Essence model construction method and system based on big data

Abstract

The invention relates to the technical field of medical data processing, in particular to a syndrome element model construction method and system based on big data, and the method comprises the steps: obtaining an original syndrome information set and a diagnosis information set; analyzing each piece of original syndrome information in the original syndrome information set to obtain an original keyword set; obtaining a plurality of original clustering groups; if the number of the original clustering groups is smaller than a first preset number threshold value, obtaining a plurality of diagnosis keywords corresponding to each piece of diagnosis information; determining a matching degree between diagnosis information corresponding to each diagnosis keyword and a plurality of pieces of original syndrome information corresponding to the plurality of intermediate keywords; and based on a preset conversion method, converting the diagnosis information of which the matching degree meets a preset condition into corresponding original syndrome information to update the original syndrome information set, and training the initial syndrome element model based on the updated original syndrome information set to obtain a target syndrome element model. The accuracy of the syndrome element model is improved by increasing the training samples of the syndrome element model.

Description

Technical Field

[0001] The present invention relates to the field of medical data processing technology, and in particular to a method and system for constructing a certificate factor model based on big data. Background Art

[0002] Syndrome differentiation based on syndrome elements is a method in the process of syndrome differentiation in traditional Chinese medicine. It determines the relevant syndrome elements by analyzing the patient's clinical symptoms and signs, and then combines the syndrome elements to form syndrome name terms that conform to traditional Chinese medicine habits to guide clinical prescription and medication. In the context of the big data era, using data mining technology to build a syndrome element model can improve the accuracy and efficiency of syndrome differentiation, and also lay the foundation for the modernization and informatization of traditional Chinese medicine.

[0003] In the study, the application of data mining technology is key, including support vector machine (SVM), rough set theory, neural network and other methods. These methods are used to analyze clinical medical record data to explore the acquisition of syndrome differentiation rules and quantitative analysis of the impact of symptoms on syndromes. For example, experimental analysis of clinical medical record data by support vector machine and rough set algorithm can explore the feasibility of these methods in TCM syndrome differentiation modeling. In addition, researchers have also tried to use TCM big data for objective, standardized and quantitative research in order to integrate the knowledge gained from the long-term clinical experience of TCM practitioners. However, there are certain limitations to the method based entirely on data-driven. Therefore, researchers propose to use standardized interpretable models to combine clinical knowledge with artificial intelligence to achieve the functional integration of modern medicine and TCM.

[0004] In practical applications, researchers collect and organize a large number of TCM clinical research literature, and use data mining techniques such as the April algorithm to mine the rules of syndrome combination, providing a basis for big data analysis for TCM syndrome standardization research. At the same time, through the Bayesian network, the diagnosis rules of TCM syndrome based on conditional probability are mined, providing a syndrome matching paradigm for the syndrome differentiation model implemented in the TCM health management system. However, model construction requires a large number of training samples for training, and insufficient training samples will lead to low accuracy and generalization ability of the syndrome model. Summary of the invention

[0005] To this end, the present invention provides a method and system for constructing a certificate factor model based on big data, so as to overcome the problems of low accuracy and generalization ability of the certificate factor model caused by insufficient training samples in the prior art.

[0006] To achieve the above objectives, on the one hand, the present invention provides a method for constructing a certificate element model based on big data, comprising:

[0007] Step S1, obtaining an original syndrome information set and a diagnosis information set, wherein the original syndrome information set includes a number of original syndrome information, and the diagnosis information set includes a number of diagnosis information;

[0008] Step S2, analyzing each original syndrome information in the original syndrome information set, determining a number of original keywords of each original syndrome information, to obtain an original keyword set;

[0009] Step S3, clustering the original keyword set to obtain a number of original cluster groups;

[0010] Step S4, if the number of the original cluster groups is less than a first preset number threshold, each piece of diagnostic information in the diagnostic information set is analyzed to obtain a number of diagnostic keywords corresponding to each piece of diagnostic information;

[0011] Step S5, determining a number of original keywords corresponding to each diagnostic keyword based on a preset matching method, and using them as intermediate keywords, and determining the matching degree between the diagnostic information corresponding to each diagnostic keyword and a number of original syndrome information corresponding to the intermediate keywords;

[0012] Step S6, based on a preset conversion method, converting the diagnostic information whose matching degree meets the preset conditions into corresponding original syndrome information to update the original syndrome information set, and training the initial syndrome factor model based on the updated original syndrome information set to obtain the target syndrome factor model.

