Medical data grading method and device, electronic equipment and medium

Abstract

The invention relates to a medical data grading method and device, electronic equipment and a medium, and the method comprises the steps: constructing a causal diagram according to specified medical data, and the causal diagram comprises a plurality of causal chains; performing a causal quantification operation on the causal graph to obtain a quantification result of each causal chain; and if it is determined that the quantification result of the target causal chain in the multiple causal chains meets a preset condition, adjusting the grading result of the reason variable in the target causal chain, and outputting the adjusted grading result. According to the method and the device, the data grading accuracy can be improved to a certain extent by adjusting the grading result of the reason variable meeting the condition by using the quantification result of the causal chain.

Description

Technical Field

[0001] The present disclosure relates to the field of medical information technology, and in particular to a medical data classification method, device, electronic device, and medium. Background Art

[0002] With the popularization and rapid development of medical informatization, medical data currently has the characteristics of large data volume, multiple data types, high value, high latitude, diversity, uncertainty and uneven data distribution. Therefore, how to accurately and effectively classify medical data to fully tap its medical value is a technical problem that needs to be solved urgently. Summary of the Invention

[0003] To overcome the problems existing in the related art, the present disclosure provides a medical data classification method, device, electronic device and medium.

[0004] According to a first aspect of an embodiment of the present disclosure, a method for grading medical data is provided, the method comprising: constructing a causal graph according to specified medical data, wherein the causal graph includes a plurality of causal chains; Performing a causal quantization operation on the causal graph to obtain a quantization result of each causal chain; If it is determined that the quantified result of the target causal chain among the multiple causal chains meets the preset condition, the grading result of the causal variables in the target causal chain is adjusted, and the adjusted grading result is output.

[0005] Optionally, performing a causal quantization operation on the causal graph includes: An average causal effect value of each causal chain in the causal graph is obtained, and the average causal effect value is used as the quantification result.

[0006] Optionally, the method further includes: If the average causal effect value is greater than a preset threshold, it is determined that the quantification result meets a preset condition.

[0007] Optionally, if it is determined that the quantified result of the target causal chain among the multiple causal chains meets a preset condition, adjusting the grading results of the causal variables in the target causal chain includes: When it is determined that the average causal effect value of the target causal chain is greater than or equal to a first threshold value and less than a second threshold value, obtaining a grading result of the outcome variable in the target causal chain; The classification result of the cause variable is adjusted according to the classification result of the result variable.

[0008] Optionally, after adjusting the grading result of the cause variable according to the grading result of the effect variable, the method further comprises: If it is determined that there is literature evidence related to the target causal chain, the classification results of the causal variables are readjusted.

[0009] Optionally, if it is determined that the quantified result of the target causal chain among the multiple causal chains meets a preset condition, adjusting the grading results of the causal variables in the target causal chain includes: When it is determined that the average causal effect value of the target causal chain is greater than or equal to a second threshold, the ranking result of the causal variable is adjusted to the ranking threshold.

[0010] Optionally, constructing a causal graph based on specified medical data includes: constructing an initial causal graph based on specified medical data, and obtaining non-causal chains in the initial causal graph; The initial causal graph is screened based on the non-causal chain to obtain a target causal graph.

[0011] According to a second aspect of an embodiment of the present disclosure, a medical data classification apparatus is provided, the apparatus comprising: a construction module configured to construct a causal graph according to specified medical data, wherein the causal graph includes a plurality of causal chains; a quantization module configured to perform a causal quantization operation on the causal graph to obtain a quantization result of each causal chain; The adjustment module is configured to adjust the grading results of the causal variables in the target causal chain if it is determined that the quantified results of the target causal chain in the multiple causal chains meet the preset conditions, and output the adjusted grading results.

[0012] According to a third aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which computer program instructions are stored. When the program instructions are executed by a processor, the steps of the medical data classification method provided in the first aspect of the present disclosure are implemented.

[0013] According to a fourth aspect of the embodiments of the present disclosure, there is provided an electronic device, including: a memory having a computer program stored thereon; A processor is used to execute the computer program in the memory to implement the steps of the medical data classification method provided in the first aspect of the present disclosure.

[0014] After obtaining a causal graph through specified medical data, the present disclosure can causally quantify the causal graph to obtain a quantified result of each causal chain. On this basis, it is determined whether there is a target causal chain that meets preset conditions in the causal chain according to the quantified result. If it is determined that the target causal chain exists, the grading result of the causal variable in the target causal chain is adjusted, and the adjusted grading result is output, thereby achieving flexibility and accuracy in the grading of medical data.

[0015] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification. Together with the following detailed description, they are used to explain the present disclosure but do not constitute a limitation of the present disclosure. In the accompanying drawings: Figure 1 The figure is a flowchart of a medical data classification method according to an exemplary embodiment.

