Recommended methods, devices, equipment, and media for prescriptions for stomach distension based on traditional Chinese medicine experience.

Abstract

This invention provides a method, apparatus, device, and medium for recommending prescriptions for gastric distension based on traditional Chinese medicine (TCM) experience, relating to the field of data processing technology. The method includes: acquiring symptom information of the target individual; assigning a quantitative value to each symptom type based on a basic quantitative value and at least one dialectical quantitative factor; aggregating the quantitative values ​​of symptoms belonging to the same syndrome type according to a preset correspondence between syndrome types and symptoms to obtain a comprehensive score for each syndrome type, and determining the dialectical result of the target individual based on each comprehensive score; and generating a recommended prescription for the target individual based on the dialectical result and at least one dialectical quantitative factor, based on the mapping relationship between syndrome types and basic prescriptions, and a preset prescription adjustment rule base. This invention can transform the diagnostic and treatment experience of renowned TCM doctors for gastric distension into a quantifiable and reusable computational model, improving the consistency of syndrome differentiation and the matching degree between prescriptions and expert experience.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, apparatus, equipment, and medium for recommending prescriptions for gastric distension based on traditional Chinese medicine experience. Background Technology

[0002] Stomach distension is a common clinical disease of the spleen and stomach system, encompassing Western medicine diseases such as chronic gastritis and functional dyspepsia. In traditional Chinese medicine, the syndrome differentiation is complex, with common syndromes including spleen and stomach deficiency-cold type, liver and stomach qi stagnation type, and spleen and stomach damp-heat type.

[0003] Currently, the diagnosis and treatment of gastric distension in Traditional Chinese Medicine (TCM) relies heavily on personal experience. The diagnostic process is typically based on non-standardized symptom descriptions, subjectively assessing the syndrome type, and then prescribing based on experience. However, this model has significant drawbacks: firstly, symptom descriptions lack objective quantitative grading, leading to low consistency in syndrome differentiation among different doctors for the same patient and a high rate of syndrome confusion; secondly, existing TCM auxiliary systems mostly use fixed prescription databases for recommendations, failing to personalize prescriptions based on multi-dimensional information such as the patient's disease duration, objective Western medical indicators (e.g., gastroscopy pathology, Helicobacter pylori levels), and individual characteristics (e.g., age, weight), resulting in unstable clinical efficacy. More critically, the experience of renowned veteran TCM doctors cannot be preserved in a quantifiable and reusable form. Young doctors need years of apprenticeship to master the core diagnostic and treatment logic, and a large amount of valuable experience is lost with the retirement of experts. Therefore, how to quantify and transform the diagnostic and treatment experience of renowned veteran TCM doctors for gastric distension, and achieve adaptive syndrome differentiation and personalized prescription recommendations based on individual symptoms and multi-dimensional characteristics, has become an urgent technical problem to be solved. Summary of the Invention

[0004] This invention provides a method, apparatus, equipment, and medium for recommending prescriptions for gastric distension based on traditional Chinese medicine experience, in order to solve the problems of reliance on subjective experience, low consistency of syndrome differentiation, and low degree of personalization when prescribing gastric distension.

[0005] In a first aspect, embodiments of the present invention provide a method for recommending prescriptions for gastric distension based on traditional Chinese medicine experience, including: Obtain symptom information of the target subject; the symptom information includes symptom type and dialectical quantitative factors; For each symptom type, a quantitative value is assigned based on a baseline quantitative value and at least one dialectical quantitative factor to determine the quantitative value of that symptom type; among which, the dialectical quantitative factor includes disease course characteristic indicators, Western medicine objective indicators, and individual characteristic parameters; According to the pre-defined correspondence between syndrome types and symptoms, the quantitative values ​​of each symptom belonging to the same syndrome type are aggregated to obtain a comprehensive score for each syndrome type, and the dialectical result of the target object is determined based on each comprehensive score. Based on the dialectical results and at least one dialectical quantification factor, a recommended prescription for the target object is generated according to the mapping relationship between the syndrome type and the basic prescription, as well as the preset prescription adjustment rule base.