[0013] Furthermore, the step S3 further includes:

[0014] Step S30: If the number of the original cluster groups is greater than or equal to a first preset number threshold, the initial syndrome factor model is trained based on the original keyword set and the original syndrome information set to obtain a target syndrome factor model.

[0015] Furthermore, the step S4 comprises:

[0016] Step S41, if the number of the original cluster groups is less than a first preset number threshold, extracting a first number of diagnostic information from the diagnostic information set as intermediate diagnostic information, and determining an intermediate diagnostic information set;

[0017] Step S42: Analyze each intermediate diagnosis information in the intermediate diagnosis information set to obtain a plurality of intermediate diagnosis keywords corresponding to each intermediate diagnosis information.

[0018] Furthermore, in the step S41, the diagnostic information in the diagnostic information set is clustered to obtain a plurality of diagnostic cluster groups, and the first number is determined according to the number of diagnostic cluster groups.

[0019] Further, in the step S41, the first quantity is determined by:

[0020] If the number of the diagnostic cluster groups is greater than the number of the original cluster groups, the number of diagnostic information extracted from each diagnostic cluster group is determined according to the ratio of the number of diagnostic information in each diagnostic cluster group to the total number of diagnostic information in the diagnostic information set; wherein the sum of the number of diagnostic information extracted from each diagnostic cluster group is the first number;

[0021] If the number of the diagnostic cluster groups is less than or equal to the number of the original cluster groups, the first number is determined based on the diagnostic cluster groups in which the amount of diagnostic information in each of the diagnostic cluster groups is greater than a second preset number threshold; wherein the sum of the amounts of diagnostic information in the diagnostic cluster groups corresponding to the amounts of diagnostic information in each of the diagnostic cluster groups greater than the second preset number threshold is the first number.

[0022] Furthermore, in step S5, the preset matching method includes:

[0023] According to the preset original keyword list, the original keywords corresponding to each of the original cluster groups are compared to determine a number of important keywords; the important keywords are any original keywords other than the original keywords corresponding to each of the original cluster groups in the preset original keyword list;

[0024] The diagnostic keyword corresponding to each important keyword is determined according to the preset original keyword comparison table.

[0025] Furthermore, in step S5, determining the matching degree between the diagnosis information corresponding to each diagnosis keyword and the plurality of original syndrome information corresponding to the plurality of intermediate keywords includes:

[0026] Step S51, obtaining a first keyword vector according to a plurality of diagnostic keywords of the diagnostic information corresponding to each of the diagnostic keywords;

[0027] Step S52, obtaining a plurality of second keyword vectors according to a plurality of original keywords of each original syndrome information corresponding to each intermediate keyword;

[0028] Step S53: determining the matching degree between the first keyword vector and each of the second keyword vectors to obtain a plurality of first matching degrees.

[0029] Further, in step S6, if the first matching degree is greater than a preset matching degree threshold, it is determined that the preset condition is met.

[0030] Furthermore, in step S6, the preset conversion method includes:

[0031] Based on the preset original keyword comparison table, each diagnostic keyword of the diagnostic information is converted into a corresponding original keyword, and filled into the corresponding slot of the preset syndrome template to obtain the original syndrome information corresponding to the diagnostic information.

