrome element weight library, the syndrome-element normalized weights and symptom normalized weights of the candidate symptoms to each syndrome element can be obtained, and then the syndrome-element normalized weights and symptom normalized weights of the candidate symptoms to each syndrome element are weighted averaged to obtain a comprehensive weight, and then the weights are accumulated to obtain the final weight of each syndrome element.

[0085] Step S80, based on the dynamic weights from the symptoms to the syndrome elements, syndrome element statistics are performed on each medical case in the medical case data set to obtain all syndrome elements and corresponding values ​​in each medical case, and generate a syndrome element value vector for each medical case.

[0086] More specifically, according to the syndrome normalization weight and the symptom normalization weight, syndrome statistics are performed on each medical case in the medical case data set to obtain all syndromes and corresponding values ​​in each medical case, and a corresponding syndrome value vector is generated according to all syndromes and corresponding values ​​in each medical case. The syndrome value vector is specifically expressed as:

[0087] Formula (6-1)

[0088] Formula (6-2)

[0089] In formula (6), w m is the numerical vector of the syndrome element calculated for the mth medical case in the medical case data set; m∈[1,…,M], M is the size of the medical case data set, and its value is 5125; j∈[1,…,K], K is the total number of syndrome elements in the medical case data set (i.e., the size of the total syndrome element set), and its value is 53; w j is the prime vector w m The value calculated for the jth certificate element in ; Indicates symptoms i To the certificate element j Symptom normalized weights; Symptoms i To the certificate element j The normalized weight of the syndrome factors; |sc| represents the number of symptoms in the mth medical case.

[0090] Step S90, optimizing the syndrome element numerical vector of each medical case according to a preset optimization rule library to obtain the syndrome element numerical optimization vector of each medical case.

[0091] It is understandable that when the correlation degree and global weight of the syndrome factor are very large, it will cause great interference to the determination of the weight of the symptom to the syndrome factor. In order to reduce this influence, the optimization value of each syndrome factor is pre-set manually, which is specifically expressed as:

[0092] Formula (7),

[0093] In formula (7), represents the optimized vector of the certificate element, x i represents the optimized value of the i-th certificate element, K represents the size of the total set of certificate elements, and its value is 53.

[0094] Furthermore, the weight optimization vector is used to reduce and increase the weight of the syndrome factor calculation values ​​that cause disturbances in the results, and the final result of the medical case syndrome factor weight calculation is obtained as follows:

[0095] Formula (8)

[0096] In formula (8), represents the optimal vector of the certificate value, Represents a numeric vector of prime numbers.

[0097] Step S110, iteratively training a fully connected neural network model according to the certificate element numerical optimization vector of each medical case in the medical case data set and the expert certificate name to obtain a certificate name recognition model.

[0098] More specifically, first, the optimized vector of the certificate element value of each medical case is used as a training sample to construct a training data set, and then an initialized fully connected neural network model is obtained. The initialized fully connected neural network model takes the certificate element value of each medical case as input and the judgment result of whether it is an expert certificate as output. The number of nodes in the input layer is K, the number of nodes in the output layer is 1, and the number of nodes in the hidden layer is:

[0099] Formula (9)

[0100] In formula (9), N i is the number of neurons in the input layer, N o is the number of neurons in the output layer, N s is the number of samples in the training data set, and α∈[2,10] is a custom variable. The activation functions of the first two hidden layers of the fully connected neural network model are ReLU function and sigmoid function respectively.

[0101] Finally, the fully connected neural network model is trained according to the training data set until the loss error calculated based on the expert certificate name label and the certificate name output by the model reaches the preset value, or reaches the preset maximum number of iterations, and the fully connected neural network model after training is obtained and determined as the certificate name recognition model.

[0102] It should be noted that using a training data set to train a fully connected neural network model to obtain the required network model is a common method in this field and will not be described in detail here.

[0103] In summary, the TCM syndrome-assisted identification method of the present embodiment, before step S10, first generates a symptom and syndrome network based on the naive relationship between symptoms and syndrome elements in the medical case data set, and then obtains the dynamic weights from symptoms to syndrome elements through multi-layer normalization processing based on the symptom and syndrome network, and further uses the dynamic weights from symptoms to syndrome elements to perform syndrome element statistics on each medical case in the medical case data set, and uses a preset optimization rule library to optimize the weights by downgrading or upgrading, thereby improving the universality and robustness of the syndrome element weight calculation, and finally uses the syndrome element weight optimization vector of each medical case to train the fully connected neural network model to obtain a syndrome name recognition model, thereby improving the performance of the syndrome name recognition model.