[0006] In one possible implementation, according to a pre-defined correspondence between syndrome types and symptoms, the quantitative values ​​of each symptom belonging to the same syndrome type are aggregated to obtain a comprehensive score for each syndrome type. Based on these comprehensive scores, the dialectical result of the target object is determined, including: The quantitative values ​​of each symptom and objective indicators from Western medicine are input into a trained Transformer classification model to obtain the probability values ​​of each syndrome type, which serve as the comprehensive score for each syndrome type. The dialectical result of the target object is the one with the highest probability value.

[0007] In one possible implementation, according to a pre-defined correspondence between syndrome types and symptoms, the quantitative values ​​of each symptom belonging to the same syndrome type are aggregated to obtain a comprehensive score for each syndrome type. Based on these comprehensive scores, the dialectical result of the target object is determined, including: Based on the inclusion relationship between syndrome types and symptom types, the quantitative values ​​of each symptom type are summed to obtain the comprehensive score of each syndrome type; The certificate type with the highest overall score is identified as the primary certificate type; If the ratio of the comprehensive score of the second-highest certificate type to the highest comprehensive score reaches a preset threshold, then the second-highest certificate type will be identified as a certificate type with both comprehensive and dual certifications.

[0008] In one possible implementation, for each symptom type, a quantitative value is determined by assigning a value based on a base quantitative value and at least one dialectical quantitative factor, including: Determine the seizure characteristic correction coefficient based on the duration of the disease in the target population; Based on the objective indicators corresponding to each symptom type in Western medicine, determine the objective indicator correction coefficient. For each symptom type, the symptom level is determined based on the frequency of onset, duration of each episode, and degree of impact on daily life, combined with corresponding objective indicators from Western medicine, and a corresponding symptom grading coefficient is determined. For each symptom type, an individual characteristic correction coefficient is determined based on the individual age of the target subject; For each symptom type, a comorbidity correction coefficient is determined based on the comorbidity relationship between other symptom types and that symptom type. For each symptom type, the base quantitative value, symptom grading coefficient, onset characteristic correction coefficient, objective indicator correction coefficient, individual characteristic correction coefficient, and comorbidity correction coefficient corresponding to that symptom type are multiplied together to obtain the quantitative value of that symptom type.

[0009] In one possible implementation, obtaining symptom information of the target object includes: The original symptom text of the target object is input into a large language model, and standardized terms are obtained through semantic mapping. The large language model is fine-tuned based on the correspondence between non-standardized symptoms and standardized terms.

[0010] In one possible implementation, based on the dialectical result and at least one dialectical quantification factor, and based on the mapping relationship between syndrome type and basic prescription, as well as a pre-defined prescription adjustment rule base, a recommended prescription for the target object is generated, including: Based on the mapping relationship between syndrome types and basic prescriptions, the drug types and dosage ranges corresponding to the syndrome differentiation results are used as basic prescriptions. Based on at least one dialectical quantitative factor, the basic prescription is adjusted according to a preset prescription adjustment rule base to obtain a recommended prescription for the target subject; wherein, the prescription adjustment rule base includes adding or subtracting drug types according to concurrent symptoms and special symptoms, and adjusting drug dosages according to symptom grading, disease course or individual characteristics.

[0011] In one possible implementation, before generating a recommended prescription for the target object based on the dialectical result and at least one dialectical quantification factor, the mapping relationship between syndrome type and basic prescription, and a pre-defined prescription adjustment rule base, the following is also included: Extract prescription adjustment rules from multiple historical prescriptions to form an initial rule base; The Apriori association rule algorithm was used to mine the initial rule base, and prescription adjustment rules with support greater than the first preset value and confidence greater than the second preset value were selected to form the prescription adjustment rule base.