[0032] On the other hand, the present invention also provides a certificate element model construction system, comprising:

[0033] An information acquisition module, used to acquire an original syndrome information set and a diagnosis information set, wherein the original syndrome information set includes a number of original syndrome information, and the diagnosis information set includes a number of diagnosis information;

[0034] an analysis and determination module, which is connected to the information acquisition module and is used to analyze each original syndrome information in the original syndrome information set, determine a number of original keywords for each original syndrome information, and obtain an original keyword set;

[0035] a cluster comparison module, which is connected to the analysis and determination module and the information acquisition module respectively, and is used to cluster the original keyword set to obtain a plurality of original cluster groups; if the number of the original cluster groups is less than a first preset number threshold, each diagnostic information in the diagnostic information set is analyzed to obtain a plurality of diagnostic keywords corresponding to each diagnostic information;

[0036] A matching determination module, which is connected to the cluster comparison module and the information acquisition module respectively, and is used to determine a number of original keywords corresponding to each diagnostic keyword based on a preset matching method, and use them as intermediate keywords to determine the matching degree between the diagnostic information corresponding to each diagnostic keyword and a number of original syndrome information corresponding to the intermediate keywords;

[0037] An update training module is respectively connected to the matching determination module and the information acquisition module, and is used to convert the diagnostic information whose matching degree meets the preset conditions into corresponding original syndrome information based on a preset conversion method, so as to update the original syndrome information set, and train the initial syndrome factor model based on the updated original syndrome information set to obtain the target syndrome factor model.

[0038] Compared with the prior art, the beneficial effect of the present invention is that the present invention obtains the original keyword set by analyzing the original syndrome information in the original syndrome information set, which can improve the accuracy and efficiency of subsequent data processing. By clustering the original keyword set, a number of original clustering groups are obtained. If the number of original clustering groups is small, it indicates that the data in the original syndrome information set is not comprehensive and the training samples are relatively small. Therefore, the data is supplemented by the diagnostic information in the diagnostic information set, so as to improve the accuracy and generalization ability of constructing the syndrome element model. By analyzing the diagnostic information, a number of diagnostic keywords corresponding to each diagnostic information are determined, and the original syndrome information corresponding to the diagnostic keyword is determined according to the original keyword corresponding to the diagnostic keyword. The diagnostic information that meets the conditions is converted into the corresponding original syndrome information through the matching degree between the original syndrome information and the diagnostic information to update the original syndrome information set, which can improve the accuracy of the training sample conversion, avoid introducing noise to reduce the accuracy of the model, and thus increase the training samples to improve the accuracy and generalization ability of the syndrome element model.

[0039] Furthermore, in the present invention, if the number of original clustering groups is greater than or equal to a first preset number threshold, it indicates that the types of original syndrome information in the original syndrome information set are relatively comprehensive and can meet the construction requirements of the syndrome factor model. Therefore, the initial syndrome factor model can be directly trained based on the original keyword set and the original syndrome information set, which can improve the efficiency of model construction.

[0040] Furthermore, in the present invention, if the number of original clustering groups is less than a first preset number threshold, indicating that the types of original syndrome information in the original syndrome information set are not comprehensive enough and do not meet the construction requirements of the syndrome factor model, a first number of diagnostic information is extracted from the diagnostic information set for data supplementation, which can reduce the subsequent data processing volume and improve the efficiency of model construction.

[0041] Furthermore, the present invention can improve data processing efficiency by converting each diagnostic keyword of the diagnostic information into a corresponding original keyword based on a preset original keyword comparison table and filling the original keyword into a corresponding slot of a preset syndrome template. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of the method for constructing a certificate element model based on big data of the present invention;

[0043] Figure 2 Schematic diagram of the process of step S4 of the embodiment of the present invention;

[0044] Figure 3 Schematic diagram of the process of step S5 of the embodiment of the present invention;

[0045] Figure 4 This is a structural block diagram of the certificate model construction system of the present invention. DETAILED DESCRIPTION

[0046] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0047] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0048] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0049] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0050] See also Figure 1 As shown, it is a flow chart of a method for constructing a certificate element model based on big data of the present invention. The present invention provides a method for constructing a certificate element model based on big data, comprising:

[0051] Step S1, obtaining an original syndrome information set and a diagnosis information set, wherein the original syndrome information set includes a number of original syndrome information, and the diagnosis information set includes a number of diagnosis information;