[0017] Figure 2 The figure is an example diagram of a cause-effect diagram in a medical data classification method according to an exemplary embodiment.

[0018] Figure 3 This is an example diagram of a process for constructing a cause-effect graph in a medical data hierarchical method according to an exemplary embodiment.

[0019] Figure 4 The figure is a flowchart showing another method for classifying medical data according to an exemplary embodiment.

[0020] Figure 5 This is an example diagram of screening an initial cause-effect diagram in another medical data classification method according to an exemplary embodiment.

[0021] Figure 6 This is an example diagram of grading incremental causal chains in another medical data grading method according to an exemplary embodiment.

[0022] Figure 7 This is a specific example diagram of obtaining a grading result in another medical data grading method according to an exemplary embodiment.

[0023] Figure 8 The figure is a block diagram of a medical data classification device according to an exemplary embodiment.

[0024] Figure 9 It is a block diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0025] The following describes the specific embodiments of the present disclosure in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to illustrate and explain the present disclosure and are not intended to limit the present disclosure.

[0026] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all possible embodiments consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.

[0027] In the description of the present disclosure, terms such as "first" and "second" are used to distinguish similar objects and are not necessarily to be understood as implying a specific order or precedence. In addition, in the description with reference to the accompanying drawings, the same reference numerals in different drawings represent the same elements unless otherwise indicated.

[0028] Although operations or steps are described in a particular order in the drawings in the embodiments of the present disclosure, this should not be understood as requiring that these operations or steps be performed in the particular order shown or in a serial order, or that all of the operations or steps shown be performed to obtain a desired result. In the embodiments of the present disclosure, these operations or steps may be performed serially; these operations or steps may also be performed in parallel; or some of these operations or steps may be performed.

[0029] Existing medical data classification methods cannot capture newly generated associations, and the sensitivity weights of each field in medical data are fixed and will not be dynamically adjusted as special causal chains emerge. This will, to a certain extent, reduce the privacy of important data.

[0030] Figure 1 FIG. 1 is a flow chart showing a method for grading medical data according to an exemplary embodiment. Figure 1 As shown, the method may include the following steps.

[0031] In step S110 , a causal graph is constructed according to the specified medical data.

[0032] In the disclosed embodiments, a causal graph may include multiple causal chains, which may include first-type causal chains and second-type causal chains. The first-type causal chains can be obtained through medical ontology injection, such as by retaining causal chains defined in SNOMED CT data (a widely used standardized clinical medical terminology system) or OMIM (an authoritative genetic disease database maintained by Johns Hopkins University). Here, the first-type causal chain may also be referred to as a basic causal chain.

[0033] Optionally, the second type of causal chain can be obtained by mining new associations through a causal discovery algorithm. For example, the embodiment of the present disclosure can mine new causal chains through a PC algorithm (Peter-Clark Algorithm) and / or an FCI algorithm (FastCausal Inference Algorithm).

[0034] The disclosed embodiments can select a corresponding causal discovery algorithm based on the actual situation of the specified medical data. For example, when the dimension of the specified medical data is determined to be higher than a preset dimension and there is no confounding data, the PC algorithm can be used as the causal discovery algorithm. For another example, when it is determined that the specified medical data contains unobserved variables or potential causal paths that need to be labeled, the FCI algorithm can be used as the causal discovery algorithm.

[0035] It should be noted that the designated medical data may include existing general medical data and medical data generated in real time, wherein the existing general medical data can be used to construct the first type of causal chain; the medical data generated in real time can be used to construct the second type of causal chain.

[0036] Here, the designated medical data may include multiple fields, each field may correspond to a basic information of the user, such as name, age, gender, address, ID card and other medical indicators, among which the multiple medical indicators may include red blood cell count, BRCA1 gene mutation, blood sugar content and R346T mutation, etc.

[0037] The causal diagram constructed by specifying medical data can be Figure 2 As shown, based on Figure 2 It is known that a causal diagram may include multiple causal chains, and each causal chain may include a cause variable and a result variable, wherein the cause variable is the cause in the causal chain, which may be an independent variable; the result variable is the result in the causal chain, which may be a dependent variable.

[0038] For example, Figure 2 The causal chain consisting of A and B in the example can be called the first causal chain, and the causal chain consisting of C and D can be called the second causal chain. A in the first causal chain can be the causal variable, such as a smoking history ("yes"), and B can be the outcome variable, such as cancer stage ("stage IV"). C in the second causal chain can be the causal variable, such as a BRCA1 gene mutation, and D can be the outcome variable, such as breast / ovarian cancer. Research has found that mutation carriers have a 5-8 times increased risk of cancer.