[0012] Secondly, embodiments of the present invention provide a prescription recommendation device for gastric distension based on traditional Chinese medicine experience, comprising: The symptom acquisition module is used to acquire symptom information of the target object; the symptom information includes symptom type and dialectical quantification factor. The symptom quantification module is used to assign a quantitative value to each symptom type based on a base quantification value and at least one dialectical quantification factor, thereby determining the quantification value of that symptom type; among which, the dialectical quantification factor includes disease course characteristic indicators, Western medicine objective indicators, and individual characteristic parameters; The comprehensive dialectical module is used to aggregate the quantitative values ​​of each symptom belonging to the same syndrome according to the preset correspondence between syndrome types and symptoms, obtain the comprehensive score of each syndrome type, and determine the dialectical result of the target object based on each comprehensive score. The prescription recommendation module is used to generate recommended prescriptions for the target object based on the dialectical results and at least one dialectical quantification factor, the mapping relationship between the syndrome type and the basic prescription, and a preset prescription adjustment rule base.

[0013] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect or any possible implementation thereof.

[0014] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect or any possible implementation thereof.

[0015] This invention provides a method, apparatus, device, and medium for recommending prescriptions for gastric distension based on traditional Chinese medicine (TCM) experience. By introducing a symptom quantification mechanism, it incorporates symptom type, disease duration, Western medical objective indicators, and individual characteristics into the quantification process. This allows the quantified value of each symptom to comprehensively reflect its severity and the patient's overall background. The symptom quantification values ​​are aggregated according to syndrome type to generate a comprehensive score, which is then ranked to determine the syndrome differentiation result. This transforms the previously vague, experience-dependent syndrome differentiation process into a calculable and reproducible numerical comparison. Based on the syndrome differentiation result and the quantified factors, a recommended prescription is generated by combining the syndrome type-basic prescription mapping relationship and a prescription adjustment rule base. This represents a shift from fixed prescription library recommendations to dynamic prescription adjustments based on individual multidimensional characteristics, enabling recommended prescriptions to adapt to patients with different disease durations, objective indicators, and individual characteristics. Compared to existing technologies, this invention transforms the diagnostic and treatment experience of renowned TCM doctors for gastric distension into a quantifiable and reusable computational model. This improves the consistency of syndrome differentiation and the matching degree between prescriptions and expert experience, thereby solving the technical problems of difficulty in inheriting TCM experience in gastric distension, low consistency in diagnosis and treatment, and insufficient personalized prescriptions. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the implementation of a method for recommending prescriptions for gastric distension based on traditional Chinese medicine experience, as provided in this embodiment of the invention. Figure 2 This is a schematic diagram of the device for recommending prescriptions for stomach ulcers based on traditional Chinese medicine experience, provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0017] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0018] See Figure 1 The document illustrates a flowchart of the method for recommending prescriptions for stomach distension based on traditional Chinese medicine experience, as provided in an embodiment of the present invention. Details are as follows: Step 101: Obtain symptom information of the target subject; wherein, symptom information includes symptom type and dialectical quantitative factor.

[0019] In this embodiment, symptom type refers to abnormal clinical manifestations reported by the patient or observed by the doctor, such as a feeling of fullness, coldness, or pain in the epigastric region, as well as abnormalities in appetite and bowel movements. These symptom types are usually recorded in outpatient medical records or chief complaint texts in natural language.

[0020] Dialectical quantitative factors refer to objective clinical parameters that can influence the conclusions of TCM diagnosis. These parameters are not the symptoms themselves, but rather the quantification of the symptoms or the characterization of the disease background. Specific types of dialectical quantitative factors can include disease course characteristic indicators, objective Western medicine testing indicators, and individual patient characteristic parameters. Disease course characteristic indicators include the length of time from the first appearance of related symptoms to the current medical visit; objective Western medicine testing indicators include the degree of inflammation or atrophic changes grade in the gastroscopy pathology report, the value of the Helicobacter pylori breath test, the gastric emptying rate test results, and the results of routine stool examination; individual characteristic parameters include the patient's age, weight, and liver and kidney function status.