[0052] Step S2, analyzing each original syndrome information in the original syndrome information set, determining a number of original keywords of each original syndrome information, to obtain an original keyword set;

[0053] In implementation, the syndrome in the present invention refers to the description of the reaction state of the essential organic connection such as the location, cause, nature, course of disease and the strength of the body's disease resistance at a certain stage in the disease process by traditional Chinese medicine. A case includes corresponding syndrome information, wherein there may be multiple reaction states in the syndrome information, that is, multiple syndromes, and the original keyword may be the syndrome element corresponding to each syndrome information, for example: the original syndrome information is "fever, chills, headache, body pain, no sweat, floating and tight pulse, thin white tongue coating", and the corresponding original keyword (syndrome element) is "wind-cold exterior"; the original syndrome information is "weak and cold pain in the waist and knees, fear of cold and cold limbs, pale or dark complexion, fatigue, frequent urination at night, slow and weak pulse, pale and fat tongue with white coating", and the corresponding original keyword (syndrome element) is "kidney yang deficiency".

[0054] In actual application, syndrome factors include disease location factors and disease properties factors. By presetting 19 disease location factor dimensions: exterior, semi-exterior and semi-interior, heart, spirit, liver, gallbladder, spleen, stomach, lung, chest and diaphragm, large intestine, kidney, bladder, anus, uterus, seminal chamber, skin, meridians, tendons and bones / joints, and 27 disease properties factor dimensions: wind, cold, heat, dampness, dryness, fire, damp heat, food accumulation, worm accumulation, yang deficiency, loss of yang, yin deficiency, loss of yin, qi deficiency, qi collapse, unstable qi, qi loss, qi stagnation, qi reversal, qi closure, blood deficiency, blood heat, blood stasis, fluid deficiency, water retention, drink, and phlegm, 190 syndrome factors are obtained by cross-reference, for example: qi stagnation and blood stasis, kidney yang deficiency, kidney yin deficiency, spleen dampness, etc.

[0055] Step S3, clustering the original keyword set to obtain a number of original cluster groups;

[0056] It should be noted that those skilled in the art are aware that any method capable of performing data clustering in the prior art falls within the protection scope of the present invention and will not be described in detail herein.

[0057] Specifically, the step S3 further includes:

[0058] Step S30: If the number of the original cluster groups is greater than or equal to a first preset number threshold, the initial syndrome factor model is trained based on the original keyword set and the original syndrome information set to obtain a target syndrome factor model.

[0059] In implementation, actual implementers may set the first preset quantity threshold based on the preset number of certificate elements according to actual conditions. Preferably, the first preset quantity threshold is set between 160 and 180.

[0060] It should be noted that those skilled in the art are aware that any model in the prior art that can output original keywords (syndrome elements) by training input of original syndrome information falls within the protection scope of the present invention and will not be described in detail here.

[0061] In the present invention, if the number of original clustering groups is greater than or equal to the first preset number threshold, it indicates that the types of original syndrome information in the original syndrome information set are relatively comprehensive and can meet the construction requirements of the syndrome factor model. Therefore, the initial syndrome factor model based on the original keyword set and the original syndrome information set can be directly trained, which can improve the model construction efficiency and reduce the amount of data processing.

[0062] Step S4, if the number of the original cluster groups is less than a first preset number threshold, each piece of diagnostic information in the diagnostic information set is analyzed to obtain a number of diagnostic keywords corresponding to each piece of diagnostic information;

[0063] See also Figure 2 As shown, it is a schematic diagram of the process of step S4 of an embodiment of the present invention. Specifically, step S4 includes:

[0064] Step S41, if the number of the original cluster groups is less than a first preset number threshold, extracting a first number of diagnostic information from the diagnostic information set as intermediate diagnostic information, and determining an intermediate diagnostic information set;

[0065] Step S42: Analyze each intermediate diagnosis information in the intermediate diagnosis information set to obtain a plurality of intermediate diagnosis keywords corresponding to each intermediate diagnosis information.