[0039] In order to better illustrate the construction process of the causal graph, the embodiment of the present disclosure provides the following Figure 3 The example diagram shown is based on Figure 3 As can be seen, the disclosed embodiments can first acquire raw medical data, preprocess it, and then perform a medical ontology injection operation. Constrained by prior causal factors, this generates a candidate set of causal graphs. Furthermore, the causal chains in the candidate set are verified by experts. If the expert verification passes, the final causal graph is generated. If the expert verification fails (rejects), it indicates that the corresponding causal chain does not conform to the actual situation and can be deleted.

[0040] At the same time, the embodiments of the present disclosure can monitor new causal chains through data-driven discovery and causal discovery algorithms. Among them, data-driven discovery can mine new associations through causal discovery algorithms.

[0041] For example, the causal chain can be: HER2 protein overexpression → increased effectiveness of trastuzumab treatment; persistent hyperglycemia (HbA1c>9%) → diabetic nephropathy. Studies have shown that every 1% increase in HbA1c → the risk of kidney disease increases by 40%.

[0042] It should also be noted that the same causal chain may have one or more causal variables. That is, one causal variable can correspond to one outcome variable, or multiple causal variables can correspond to one outcome variable. For example, in the first causal chain, the causal variable is A, and the outcome variable is B. In another example, in the second causal chain, the causal variables are C, D, and E, and the outcome variable is F.

[0043] In step S120 , a causal quantization operation is performed on the causal graph to obtain a quantization result of each causal chain.

[0044] As an optional approach, after obtaining a causal graph, embodiments of the present disclosure may perform a causal quantification operation on the causal graph to obtain a quantified result for each causal chain in the causal graph. Specifically, embodiments of the present disclosure may obtain a quantified result for each causal chain based on at least one of the following methods: average causal effect, conditional average causal effect, individual causal effect, instrumental variable method, and doubly robust method.

[0045] After obtaining the quantitative results of each causal chain, the embodiment of the present disclosure can determine whether there is a target causal chain among the multiple causal chains whose quantitative results meet the preset conditions. If it is determined that the quantitative results of the target causal chain among the multiple causal chains meet the preset conditions, the grading results of the cause variables in the target causal chain are adjusted, that is, step S130 is entered.

[0046] In step S130 , if it is determined that the quantified result of the target causal chain among the multiple causal chains meets the preset condition, the grading results of the causal variables in the target causal chain are adjusted, and the adjusted grading results are output.

[0047] As described above, when it is determined that a target causal chain exists in a causal chain whose quantified results meet preset conditions, embodiments of the present disclosure can adjust the grading results of the causal variables of the target causal chain that meets the preset conditions. Adjustment of the causal variable grading results can be based on the grading results of the outcome variables, or can also be based on the default grading results.

[0048] In some implementations, adjustments to the causal variable classification results can be determined based on the quantification results. Different quantification results may result in different corresponding adjustment methods. For example, if the quantification result of a first causal chain is A, and this quantification result A satisfies a first condition, the first adjustment method can be used to adjust the causal variable classification results of the first causal chain.

[0049] For another example, if the quantified result of the second causal chain is B, and the quantified result B satisfies the second condition, the second adjustment method can be used to adjust the grading result of the cause variable of the second causal chain.

[0050] It should be noted that before adjusting the grading results of the causal variables in the target causal chain, the embodiments of the present disclosure may first obtain an initial grading of the causal variables, which may be the default grading defined by SNOMED CT / OMIM. Based on this, the embodiments of the present disclosure may adjust the grading of the causal variables in the target causal chain from the initial grading to the target grading.

[0051] In addition, before adjusting the grading result of the causal variable in the target causal chain, the embodiment of the present disclosure may first determine whether the grading result of the causal variable is less than a preset grading level. If the grading result of the causal variable is determined to be less than the preset grading level, the embodiment of the present disclosure may adjust the grading result of the causal variable in the target causal chain. Conversely, if the grading result of the causal variable is determined to exceed the preset grading level, the embodiment of the present disclosure may maintain the grading result of the causal variable unchanged.

[0052] After obtaining a causal graph through specified medical data, the embodiment of the present disclosure can perform causal quantification on the causal graph to obtain a quantified result of each causal chain. On this basis, it is determined whether there is a target causal chain that meets preset conditions in the causal chain according to the quantified result. If it is determined that the target causal chain exists, the grading result of the causal variable in the target causal chain is adjusted, and the adjusted grading result is output, thereby achieving flexibility and accuracy in the grading of medical data.

[0053] Figure 4 is a flowchart showing another method for grading medical data according to an exemplary embodiment. Figure 4 As shown, the method may include the following steps.

[0054] In step S210 , a causal graph is constructed based on the specified medical data.