[0021] These symptom information can be obtained by receiving structured consultation data entered by doctors through a human-computer interaction interface, or by reading relevant fields from electronic medical records and test reports from the hospital information system through an interface. Alternatively, natural language processing technology can be used to extract information from unstructured text. This embodiment only requires obtaining the symptom type and its corresponding dialectical quantification factor, without limiting the specific acquisition techniques.

[0022] Step 102: For each symptom type, a quantitative value is assigned based on the basic quantitative value and at least one dialectical quantitative factor to determine the quantitative value of the symptom type; wherein, the dialectical quantitative factor includes disease course characteristic indicators, Western medicine objective indicators and individual characteristic parameters.

[0023] In this embodiment, a specific process for quantifying each symptom type based on a base quantification value and a dialectical quantification factor includes determining multiple correction coefficients and performing multiplication operations.

[0024] Specifically, the correction coefficient for the onset characteristics is first determined based on the duration of the disease in the target population. For example, when the disease duration reaches or exceeds one year, it indicates a high degree of chronicity, and the correction coefficient is set to a value greater than 1 (such as 1.2); otherwise, it is 1.

[0025] Secondly, for each symptom type, the system obtains the corresponding objective indicator test values ​​from Western medicine and determines the objective indicator correction coefficient according to a preset mapping table. For example, for epigastric fullness symptoms associated with gastroscopy pathology results, if the result is moderate or above chronic atrophic gastritis, the correction coefficient is set to 1.5; if it is mild atrophy, it is set to 1.3; if it is only moderate superficial gastritis, it is set to 1.2; and if there is no obvious abnormality, it is 1.

[0026] Next, the system determines the symptom level (mild, moderate, or severe) based on the frequency of symptom attacks, the duration of each episode, and the degree of impact on daily life, combined with the severity of relevant objective indicators from Western medicine. This determines the corresponding symptom grading coefficient, for example, 1 for mild, 2 for moderate, and 3 for severe. Specific grading standards are shown in Table 1.

[0027] Table 1

[0028] Secondly, the system determines the individual characteristic correction coefficient based on the individual age of the target object: for example, when the age is greater than 65 years old, the overall correction coefficient for all spleen and stomach deficiency-cold related symptoms is set to 1.1, otherwise it is 1.

[0029] Secondly, when other symptom types that are concomitantly related to the current symptom type exist, a concomitant association correction factor is determined. For example, if a patient has both "fatigue" and "cold stomach and cold limbs", the quantification value of "cold stomach and cold limbs" is multiplied by a correction factor of 1.1.

[0030] Finally, the system multiplies the base quantitative value, symptom grading coefficient, attack characteristic correction coefficient, objective indicator correction coefficient, individual characteristic correction coefficient, and concurrent symptom association correction coefficient corresponding to the symptom type to obtain the final quantitative value of the symptom type. The multiplication operation reflects the independent contribution and cumulative effect among the factors, which is consistent with the logic of comprehensive judgment in traditional Chinese medicine.

[0031] Step 103: According to the preset correspondence between syndrome types and symptoms, aggregate the quantitative values ​​of each symptom belonging to the same syndrome type to obtain a comprehensive score for each syndrome type, and determine the dialectical result of the target object based on each comprehensive score.

[0032] In this embodiment, the system pre-stores the correspondence between TCM syndrome types and symptom types. Taking gastric distension as an example, the spleen and stomach deficiency-cold type is usually associated with symptoms such as epigastric fullness and discomfort, cold stomach and limbs, loss of appetite, loose stools, fatigue, pale tongue with white coating, and deep and slow pulse; the liver and stomach qi stagnation type is associated with symptoms such as epigastric distension and pain radiating to the hypochondria, frequent belching, worsening of mood swings, red tongue with thin white coating, and wiry pulse; the spleen and stomach damp-heat type is associated with symptoms such as burning pain in the epigastric region, bitter taste in the mouth and bad breath, sticky and unsatisfactory stools, red tongue with yellow and greasy coating, and slippery and rapid pulse.

[0033] For each syndrome type, the system aggregates the quantitative values ​​of all associated symptoms. Aggregation can be summation, weighted summation, averaging, or other mathematical operations that reflect the overall contribution. The result of the aggregation is a comprehensive score for that syndrome type. The higher the comprehensive score, the better the patient's clinical manifestations match the syndrome type.