[0066] In practice, the diagnostic information may be examination results during the diagnosis of medical conditions, doctor's diagnosis and treatment information, patient's self-reported symptoms, etc.

[0067] In the present invention, if the number of original clustering groups is less than a first preset number threshold, indicating that the types of original syndrome information in the original syndrome information set are not comprehensive enough and do not meet the construction requirements of the syndrome factor model, a first number of diagnostic information is extracted from the diagnostic information set for data supplementation, which can reduce the subsequent data processing volume and improve the efficiency of model construction.

[0068] Specifically, in step S41, the diagnostic information in the diagnostic information set is clustered to obtain a plurality of diagnostic cluster groups, and the first number is determined according to the number of diagnostic cluster groups.

[0069] Specifically, in step S41, the first quantity is determined in the following manner:

[0070] If the number of the diagnostic cluster groups is greater than the number of the original cluster groups, the number of diagnostic information extracted from each diagnostic cluster group is determined according to the ratio of the number of diagnostic information in each diagnostic cluster group to the total number of diagnostic information in the diagnostic information set; wherein the sum of the number of diagnostic information extracted from each diagnostic cluster group is the first number;

[0071] In implementation, if there are W pieces of diagnostic information in the diagnostic information set, after clustering, n diagnostic cluster groups are obtained, and the number of diagnostic information in each diagnostic cluster group is K1, K2, ..., K i , …, K n , the number of original cluster groups is a, then the number of clusters extracted in the i-th diagnostic cluster group is R i =K i ×(K i / W), the first number P = ∑ n i＝1 R i ; Among them, n>a, i=1, 2,..., n; K i is the number of diagnostic information in the ith diagnostic cluster group; if there is a decimal, the extracted number is the integer part plus 1. For example, if there are 50 diagnostic information in the diagnostic information set, 5 diagnostic cluster groups are obtained after clustering, and the number of diagnostic information in each diagnostic cluster group is 15, 10, 10, 10, and 5 respectively, and the number of original cluster groups is 3, then 15×(15 / 50)=4.5; 10×(10 / 50)=2; 10×(10 / 50)=2; 10×(10 / 50)=2; 5×(5 / 50)=0.5; the numbers extracted in each diagnostic cluster group are: 5, 2, 2, 2, and 1 respectively, then the first number is 5+2+2+2+1=12.

[0072] If the number of the diagnostic cluster groups is less than or equal to the number of the original cluster groups, the first number is determined based on the diagnostic cluster groups in which the amount of diagnostic information in each of the diagnostic cluster groups is greater than a second preset number threshold; wherein the sum of the amounts of diagnostic information in the diagnostic cluster groups corresponding to the amounts of diagnostic information in each of the diagnostic cluster groups greater than the second preset number threshold is the first number.

[0073] In actual application scenarios, actual implementers may set the second preset quantity threshold based on the quantity in the diagnosis information set and the original syndrome information set according to actual conditions.

[0074] Step S5, determining a number of original keywords corresponding to each diagnostic keyword based on a preset matching method, and using them as intermediate keywords, and determining the matching degree between the diagnostic information corresponding to each diagnostic keyword and a number of original syndrome information corresponding to the intermediate keywords;

[0075] Specifically, in step S5, the preset matching method includes:

[0076] According to the preset original keyword list, the original keywords corresponding to each of the original cluster groups are compared to determine a number of important keywords; the important keywords are any original keywords other than the original keywords corresponding to each of the original cluster groups in the preset original keyword list;

[0077] In implementation, actual implementers can set a preset original keyword list based on the 190 certificate elements obtained by cross-section, and determine the original keywords in the preset original keyword list except the original keywords corresponding to each original clustering group as important keywords.

[0078] The diagnostic keyword corresponding to each important keyword is determined according to the preset original keyword comparison table.

[0079] In actual application scenarios, actual implementers may set a preset original keyword comparison table based on the expert system according to actual conditions, wherein each original keyword in the preset original keyword comparison table has a number of corresponding diagnostic keywords.