[0055] In the process of constructing a causal graph based on specified medical data, the embodiment of the present disclosure can first construct an initial causal graph and determine whether there is a non-causal chain in the initial causal graph. If it is determined that there is a non-causal chain in the initial causal graph, the initial causal graph is screened according to the non-causal chain to obtain a target causal graph.

[0056] A non-causal chain refers to a connection between events without a clear causal relationship. For example, medical insurance type and cancer stage are only statistically correlated, without a clear causal relationship, making this a non-causal chain or a false causal chain. Another example is a false correlation between hospital bed number and treatment prognosis. Data collection bias can lead to the appearance of an association between the two, but in reality, there is no clear causal relationship. Another example is a true causal chain, where a smoking history and cancer stage have a causal relationship.

[0057] These non-causal chains can lead to misleading false correlations. Related technologies rely primarily on statistical features for classification, which can lead to non-causally related fields being mistakenly identified as sensitive data. For example, due to a statistical correlation between "Medical Insurance Type" and "Cancer Stage" at a certain hospital, the medical insurance field was mistakenly identified as high-risk data. The hospital is located in a rural area, and most of its patients use the New Rural Cooperative Medical Scheme (NCMS). These patients delay seeking medical treatment for financial reasons, resulting in later cancer stages. Data analysis shows that among patients covered by the NCMS, 68% have advanced cancer (stages III-IV), while only 29% of patients covered by the Urban Employee Medical Insurance have advanced cancer.

[0058] Related technologies, based on the random forest algorithm, rank the feature importance score of "medical insurance type" in the top 3, causing the system to classify it as L3 sensitive data. Therefore, the correlation between medical insurance type and cancer stage is spurious and not causal. To reduce this impact, the embodiments of the present disclosure can screen for non-causal chains during the construction of the causal graph to remove spurious causal chains.

[0059] In the embodiment of the present disclosure, the acquisition process of the non-causal chain can be as follows: Figure 5 So, based on Figure 5 After acquiring the original data (specified medical data), the disclosed embodiments can intervene in the original data through the causal intervention module and generate counterfactual samples. This process can be referred to as a counterfactual verification process. The causal intervention module can be used to select causal variables and perform intervention operations.

[0060] For example, the counterfactual sample may be {“medical insurance type”: “urban”, “cancer stage”: “stage III”}.

[0061] It should be noted that the screening of the initial causal graph can be achieved through the outcome causal model (SCM). That is, by screening non-causal chains, the disclosed embodiments can eliminate misleading statistical correlations and classify data based on true causal chains. In other words, the disclosed embodiments can fundamentally distinguish between causal and statistical correlations in data, ensuring the medical rationality and dynamic adaptability of classification decisions.

[0062] After generating counterfactual samples, embodiments of the present disclosure can perform causal structure verification, which verifies the causal chain using medical rules. If the causal chain is determined to violate the rules, it is intercepted and recorded. Here, causal structure verification can be used to intercept generated samples that violate causal logic, such as BRCA1 wild-type and breast cancer positive.

[0063] Conversely, if the causal chain is determined to comply with the rules, then the verified causal chain is determined to comply with the counterfactual data. Medical rule verification can be achieved through the globally used standardized clinical medical terminology system (SNOMED CT logic), clinical guideline constraints, and temporal relationship verification.

[0064] The disclosed embodiments can intervene on non-causal variables while keeping the causal variables unchanged. For example, the intervention on the non-causal variable could be: do(Medical Insurance Type = Urban) while keeping the causal variable "Cancer Diagnosis" unchanged. Based on counterfactual verification, the disclosed embodiments can generate causally consistent counterfactual data to determine the robustness of the hierarchical model.

[0065] As an example, consider the input data set {"Medical Insurance Type":"New Rural Cooperative Medical System", "Cancer Stage":"Stage IV", "Smoking History":"Yes"}. Causal analysis reveals that "Medical Insurance Type" and "Cancer Stage" are only statistically correlated (correlation coefficient 0.68), with no causal path. The link "Smoking History → Cancer Stage" is confirmed to be a true causal chain (ACE=0.79). In this case, the classification results are: "Medical Insurance Type" is classified as L1 (non-sensitive), and "Smoking History" is classified as L2 (moderately sensitive).

[0066] In other words, when a causal chain is determined to be non-causal, the disclosed embodiments can directly delete the causal chain from the causal graph, or retain the causal chain while lowering the classification of the causal variables in the causal chain. For example, the classification of the medical insurance type can be lowered from L3 to L1.

[0067] In other words, before adjusting the grading result of the causal variable in the target causal chain, the embodiment of the present disclosure can determine whether the target causal chain is a non-causal chain. If it is a non-causal chain, the grading result of the causal variable is lowered.