[0034] Subsequently, the system determines the dialectical result of the target object based on the comprehensive score of each syndrome type. The determination method can be: taking the syndrome type with the highest comprehensive score as the primary dialectical result, or when the scores of multiple syndrome types are close, outputting several of them simultaneously as a composite syndrome type. The dialectical result may include a primary syndrome type and possible concurrent syndromes. By elevating the quantitative information of individual symptoms to the syndrome type level, this embodiment achieves automated reasoning from symptoms to syndrome types.

[0035] Step 104: Based on the dialectical results and at least one dialectical quantification factor, and based on the mapping relationship between syndrome type and basic prescription, as well as the preset prescription adjustment rule base, generate a recommended prescription for the target object.

[0036] In this embodiment, the basic prescription refers to the core drug composition and its commonly used dosage range recommended for a specific syndrome type. These drug combinations are derived from the clinical experience summaries of renowned traditional Chinese medicine practitioners. Simultaneously, the system also maintains a prescription adjustment rule base, which contains empirical rules for personalized fine-tuning of the basic prescription under various circumstances.

[0037] After obtaining the diagnostic results, the system first extracts the corresponding basic prescription as the initial prescription based on the mapping relationship between the syndrome type and the basic prescription. Then, based on one or more diagnostic quantitative factors involved in steps 101 and 102 (such as the duration of illness, severity of symptoms, values ​​of objective test indicators, patient weight or age, etc.), the system queries the prescription adjustment rule base for matching rules, adjusts the initial prescription by adding or removing drugs or adjusting the dosage, and finally generates a recommended prescription. The adjustment rules may include: adding drugs for specific concurrent syndromes, deleting unnecessary drugs, increasing or decreasing the dosage of certain drugs, etc. The generated recommended prescription can be output to the doctor's terminal for the attending physician to refer to or further modify and confirm based on it.

[0038] This invention introduces a symptom quantification mechanism, incorporating symptom type, disease duration, objective Western medicine indicators, and individual characteristics into the quantification process. This allows the quantified value of each symptom to comprehensively reflect its severity and the patient's overall background. Symptom quantification values ​​are aggregated according to syndrome type to generate a comprehensive score, which is then ranked to determine the syndrome differentiation result. This transforms the previously vague, experience-dependent syndrome differentiation process into a calculable and reproducible numerical comparison. Based on the syndrome differentiation result and the quantified factors, a recommended prescription is generated by combining the syndrome type-basic prescription mapping relationship and a prescription adjustment rule base. This represents a shift from fixed prescription library recommendations to dynamic prescription adjustments based on individual multidimensional characteristics, enabling recommended prescriptions to adapt to the needs of patients with different disease durations, objective indicators, and individual characteristics. Compared to existing technologies, this invention transforms the diagnostic and treatment experience of renowned traditional Chinese medicine practitioners in treating gastric distension into a quantifiable and reusable computational model. This improves the consistency of syndrome differentiation and the matching degree between prescriptions and expert experience, thereby solving the technical problems of difficulty in inheriting traditional Chinese medicine experience in treating gastric distension, low consistency in diagnosis and treatment, and insufficient personalized prescriptions.

[0039] In one possible implementation, according to a pre-defined correspondence between syndrome types and symptoms, the quantitative values ​​of each symptom belonging to the same syndrome type are aggregated to obtain a comprehensive score for each syndrome type. Based on these comprehensive scores, the dialectical result of the target object is determined, including: The quantitative values ​​of each symptom and objective indicators from Western medicine are input into a trained Transformer classification model to obtain the probability values ​​of each syndrome type, which serve as the comprehensive score for each syndrome type. The dialectical result of the target object is the one with the highest probability value.

[0040] In this embodiment, one specific way to determine the dialectical result is by using the Transformer classification model.