[0104] In a preferred embodiment, when the symptom-syndrome weight database contains the comprehensive weights from symptoms to syndromes, step S20 specifically includes the following steps:

[0105] Step S201, obtaining the comprehensive weight of each symptom to syndrome factor from the symptom and syndrome factor weight library.

[0106] Step S202: according to the comprehensive weight, the weights are accumulated according to the certificate elements to obtain the accumulated weight of each certificate element.

[0107] In step S202, the accumulated weight of each certificate element can be expressed as:

[0108] Formula (10)

[0109] In formula (10), w j is the cumulative weight of the jth certificate element, Indicates symptoms i To the certificate element j The comprehensive weight of ; |sc| represents the size of the symptom set.

[0110] It can be understood that the syndrome factor weights obtained by the above steps S201 and S202 in this embodiment can reduce the impact of the imbalance of the number of symptoms on the weight determination, and improve the universality and robustness of the syndrome factor weight calculation.

[0111] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0112] Based on the same inventive concept, the present invention also provides a TCM syndrome type auxiliary identification system, which is used to implement the TCM syndrome type auxiliary identification method involved above. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more TCM syndrome type auxiliary identification system embodiments provided below can refer to the limitations of the TCM syndrome type auxiliary identification method above, and will not be repeated here.

[0113] Reference Figure 5 , based on the above Figure 1 In the embodiment shown, the medical syndrome auxiliary identification system includes a symptom acquisition module 110, a syndrome element weight calculation module 120, a syndrome element weight optimization module 130 and a syndrome name identification module 140. The detailed description of each functional module is as follows:

[0114] The symptom acquisition module 110 is used to acquire a symptom set of the medical case to be identified, wherein the symptom set includes a plurality of symptoms;

[0115] The syndrome weight calculation module 120 is used to obtain the weight of each symptom to each syndrome from a preset symptom and syndrome weight library, and obtain the accumulated weight of each syndrome by accumulating the weights;

[0116] The certificate element weight optimization module 130 is used to optimize the accumulated weight of each certificate element according to a preset optimization rule base, obtain the optimized weight of each certificate element, and generate a weight optimization vector;

[0117] The certificate name identification module 140 inputs the weight optimization vector into a pre-trained certificate name identification model to obtain the certificate name of the medical case to be identified.

[0118] Further, refer to Figure 6The TCM syndrome auxiliary identification system also includes a medical case data acquisition module 150, a network generation module 160, a weight library generation module 170, a syndrome element statistics module 180, an optimization module 190 and a model training module 200. The detailed description of each functional module is as follows:

[0119] The medical case data acquisition module 150 is used to acquire a medical case data set including a plurality of medical cases, each of which includes a plurality of symptoms, a plurality of syndrome elements and at least one expert syndrome name;

[0120] The network generation module 160 is used to generate a symptom-syndrome network according to the relationship between the symptoms and syndrome elements in the medical case data; the relationship between the symptoms and syndrome elements refers to whether the symptoms and syndrome elements appear in the same medical case;

[0121] A weight library generation module 170 is used to obtain the dynamic weights of the symptoms to the syndrome elements through multi-layer normalization processing according to the symptom and syndrome element network, and generate a symptom and syndrome element weight library;

[0122] The syndrome element statistics module 180 is used to perform syndrome element statistics on each medical case in the medical case data set according to the dynamic weights from the symptoms to the syndrome elements, obtain all syndrome elements and corresponding values ​​in each medical case, and generate a syndrome element value vector for each medical case;

[0123] An optimization module 190, configured to optimize the syndrome element numerical vector of each medical case according to a preset optimization rule base to obtain the syndrome element numerical optimization vector of each medical case;

[0124] The model training module 200 is used to train a fully connected neural network model based on the certificate element numerical optimization vector of each medical case in the medical case data set and the expert certificate name to obtain a certificate name recognition model.

[0125] Furthermore, the network generation module 160 includes the following submodules, and the detailed description of each functional submodule is as follows:

[0126] A medical record acquisition submodule, used to acquire the first medical record in the medical record data set;

[0127] A node determination submodule, used to determine each symptom in the first medical case as a first node in the network to be constructed, and to determine each syndrome element in the first medical case as a second node in the network to be constructed;

[0128] An initial network generation submodule, used to generate an edge between the first node and the second node to obtain an initial symptom and syndrome factor network;

[0129] The network update submodule is used to obtain the next medical case in the medical case data set, and perform node update and edge update on the initial symptom and syndrome network according to the symptoms and syndrome elements in the next medical case, until the edge update is completed according to all the medical cases, so as to obtain the final symptom and syndrome network.