[0068] In step S220 , the average causal effect value of each causal chain in the causal graph is obtained.

[0069] As an optional method, after constructing a causal graph, the embodiment of the present disclosure can obtain the average causal effect (ACE) of each causal chain in the causal graph. The average causal effect value can be the average causal effect of the cause variable X on the result variable Y. The specific calculation formula is as follows: ; Here, X is the causal variable, such as smoking history, and Y is the outcome variable, such as lung cancer diagnosis. do(X=x) indicates an intervention on X. Calculating the average causal effect value can eliminate the influence of confounding factors and obtain a more accurate estimate of the causal effect.

[0070] In step S230 , if the average causal effect value is greater than a preset threshold, it is determined that the quantification result meets a preset condition.

[0071] As can be seen from the above description, after obtaining the quantitative results of a causal chain, embodiments of the present disclosure can determine whether the quantitative results meet preset conditions. If so, the classification results of the causal variables of the target causal chain that meets the preset conditions can be adjusted. Therefore, when the quantitative result is an average causal effect value, embodiments of the present disclosure can determine whether the average causal effect value is greater than a preset threshold. If the average causal effect value of a causal chain is greater than the preset threshold, the causal chain can be used as the target causal chain, i.e., the quantitative results of the causal chain are determined to meet the preset conditions.

[0072] Specifically, the preset thresholds may include a first threshold and a second threshold. When the average causal effect value of a causal chain is determined to be greater than or equal to the first threshold and less than the second threshold, the causal chain is determined to meet the preset conditions and can be used as the target causal chain. At this point, the grading results of the outcome variables in the target causal chain can be obtained. Based on this, the grading results of the cause variables are adjusted according to the grading results of the outcome variables.

[0073] In step S240 , the grading results of the causal variables in the target causal chain are adjusted, and the adjusted grading results are output.

[0074] In some embodiments, when it is determined that the average causal effect value of the target causal chain is greater than or equal to a first threshold and less than a second threshold, the embodiment of the present disclosure can obtain the grading result of the result variable in the target causal chain and adjust the grading result of the cause variable according to the grading result of the result variable.

[0075] Furthermore, after adjusting the causal variable classification results based on the outcome variable classification results, the disclosed embodiments can also determine whether there is any documentary evidence related to the target causal chain in the existing documentary evidence. If there is documentary evidence related to the target causal chain, the causal variable classification results of the target causal chain can be further adjusted.

[0076] As an example, the average causal effect value of the first causal chain in a causal diagram is 0.4. Comparison shows that this value is greater than a first threshold of 0.3 and less than a second threshold of 0.6. Therefore, the disclosed embodiment can adjust the classification of causal variable A in the first causal chain, for example, from L1 to L2. Subsequently, if there is documentary evidence related to the first causal chain, the classification of causal variable A can be adjusted to L3; otherwise, it remains at L2.

[0077] Here, the adjustment of the causal variable classification can be made according to the default rules. For example, if it is determined that the causal chain meets the preset conditions, the classification of the causal variables in the causal chain can be increased by two levels. The adjustment of the causal variable classification can also be made based on the classification of the result variables in the same causal chain. That is, if it is determined that the causal chain meets the preset conditions, the classification of the causal variables in the causal chain can be adjusted to the classification of the result variables. For example, if the classification of the causal variable is L1 and the classification of the result variable is L3, then when it is determined that the quantitative result meets the preset conditions, the classification of the causal variable can be adjusted to L3.

[0078] Optionally, when it is determined that the average causal effect value of the target causal chain is greater than or equal to a second threshold, embodiments of the present disclosure may adjust the grading result of the causal variable to the grading threshold, i.e., the grading of the causal variable may be set to the highest grading level corresponding to the medical data grading system. This grading threshold may be the maximum grading level. For example, when it is determined that the average causal effect value is greater than or equal to 0.6, embodiments of the present disclosure may set the causal variable grading to the system's highest level, L3. Different grading systems correspond to different highest grading levels.

[0079] When it is determined that the average causal chain of the causal chain is less than the first threshold, the classification of the cause variables is kept unchanged, that is, the causal chain is not used as the target causal chain.

[0080] To sum up, when it is determined that the causal effect value of the cause variable on the result variable reaches a preset threshold, the embodiment of the present disclosure can adjust the grade of the cause variable, and the adjusted grade can be greater than or equal to the grade of the result variable, so that a causal protection chain can be formed.