[0041] The Transformer classification model requires training before application. First, medical records from renowned traditional Chinese medicine (TCM) clinics are collected. Each record includes quantified values ​​of each symptom, objective Western medicine indicators, and expert-annotated diagnostic results. This data is used as the training set to train a multi-classification model based on the Transformer architecture. This model, with multi-head self-attention as its core, can automatically learn the non-linear interactions between symptoms and between symptoms and objective indicators. At the model input layer, the quantified values ​​of all symptoms and the values ​​of each objective Western medicine indicator are concatenated into a feature vector. At the model output layer, the Softmax activation function is used to output the probability values ​​of each preset syndrome. These probability values ​​can be considered as the comprehensive score for each syndrome. For new patient data, the system inputs the calculated symptom quantified values ​​and the acquired objective Western medicine indicators into the trained model, which directly outputs the probability distribution of each syndrome. The system uses the syndrome with the highest probability value as the diagnostic result for the target object. This deep learning-based method is particularly suitable for cases with complex, atypical, or intertwined symptom presentations, and can uncover deep-seated patterns that are difficult to capture using linear weighting methods.

[0042] In one possible implementation, according to a pre-defined correspondence between syndrome types and symptoms, the quantitative values ​​of each symptom belonging to the same syndrome type are aggregated to obtain a comprehensive score for each syndrome type. Based on these comprehensive scores, the dialectical result of the target object is determined, including: Based on the inclusion relationship between syndrome types and symptom types, the quantitative values ​​of each symptom type are summed to obtain the comprehensive score of each syndrome type; The certificate type with the highest overall score is identified as the primary certificate type; If the ratio of the comprehensive score of the second-highest certificate type to the highest comprehensive score reaches a preset threshold, then the second-highest certificate type will be identified as a certificate type with both comprehensive and dual certifications.

[0043] In this embodiment, another way to determine the diagnostic result is to use a score summation method and introduce a rule for determining concurrent symptoms. Specifically, the system, according to a preset inclusion relationship between syndrome types and symptom types, sums the quantitative values ​​of all symptoms included in each syndrome type to obtain a comprehensive score for each syndrome type. Then, the syndrome type with the highest comprehensive score is determined as the primary syndrome type. To handle common concurrent symptoms in clinical practice, the system also sets a preset threshold, such as 70%. When the ratio of the score of the second-highest syndrome type to the highest score reaches or exceeds this threshold, the system determines the second-highest syndrome type as a concurrent syndrome type, and the final output diagnostic result is a primary syndrome type with the second-highest syndrome type present; if the ratio is lower than the threshold, only a single primary syndrome type is output. This threshold can be set based on the clinical experience of renowned traditional Chinese medicine doctors or statistical analysis of a batch of consultation data. The advantages of this implementation method are simple calculation, strong interpretability, and doctors can intuitively see the contribution of each symptom to the syndrome type score, which is convenient for clinical verification and manual correction.

[0044] This method can also be used to generate training sample labels for Transformers.

[0045] In one possible implementation, for each symptom type, a quantitative value is determined by assigning a value based on a base quantitative value and at least one dialectical quantitative factor, including: Determine the seizure characteristic correction coefficient based on the duration of the disease in the target population; Based on the objective indicators corresponding to each symptom type in Western medicine, determine the objective indicator correction coefficient. For each symptom type, the symptom level is determined based on the frequency of onset, duration of each episode, and degree of impact on daily life, combined with corresponding objective indicators from Western medicine, and a corresponding symptom grading coefficient is determined. For each symptom type, an individual characteristic correction coefficient is determined based on the individual age of the target subject; For each symptom type, a comorbidity correction coefficient is determined based on the comorbidity relationship between other symptom types and that symptom type. For each symptom type, the base quantitative value, symptom grading coefficient, onset characteristic correction coefficient, objective indicator correction coefficient, individual characteristic correction coefficient, and comorbidity correction coefficient corresponding to that symptom type are multiplied together to obtain the quantitative value of that symptom type.

[0046] In this embodiment, the system first determines the onset characteristic correction coefficient based on the duration of the target object's disease: for example, when the course of the disease reaches or exceeds one year, it indicates that the disease is chronic, and the correction coefficient is set to a value greater than 1 (such as 1.2), otherwise it is 1.