[0080] In actual implementation, there may be several important keywords corresponding to the same diagnostic keyword.

[0081] See also Figure 3 As shown, it is a flow chart of step S5 of an embodiment of the present invention. Specifically, in step S5, determining the matching degree of the diagnosis information corresponding to each of the diagnosis keywords and the plurality of original syndrome information corresponding to the plurality of intermediate keywords includes:

[0082] Step S51, obtaining a first keyword vector according to a plurality of diagnostic keywords of the diagnostic information corresponding to each of the diagnostic keywords;

[0083] Step S52, obtaining a plurality of second keyword vectors according to a plurality of original keywords of each original syndrome information corresponding to each intermediate keyword;

[0084] Step S53: determining the matching degree between the first keyword vector and each of the second keyword vectors to obtain a plurality of first matching degrees.

[0085] In implementation, the first keyword vectors Y1, Y2, ..., Y j , …, Y m , the second keyword vector E1, E2, ..., E j ,…,E m ;

[0086] Then the first matching degree P1=(∑ m j＝1 Y j ×E j ) / (sqrt(∑ m j＝1 (Y j ) 2 )×sqrt(∑ m j＝1 (E j ) 2));

[0087] Wherein, j=1, 2, ..., m; sqrt() is a preset square root determination function.

[0088] Step S6, based on a preset conversion method, converting the diagnostic information whose matching degree meets the preset conditions into corresponding original syndrome information to update the original syndrome information set, and training the initial syndrome factor model based on the updated original syndrome information set to obtain the target syndrome factor model.

[0089] Specifically, in step S6, if the first matching degree is greater than a preset matching degree threshold, it is determined that the preset condition is met.

[0090] In actual application scenarios, actual implementers can set a preset matching threshold according to actual conditions. The larger the value of the preset matching threshold, the higher the matching requirement for the first keyword vector and the second keyword vector. Preferably, the value range of the preset matching threshold can be set to 0.8-0.95.

[0091] Specifically, in step S6, the preset conversion method includes:

[0092] Based on the preset original keyword comparison table, each diagnostic keyword of the diagnostic information is converted into a corresponding original keyword, and filled into the corresponding slot of the preset syndrome template to obtain the original syndrome information corresponding to the diagnostic information.

[0093] In actual application scenarios, actual implementers can set preset syndrome templates based on the expert system according to actual conditions.

[0094] In implementation, the preset conversion method may also include: converting each diagnostic keyword of the diagnostic information into a corresponding original keyword based on a preset original keyword comparison table, determining a number of original syndrome information corresponding to each original keyword based on the original syndrome information set, and if there is a certain original diagnostic information including all the original keywords after the conversion of each diagnostic keyword, then the original syndrome information is determined as the original syndrome information corresponding to the diagnostic information.

[0095] The present invention can improve data processing efficiency by converting each diagnostic keyword of diagnostic information into a corresponding original keyword based on a preset original keyword comparison table and filling the original keyword into a corresponding slot of a preset syndrome template.

[0096] The present invention analyzes the original syndrome information in the original syndrome information set to obtain the original keyword set, which can improve the accuracy and efficiency of subsequent data processing. By clustering the original keyword set, a number of original clustering groups are obtained. If the number of original clustering groups is small, it indicates that the data in the original syndrome information set is not comprehensive and the training samples are relatively small. Therefore, the data is supplemented by the diagnostic information in the diagnostic information set to improve the accuracy and generalization ability of constructing the syndrome element model. By analyzing the diagnostic information, a number of diagnostic keywords corresponding to each diagnostic information are determined, and the original syndrome information corresponding to the diagnostic keyword is determined according to the original keyword corresponding to the diagnostic keyword. The diagnostic information that meets the conditions is converted into the corresponding original syndrome information through the matching degree between the original syndrome information and the diagnostic information to update the original syndrome information set, which can improve the accuracy of the training sample conversion, avoid introducing noise to reduce the accuracy of the model, and thus increase the training samples to improve the accuracy and generalization ability of the syndrome element model.