[0081] As an example, embodiments of the present disclosure can first discover a causal chain (X->Y), which can also be called a causal pair, which can be obtained through data-driven and ontological constraints. On this basis, ACE calculation and threshold judgment are performed on the causal chain to achieve dual robust estimation and medical reinforcement. In this process, if it is determined that 0.3≤ACE≤0.6, the classification of X can inherit the classification of Y. Conversely, if it is determined that ACE<0.3, the original classification is maintained. Alternatively, if it is determined that ACE≥0.6, the classification of the causal variable X can be directly adjusted to the highest classification.

[0082] It should be noted that when it is determined that the average causal effect value of the target causal chain is greater than or equal to a first threshold and less than a second threshold, the disclosed embodiment may also determine whether the grading result of the cause variable is less than the grading result of the effect variable. If so, the grading of the effect variable may be used as the grading of the cause variable. Otherwise, the grading of the cause variable may be adjusted to the target grading, and a cross-level review prompt may be output to confirm the target grading through this prompt.

[0083] In addition, the medical data classification method shown in the embodiment of the present disclosure can be implemented through a dedicated classification model, that is, the classification of each cause variable and result variable can be obtained through the classification model, and in the process of grading the cause variables, if it is determined that the causal chain corresponding to the cause variable meets the preset conditions, the classification result of the cause variable can be adjusted from the primary classification to the target classification.

[0084] After adjusting the grading results of the causal variables, the embodiment of the present disclosure may output the adjusted grading results.

[0085] In other implementations, the disclosed embodiments can monitor new causal chains in real time to discover potential causal chains, which can be referred to as incremental causality. For example, potential causal relationships, such as newly discovered gene-disease relationships, can be extracted from new medical data every 24 hours. This allows for real-time capture of changes in data sensitivity caused by new medical discoveries.

[0086] The monitoring process of the new causal chain can be as follows: Figure 6 As shown, based on Figure 6 As can be seen, the disclosed embodiments can acquire data streams in real time and analyze incremental causal relationships to update the causal graph. Based on this, new causal chains are graded using an online grading model to enable new data prediction. Incremental causal discovery is used to detect new associations, modify causal edges, and verify new causal chains through medical evidence. The online grading model can be used to calculate dynamic weights and adjust the grading of causal variables based on grading thresholds.

[0087] Through the above-mentioned monitoring process, the disclosed embodiment can capture new medical discoveries in real time to automatically adjust the sensitivity grading, such as changes in the pathogenicity of virus variants as new medical discoveries. The disclosed embodiment adjusts the grading results of the causal variables of the target causal chain that meet the preset conditions, which can not only ensure the security of the data, but also ensure the security of data with high scientific research value. For example, the R346T gene mutation will lead to an increased probability of stroke, which has strong scientific research value, and the weight (grading / level) of the gene mutation can be increased. Therefore, the disclosed embodiment can be applicable to both scenarios of data scientific research value and personal privacy (personal information leakage).

[0088] As a specific implementation method, Figure 7 As shown, after acquiring multi-source medical data (specified medical data), the embodiment of the present disclosure can pre-process it and obtain a causal graph based on a causal analysis engine, which can include medical ontology injection, causal discovery, and expert correction.

[0089] On this basis, a dynamic hierarchical decision module is used to grade and adjust the causal chains in the causal diagram and output the graded results. The graded results can be output via an API, blockchain, or database. Furthermore, the dynamic hierarchical decision module can include operations such as causal effect calculation, graded threshold determination, and dynamic weight adjustment.

[0090] For example, if the causal discovery engine captures a new causal chain, such as the S:R346T mutation leading to an increased risk of severe illness, the classification (level) of the relevant viral genetic data can be automatically upgraded from L2 to L3. Furthermore, the classification model can also complete adaptive adjustments within a specified timeframe, such as within 17 minutes.

[0091] In summary, by integrating medical ontology constraints with data-driven approaches, the disclosed embodiments can construct causal graphs to distinguish true causal chains from false correlations. Furthermore, by generating causally consistent counterfactual data, the interference of non-causal associations in hierarchical decision-making can be eliminated. Furthermore, the hierarchical model can be adaptively updated based on a dynamic causal evolution mechanism.

[0092] After obtaining a causal graph through specified medical data, the embodiment of the present disclosure can perform causal quantification on the causal graph to obtain a quantified result of each causal chain. On this basis, it is determined whether there is a target causal chain that meets preset conditions in the causal chain according to the quantified result. If it is determined that the target causal chain exists, the grading result of the causal variable in the target causal chain is adjusted, and the adjusted grading result is output, thereby achieving flexibility and accuracy in the grading of medical data.

[0093] Figure 8FIG. 1 is a block diagram of a medical data classification device according to an exemplary embodiment. Figure 8 The medical data grading device 300 may include a construction module 310 , a quantification module 320 and an adjustment module 330 .