[0047] Secondly, for each symptom type, the system obtains the corresponding Western medicine objective indicator test value and determines the objective indicator correction coefficient according to the preset mapping table. For example, the epigastric fullness symptom is associated with the gastroscopy pathology result. If the result is moderate or above chronic atrophic gastritis, the correction coefficient is set to 1.5; if it is mild atrophy, it is set to 1.3; if it is only moderate superficial gastritis, it is set to 1.2; if there is no obvious abnormality, it is 1.

[0048] Furthermore, the system determines the symptom level (mild, moderate, or severe) based on the frequency of symptom onset, duration of each episode, and degree of impact on daily life, combined with the severity of relevant objective indicators from Western medicine. It then determines the corresponding symptom grading coefficient, for example, 1 for mild, 2 for moderate, and 3 for severe.

[0049] Secondly, the system determines the individual characteristic correction coefficient based on the individual age of the target object: for example, when the age is greater than 65 years old, the overall correction coefficient for all spleen and stomach deficiency-cold related symptoms is set to 1.1, otherwise it is 1.

[0050] Furthermore, when there are other symptom types that are concomitantly associated with the current symptom type, a concomitant association correction factor is determined: for example, if a patient has both "fatigue" and "cold stomach and cold limbs", the quantitative value of "cold stomach and cold limbs" is multiplied by a correction factor of 1.1.

[0051] Finally, the system multiplies the basic quantitative value, symptom grading coefficient, onset characteristic correction coefficient, objective indicator correction coefficient, individual characteristic correction coefficient, and comorbidity correction coefficient corresponding to the symptom type together to obtain the final quantitative value of the symptom type.

[0052] In one possible implementation, obtaining symptom information of the target object includes: The original symptom text of the target object is input into a large language model, and standardized terms are obtained through semantic mapping. The large language model is fine-tuned based on the correspondence between non-standardized symptoms and standardized terms.

[0053] In this embodiment, since patients' original symptom descriptions are usually presented in non-standardized natural language, such as "a cold feeling in the stomach," "bloating," or "bloating after eating," these expressions cannot be directly used as standardized symptom types. Therefore, the system employs a fine-tuned large language model. The base model is a pre-trained TCM large language model, which is pre-trained on large-scale TCM literature and medical record data and possesses semantic understanding capabilities in the TCM field. Then, the base model is fine-tuned using the correspondence between non-standardized symptoms and standardized terms annotated by renowned TCM doctors. This fine-tuning can employ Few-Shot learning, meaning that providing only a small number of high-quality annotated examples allows the model to quickly adapt to the symptom mapping task.

[0054] In practical use, the original symptom description text of the target subject is input into the fine-tuned large language model. The model outputs one or more standardized terms with the highest matching degree and their corresponding confidence scores. The system sets a confidence threshold, such as 90%. When the output confidence score reaches or exceeds the threshold, the standardized term is automatically used to complete the mapping; when the confidence score is below the threshold, the system pushes the mapping request to a manual confirmation interface, where clinicians manually select or correct the standardized term. The manually confirmed data can be used periodically for incremental training of the model to continuously improve the mapping accuracy.

[0055] Common mapping examples include: "a cold feeling in the stomach, a cold abdomen, and a feeling of coldness in the stomach" are mapped to "cold stomach and cold limbs"; "bloating, abdominal distension, feeling bloated after eating, and feeling full after meals" are mapped to "a feeling of fullness and discomfort in the stomach"; "stomach pain, intermittent stomach pain, and dull stomach pain" are mapped to "stomach pain"; and "not wanting to eat and having no appetite" are mapped to "loss of appetite," etc.