[0130] Furthermore, the weight library generation module 170 includes the following submodules, and the detailed description of each functional submodule is as follows:

[0131] A global weight calculation submodule, used for calculating the global weight of the symptom to the syndrome element according to the degree of the edge between the symptom and the syndrome element in the symptom-syndrome element network;

[0132] A normalized weight processing submodule, used for respectively calculating the syndrome-element normalized weights of the symptoms and the syndrome-element and the symptom normalized weights according to the global weights from the symptoms to the syndrome-element;

[0133] A comprehensive weight processing submodule, used for performing weighted average processing on the syndrome factor normalized weight and the symptom normalized weight to obtain a comprehensive weight from the symptom to the syndrome factor;

[0134] The weight library generation submodule is used to generate a symptom and syndrome factor weight library based on the comprehensive weights from the symptoms to the syndrome factors.

[0135] Furthermore, the syndrome statistics module 180 is also used to perform syndrome statistics on each medical case in the medical case data set according to the syndrome normalization weight and the symptom normalization weight.

[0136] Furthermore, the certificate weight calculation module 120 includes the following submodules, and the detailed description of each functional submodule is as follows:

[0137] A comprehensive weight extraction submodule, used for obtaining the comprehensive weight of each candidate symptom to each syndrome element from the symptom and syndrome element weight library;

[0138] The weight accumulation submodule is used to accumulate the weights according to the certificate elements according to the comprehensive weight to obtain the accumulated weight of each certificate element.

[0139] Each module in the above-mentioned TCM syndrome auxiliary identification system can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0140] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor 1001, a memory 1005, a user interface 1003 and a network interface 1004 connected via a system bus 1002. Among them, the processor 1001 of the computer device is used to provide computing and control capabilities. The memory 1005 of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store medical record data, syndrome element weights, etc. The network interface 1004 of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for auxiliary identification of Chinese medicine syndrome types is implemented.

[0141] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0142] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of the method for auxiliary identification of TCM syndrome types as described in any of the above embodiments are implemented.

[0143] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the TCM syndrome-assisted identification method as described in any of the above embodiments are implemented.

[0144] The specific embodiments of the computer device and computer-readable storage medium of the present invention are basically the same as the embodiments of the above-mentioned TCM syndrome-assisted identification method, and will not be described in detail here.

[0145] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method or system. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method or system including the element.

[0146] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0147] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0148] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for auxiliary identification of TCM syndromes, characterized in that: include: Acquire a symptom set of a medical case to be identified, wherein the symptom set includes a plurality of candidate symptoms; Obtaining the weight of each candidate symptom to each syndrome from a preset symptom and syndrome weight library, and obtaining the accumulated weight of each syndrome by accumulating the weights; Optimizing the accumulated weight of each of the certificate elements according to a preset optimization rule base to obtain an optimized weight of each of the certificate elements and generate a weight optimization vector; The weight optimization vector is input into a pre-trained syndrome name recognition model to obtain the syndrome name of the medical case to be identified.

2. The TCM syndrome-aided identification method according to claim 1, characterized in that: Also includes: Acquire a medical case data set including multiple medical cases, each of which includes multiple symptoms, multiple syndrome elements and at least one expert syndrome name; According to the relationship between symptoms and syndrome elements in the medical case data set, a symptom and syndrome element network is generated; the relationship between symptoms and syndrome elements refers to whether the symptoms and syndrome elements appear in the same medical case; According to the symptom and syndrome element network, a dynamic weight from the symptom to the syndrome element is obtained through multi-layer normalization processing, and a symptom and syndrome element weight library is generated; According to the dynamic weights from the symptoms to the syndrome elements, syndrome element statistics are performed on each medical case in the medical case data set to obtain all syndrome elements and corresponding values ​​in each medical case, and generate syndrome element value vectors for each medical case; Optimizing the syndrome element numerical vector of each medical case according to a preset optimization rule library to obtain the syndrome element numerical optimization vector of each medical case; According to the numerical optimization vector of the syndrome element of each medical case in the medical case data set and the expert syndrome name, a fully connected neural network model is iteratively trained to obtain a syndrome name recognition model.

3. The TCM syndrome-aided identification method according to claim 2, characterized in that: The generating of a symptom and syndrome network according to the relationship between symptoms and syndrome elements in the medical record data set includes: Obtaining the first medical case in the medical case data set; Determine each symptom in the first medical case as a first node in the network to be constructed, and determine each syndrome element in the first medical case as a second node in the network to be constructed; Generate an edge between the first node and the second node to obtain an initial symptom and syndrome network; The next medical case in the medical case data set is obtained, and nodes and edges of the initial symptom and syndrome network are updated according to the symptoms and syndromes in the next medical case, until the edge update is completed according to all the medical cases, so as to obtain the final symptom and syndrome network.