[0094] The construction module 310 is configured to construct a causal graph according to specified medical data, wherein the causal graph includes a plurality of causal chains; The quantization module 320 is configured to perform a causal quantization operation on the causal graph to obtain a quantization result of each causal chain; The adjustment module 330 is configured to adjust the grading results of the causal variables in the target causal chain if it is determined that the quantified results of the target causal chain in the multiple causal chains meet the preset conditions, and output the adjusted grading results.

[0095] In some embodiments, the quantification module 320 may be further configured to obtain an average causal effect value of each causal chain in the causal graph, and use the average causal effect value as the quantification result.

[0096] In some embodiments, the medical data classification apparatus 300 may further include: The determination module is configured to determine whether the quantification result meets a preset condition if the average causal effect value is greater than a preset threshold.

[0097] In some implementations, the adjustment module 330 may include: an acquisition submodule configured to acquire a grading result of the outcome variable in the target causal chain when it is determined that the average causal effect value of the target causal chain is greater than or equal to a first threshold and less than a second threshold; The grading submodule is configured to adjust the grading result of the cause variable according to the grading result of the effect variable.

[0098] In some embodiments, the medical data classification apparatus 300 may further include: The secondary adjustment module is configured to adjust the classification result of the causal variable again if it is determined that there is literature evidence related to the target causal chain.

[0099] In some embodiments, the adjustment module 330 may be further configured to adjust the ranking result of the causal variable to a ranking threshold when it is determined that the average causal effect value of the target causal chain is greater than or equal to a second threshold.

[0100] In some embodiments, the construction module 310 is further configured to construct an initial causal graph based on specified medical data, obtain non-causal chains in the initial causal graph, and screen the initial causal graph based on the non-causal chains to obtain a target causal graph.

[0101] After obtaining a causal graph through specified medical data, the present disclosure can causally quantify the causal graph to obtain a quantified result of each causal chain. On this basis, it is determined whether there is a target causal chain that meets preset conditions in the causal chain according to the quantified result. If it is determined that the target causal chain exists, the grading result of the causal variable in the target causal chain is adjusted, and the adjusted grading result is output, thereby achieving flexibility and accuracy in the grading of medical data.

[0102] Regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0103] Figure 9 FIG. 7 is a block diagram of an electronic device 700 according to an exemplary embodiment. Figure 9 As shown, the electronic device 700 may include: a processor 701 , a memory 702 , and may further include one or more of a multimedia component 703 , an input / output (I / O) interface 704 , and a communication component 705 .

[0104] The processor 701 is used to control the overall operation of the electronic device 700 to complete all or part of the steps in the aforementioned medical data classification method. The memory 702 is used to store various types of data to support the operation of the electronic device 700. This data may include, for example, instructions for any application or method operating on the electronic device 700, as well as application-related data such as contact information, sent and received messages, images, audio, video, etc. The memory 702 can be implemented by any type of volatile or non-volatile storage device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 703 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 702 or transmitted via the communication component 705. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 704 provides an interface between the processor 701 and other interface modules, which may be a keyboard, a mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IOT, eMTC, or other 5G, etc., or a combination of one or more thereof, is not limited here. Therefore, the corresponding communication component 705 may include: a Wi-Fi module, a Bluetooth module, an NFC module, etc.

[0105] In an exemplary embodiment, the electronic device 700 can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the above-mentioned medical data classification method.

[0106] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When executed by a processor, the program instructions implement the steps of the aforementioned medical data classification method. For example, the computer-readable storage medium may be the aforementioned memory 702 including the program instructions. The program instructions may be executed by the processor 701 of the electronic device 700 to implement the aforementioned medical data classification method.

[0107] In another exemplary embodiment, a computer program product is further provided. The computer program product includes a computer program executable by a programmable device, and has a code portion for performing the above-mentioned medical data classification method when executed by the programmable device.

[0108] The preferred embodiments of the present disclosure are described in detail above in conjunction with the accompanying drawings. However, the present disclosure is not limited to the specific details of the above embodiments. Within the technical concept of the present disclosure, various simple modifications can be made to the technical solutions of the present disclosure, and these simple modifications all fall within the scope of protection of the present disclosure.

[0109] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any appropriate manner without contradiction. In order to avoid unnecessary repetition, the present disclosure will not further describe various possible combinations.

[0110] In addition, the various embodiments of the present disclosure may be arbitrarily combined, and as long as they do not violate the concept of the present disclosure, they should also be regarded as the contents disclosed by the present disclosure.

[0111] Those skilled in the art will also understand that the various illustrative logical blocks and steps listed in the embodiments of this application can be implemented through electronic hardware, computer software, or a combination of both. Whether such functions are implemented through hardware or software depends on the specific application and the design requirements of the entire system. Those skilled in the art may use various methods to implement the described functions for each specific application, but such implementation should not be construed as exceeding the scope of protection of the embodiments of this application.