[0097] See also Figure 4 As shown, it is a structural block diagram of the certificate element model construction system of the present invention. The present invention also provides a certificate element model construction system, including:

[0098] An information acquisition module, used to acquire an original syndrome information set and a diagnosis information set; the original syndrome information set includes a number of original syndrome information; the diagnosis information set includes a number of diagnosis information;

[0099] an analysis and determination module, which is connected to the information acquisition module and is used to analyze each original syndrome information in the original syndrome information set, determine a number of original keywords for each original syndrome information, and obtain an original keyword set;

[0100] a cluster comparison module, which is connected to the analysis and determination module and the information acquisition module respectively, and is used to cluster the original keyword set to obtain a plurality of original cluster groups; if the number of the original cluster groups is less than a first preset number threshold, each diagnostic information in the diagnostic information set is analyzed to obtain a plurality of diagnostic keywords corresponding to each diagnostic information;

[0101] A matching determination module, which is connected to the cluster comparison module and the information acquisition module respectively, and is used to determine a number of original keywords corresponding to each diagnostic keyword based on a preset matching method, and use them as intermediate keywords to determine the matching degree between the diagnostic information corresponding to each diagnostic keyword and a number of original syndrome information corresponding to the intermediate keywords;

[0102] An update training module is respectively connected to the matching determination module and the information acquisition module, and is used to convert the diagnostic information whose matching degree meets the preset conditions into corresponding original syndrome information based on a preset conversion method, so as to update the original syndrome information set, and train the initial syndrome factor model based on the updated original syndrome information set to obtain the target syndrome factor model.

[0103] Specifically, the certificate element model construction system provided by the present invention adopts the above-mentioned certificate element model construction method based on big data to achieve the same technical effect, which will not be repeated here.

[0104] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A method for constructing a certificate element model based on big data, characterized in that: include: Step S1, obtaining an original syndrome information set and a diagnosis information set, wherein the original syndrome information set includes a number of original syndrome information, and the diagnosis information set includes a number of diagnosis information; Step S2, analyzing each original syndrome information in the original syndrome information set, determining a number of original keywords of each original syndrome information, to obtain an original keyword set; Step S3, clustering the original keyword set to obtain a number of original clustering groups; Step S4, if the number of the original cluster groups is less than a first preset number threshold, each piece of diagnostic information in the diagnostic information set is analyzed to obtain a number of diagnostic keywords corresponding to each piece of diagnostic information; Step S5, determining a number of original keywords corresponding to each diagnostic keyword based on a preset matching method, and using them as intermediate keywords, and determining the matching degree between the diagnostic information corresponding to each diagnostic keyword and a number of original syndrome information corresponding to the intermediate keywords; Step S6, based on a preset conversion method, converting the diagnostic information whose matching degree meets the preset conditions into corresponding original syndrome information to update the original syndrome information set, and training the initial syndrome factor model based on the updated original syndrome information set to obtain the target syndrome factor model.

2. The method for constructing a certificate element model based on big data according to claim 1, characterized in that: The step S3 further comprises: Step S30: If the number of the original cluster groups is greater than or equal to a first preset number threshold, the initial syndrome factor model is trained based on the original keyword set and the original syndrome information set to obtain a target syndrome factor model.

3. The method for constructing a certificate element model based on big data according to claim 2, characterized in that: The step S4 comprises: Step S41, if the number of the original cluster groups is less than a first preset number threshold, extracting a first number of diagnostic information from the diagnostic information set as intermediate diagnostic information, and determining an intermediate diagnostic information set; Step S42: Analyze each intermediate diagnosis information in the intermediate diagnosis information set to obtain a plurality of intermediate diagnosis keywords corresponding to each intermediate diagnosis information.

4. The method for constructing a certificate element model based on big data according to claim 3, characterized in that: In the step S41, the diagnostic information in the diagnostic information set is clustered to obtain a plurality of diagnostic cluster groups, and the first number is determined according to the number of the diagnostic cluster groups.