4. The TCM syndrome-aided identification method according to claim 3, characterized in that: According to the symptom and syndrome element network, the dynamic weights of the symptoms to the syndrome elements are obtained through multi-layer normalization processing, and a symptom and syndrome element weight library is generated, including: Calculating the global weight of the symptom to the syndrome element according to the degree of the edge between the symptom and the syndrome element in the symptom-syndrome element network; According to the global weight from the symptom to the syndrome factor, respectively calculating the syndrome factor normalized weight and the symptom normalized weight of the symptom and the syndrome factor; Performing weighted average processing on the normalized weight of the syndrome factor and the normalized weight of the symptom to obtain a comprehensive weight from the symptom to the syndrome factor; According to the comprehensive weights from the symptoms to the syndrome factors, a symptom and syndrome factor weight library is generated.

5. The TCM syndrome-aided identification method according to claim 4, characterized in that: The step of performing syndrome statistics on each medical case in the medical case data set according to the dynamic weights from the symptoms to the syndrome factors includes: According to the syndrome factor normalized weight and the symptom normalized weight, syndrome factor statistics are performed on each medical case in the medical case data set; the calculation formula of the syndrome factor statistics is: Formula(1-1) Formula(1-2) In formula (1), w m is the numerical vector of the syndrome element of the mth medical case in the medical case dataset; m∈[1,…,M], M is the size of the medical case dataset; j∈[1,…,K], K is the total number of syndrome elements in the medical case dataset; w j is the prime vector w m The statistical value of the jth certificate element in ; Indicates symptoms i To the certificate element j Symptom normalized weights; Symptoms i To the certificate element j The normalized weight of the syndrome factors; |sc| represents the number of symptoms in the mth medical case.

6. The TCM syndrome-aided identification method according to claim 4, characterized in that: The step of obtaining the weight of each candidate symptom to each syndrome from a preset symptom and syndrome weight library, and obtaining the accumulated weight of each syndrome by accumulating the weights, includes: Obtaining the comprehensive weight of each candidate symptom to each syndrome from the symptom and syndrome weight library; According to the comprehensive weight, the weights of the certificate elements are accumulated respectively to obtain the accumulated weight of each certificate element.

7. A TCM syndrome-type auxiliary identification system, characterized in that: include: A symptom acquisition module, used to acquire a symptom set of the medical case to be identified, wherein the symptom set includes a plurality of candidate symptoms; A syndrome weight calculation module, used for obtaining the weight of each candidate symptom to each syndrome from a preset symptom and syndrome weight library, and obtaining the accumulated weight of each syndrome by accumulating the weights; The certificate element weight optimization module is used to optimize the accumulated weight of each certificate element according to a preset optimization rule base, obtain the optimized weight of each certificate element, and generate a certificate element weight vector; The certificate name identification module inputs the weight optimization vector into a pre-trained certificate name identification model to obtain the certificate name of the medical case to be identified.

8. The TCM syndrome-type auxiliary identification system according to claim 7, characterized in that: Also includes: A medical case data acquisition module, used to acquire a medical case data set containing multiple medical cases, each of which contains multiple symptoms, multiple syndrome elements and at least one expert syndrome name; A network generation module, used for generating a symptom and syndrome network according to the relationship between symptoms and syndrome elements in the medical case data set; the relationship between symptoms and syndrome elements refers to whether the symptoms and syndrome elements appear in the same medical case; A weight library generation module is used to obtain the dynamic weights of the symptoms to the syndrome elements through multi-layer normalization processing according to the symptom and syndrome element network, and generate a symptom and syndrome element weight library; A syndrome factor statistics module, for performing syndrome factor statistics on each medical case in the medical case data set according to the dynamic weights from the symptoms to the syndrome factors, obtaining all syndrome factors and corresponding values ​​in each medical case, and generating a syndrome factor value vector for each medical case; An optimization module, used for optimizing the syndrome element numerical vector of each medical case according to a preset optimization rule library to obtain the syndrome element numerical optimization vector of each medical case; The model training module is used to train a fully connected neural network model based on the certificate element numerical optimization vector of each medical case in the medical case data set and the expert certificate name to obtain a certificate name recognition model.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the method for auxiliary identification of TCM syndrome types as described in any one of claims 1 to 6 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for auxiliary identification of TCM syndrome types according to any one of claims 1 to 6 is implemented.

Images