[0112] It should be understood that, unless otherwise specifically noted, the features of the various embodiments of the present disclosure described herein may be combined with each other. As used herein, the term "and / or" includes any one of the relevant listed items and any combination of any two or more thereof; similarly, "at least one of" includes any one of the relevant listed items and any combination of any two or more thereof.

[0113] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one of such features. In this description, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0114] Furthermore, the word "exemplary" is used herein to mean serving as an example, instance, or illustration. Any aspect or design described herein as "exemplary" is not necessarily to be construed as advantageous over other aspects or designs. Rather, the use of the word exemplary is intended to present concepts in a concrete manner. As used herein, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or." That is, unless otherwise specified or clear from the context, "X applies to A or B" is intended to mean any of the natural inclusive permutations. That is, if X applies to A; X applies to B; or X applies to both A and B, then "X applies to A or B" satisfies any of the aforementioned instances. Furthermore, the articles "a" and "an," as used in this application and the appended claims, are generally understood to mean "one or more," unless otherwise specified or clear from the context to refer to the singular form.

[0115] Likewise, although the present disclosure has been shown and described with respect to one or more implementations, equivalent variations and modifications will occur to those skilled in the art upon reading and understanding this specification and the accompanying drawings. The present disclosure includes all such modifications and variations and is limited only by the scope of the claims. With particular regard to the various functions performed by the components described above (e.g., elements, resources, etc.), unless otherwise indicated, terms used to describe such components are intended to correspond to any component (functionally equivalent) that performs the specific function of the described component, even if not structurally equivalent to the disclosed structure. In addition, although particular features of the present disclosure may have been disclosed with respect to only one of several implementations, such features may be combined with one or more other features of other implementations as may be desired and advantageous for any given or particular application. Furthermore, to the extent that the terms "include," "have," "have," "have," or variations thereof are used in the detailed description or claims, such terms are intended to be inclusive in a manner similar to the term "comprising."

[0116] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the present disclosure being indicated by the appended claims.

[0117] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A medical data classification method, characterized in that: The method comprises constructing a causal graph according to specified medical data, wherein the causal graph includes a plurality of causal chains; Performing a causal quantization operation on the causal graph to obtain a quantization result of each causal chain; If it is determined that the quantified result of the target causal chain among the multiple causal chains meets the preset condition, the grading result of the causal variables in the target causal chain is adjusted, and the adjusted grading result is output.

2. The medical data classification method according to claim 1, characterized in that: The performing a causal quantization operation on the causal graph includes: An average causal effect value of each causal chain in the causal graph is obtained, and the average causal effect value is used as the quantification result.

3. The medical data classification method according to claim 2, characterized in that: The method further comprises: If the average causal effect value is greater than a preset threshold, it is determined that the quantification result meets a preset condition.

4. The medical data classification method according to claim 2, characterized in that: If it is determined that the quantitative result of the target causal chain among the multiple causal chains meets the preset condition, adjusting the grading results of the causal variables in the target causal chain includes: When it is determined that the average causal effect value of the target causal chain is greater than or equal to a first threshold value and less than a second threshold value, obtaining a grading result of the outcome variable in the target causal chain; The classification result of the cause variable is adjusted according to the classification result of the result variable.

5. The medical data classification method according to claim 4, characterized in that: After adjusting the classification result of the cause variable according to the classification result of the result variable, the method further comprises: If it is determined that there is literature evidence related to the target causal chain, the classification results of the causal variables are readjusted.

6. The medical data classification method according to claim 2, characterized in that: If it is determined that the quantitative result of the target causal chain among the multiple causal chains meets the preset condition, adjusting the grading results of the causal variables in the target causal chain includes: When it is determined that the average causal effect value of the target causal chain is greater than or equal to a second threshold, the ranking result of the causal variable is adjusted to the ranking threshold.

7. The medical data classification method according to claim 2, characterized in that: The step of constructing a causal graph based on specified medical data includes: constructing an initial causal graph based on specified medical data, and obtaining non-causal chains in the initial causal graph; The initial causal graph is screened based on the non-causal chain to obtain a target causal graph.

8. A medical data classification device, characterized in that: The device comprises: a construction module configured to construct a causal graph according to specified medical data, wherein the causal graph includes a plurality of causal chains; a quantization module configured to perform a causal quantization operation on the causal graph to obtain a quantization result of each causal chain; The adjustment module is configured to adjust the grading results of the causal variables in the target causal chain if it is determined that the quantified results of the target causal chain in the multiple causal chains meet the preset conditions, and output the adjusted grading results.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. An electronic device, characterized in that: include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1 to 7.

Images