5. The method for constructing a certificate element model based on big data according to claim 4, characterized in that: In the step S41, the first quantity is determined by: If the number of the diagnostic cluster groups is greater than the number of the original cluster groups, the number of diagnostic information extracted from each diagnostic cluster group is determined according to the ratio of the number of diagnostic information in each diagnostic cluster group to the total number of diagnostic information in the diagnostic information set; wherein the sum of the number of diagnostic information extracted from each diagnostic cluster group is the first number; If the number of the diagnostic cluster groups is less than or equal to the number of the original cluster groups, the first number is determined based on the diagnostic cluster groups in which the amount of diagnostic information in each of the diagnostic cluster groups is greater than a second preset number threshold; wherein the sum of the amounts of diagnostic information in the diagnostic cluster groups corresponding to the amounts of diagnostic information in each of the diagnostic cluster groups greater than the second preset number threshold is the first number.

6. The method for constructing a certificate element model based on big data according to claim 5, characterized in that: In step S5, the preset matching method includes: According to the preset original keyword list, the original keywords corresponding to each of the original cluster groups are compared to determine a number of important keywords; the important keywords are any original keywords other than the original keywords corresponding to each of the original cluster groups in the preset original keyword list; The diagnostic keyword corresponding to each important keyword is determined according to the preset original keyword comparison table.

7. The method for constructing a certificate element model based on big data according to claim 6, characterized in that: In the step S5, determining the matching degree between the diagnosis information corresponding to each diagnosis keyword and the plurality of original syndrome information corresponding to the plurality of intermediate keywords includes: Step S51, obtaining a first keyword vector according to a plurality of diagnostic keywords of the diagnostic information corresponding to each of the diagnostic keywords; Step S52, obtaining a plurality of second keyword vectors according to a plurality of original keywords of each original syndrome information corresponding to each intermediate keyword; Step S53: determining the matching degree between the first keyword vector and each of the second keyword vectors to obtain a plurality of first matching degrees.

8. The method for constructing a certificate element model based on big data according to claim 7, characterized in that: In step S6, if the first matching degree is greater than a preset matching degree threshold, it is determined that the preset condition is met.

9. The method for constructing a certificate element model based on big data according to claim 8, characterized in that: In step S6, the preset conversion method includes: Based on the preset original keyword comparison table, each diagnostic keyword of the diagnostic information is converted into a corresponding original keyword, and filled into the corresponding slot of the preset syndrome template to obtain the original syndrome information corresponding to the diagnostic information.

10. A certificate element model construction system, which adopts the certificate element model construction method based on big data according to any one of claims 1 to 9, characterized in that: include: An information acquisition module, used to acquire an original syndrome information set and a diagnosis information set, wherein the original syndrome information set includes a number of original syndrome information, and the diagnosis information set includes a number of diagnosis information; an analysis and determination module, which is connected to the information acquisition module and is used to analyze each original syndrome information in the original syndrome information set, determine a number of original keywords for each original syndrome information, and obtain an original keyword set; a cluster comparison module, which is connected to the analysis and determination module and the information acquisition module respectively, and is used to cluster the original keyword set to obtain a plurality of original cluster groups; if the number of the original cluster groups is less than a first preset number threshold, each diagnostic information in the diagnostic information set is analyzed to obtain a plurality of diagnostic keywords corresponding to each diagnostic information; A matching determination module, which is connected to the cluster comparison module and the information acquisition module respectively, and is used to determine a number of original keywords corresponding to each diagnostic keyword based on a preset matching method, and use them as intermediate keywords to determine the matching degree between the diagnostic information corresponding to each diagnostic keyword and a number of original syndrome information corresponding to the intermediate keywords; An update training module is respectively connected to the matching determination module and the information acquisition module, and is used to convert the diagnostic information whose matching degree meets the preset conditions into corresponding original syndrome information based on a preset conversion method, so as to update the original syndrome information set, and train the initial syndrome factor model based on the updated original syndrome information set to obtain the target syndrome factor model.

Images