[0056] In one possible implementation, based on the dialectical result and at least one dialectical quantification factor, and based on the mapping relationship between syndrome type and basic prescription, as well as a pre-defined prescription adjustment rule base, a recommended prescription for the target object is generated, including: Based on the mapping relationship between syndrome types and basic prescriptions, the drug types and dosage ranges corresponding to the syndrome differentiation results are used as basic prescriptions. Based on at least one dialectical quantitative factor, the basic prescription is adjusted according to a preset prescription adjustment rule base to obtain a recommended prescription for the target subject; wherein, the prescription adjustment rule base includes adding or subtracting drug types according to concurrent symptoms and special symptoms, and adjusting drug dosages according to symptom grading, disease course or individual characteristics.

[0057] In this embodiment, the specific process for generating recommended prescriptions includes two stages: basic prescription selection and rule adjustment. First, based on the diagnostic results determined in step 103, a mapping table between the syndrome type and the basic prescription is queried. This mapping table is obtained by analyzing a large number of historical prescriptions from renowned traditional Chinese medicine practitioners. For example, for the spleen and stomach deficiency-cold type, the basic prescription usually includes Lindera strychnifolia, Angelica sinensis, Ligusticum chuanxiong, Paeonia lactiflora, and Atractylodes macrocephala, with each drug having a commonly used dosage range. The system uses the drug type corresponding to the diagnostic results and its intermediate dosage range as the basic prescription. Taking the spleen and stomach deficiency-cold type, liver and stomach qi stagnation type, and spleen and stomach damp-heat type as examples, their frequently used drugs and dosage ranges are shown in Table 2.

[0058] Table 2

[0059] Secondly, based on at least one dialectical quantitative factor involved in steps 101 and 102 (such as disease course, symptom grade, objective indicator values, weight, liver and kidney function, etc.), the system calls upon a preset prescription adjustment rule base to adjust the basic prescription. The prescription adjustment rule base contains various rule types: for example, rules for increasing or decreasing drug types based on concurrent symptoms and special symptoms—adding Amomum villosum when there is severe postprandial abdominal distension, and increasing the dosage of Paeonia lactiflora when there is significant epigastric cramping pain; rules for adjusting drug dosage based on symptom grade—increasing the dosage of Paeonia lactiflora to a medium-high dose range when there is moderate epigastric cramping pain; rules for adjusting drug dosage based on disease course or individual characteristics—adding Salvia miltiorrhiza to invigorate blood circulation when the disease course exceeds one year, increasing the upper limit of all drug dosages by a certain percentage when the weight exceeds 80 kg, and reducing the dosage of toxic drugs by a certain percentage when liver and kidney function is abnormal. Rules in the rule base can be applied cumulatively, meaning a basic prescription may trigger multiple rules simultaneously. The system executes all matching rules sequentially, ultimately generating a completely personalized recommended prescription. Table 3 lists several typical dose adjustment triggering conditions and their corresponding rules.

[0060] Table 3

[0061] After the prescription is output to the physician's terminal, the doctor can make manual adjustments to it for clinical treatment.

[0062] In one possible implementation, before generating a recommended prescription for the target object based on the dialectical result and at least one dialectical quantification factor, the mapping relationship between syndrome type and basic prescription, and a pre-defined prescription adjustment rule base, the following is also included: Extract prescription adjustment rules from multiple historical prescriptions to form an initial rule base; The Apriori association rule algorithm was used to mine the initial rule base, and prescription adjustment rules with support greater than the first preset value and confidence greater than the second preset value were selected to form the prescription adjustment rule base.

[0063] In this embodiment, a prescription adjustment rule base needs to be built before implementing step 104. The building process includes the following steps: First, collect multiple historical prescriptions from renowned veteran TCM doctors. This prescription data should include complete information on patient symptoms, diagnostic results, and the medications and dosages used. Organize these prescriptions, extracting adjustments relative to the baseline prescription to form an initial rule base. This initial rule base can be derived from expert interview records, directly transforming the oral medication experience of renowned TCM doctors into structured rules.

[0064] Second, the Apriori association rule algorithm is used to mine the initial rule base or directly from historical prescription data. The Apriori algorithm can discover symptom-drug-dosage combinations that frequently appear in the same prescription. Support thresholds (e.g., 30%) and confidence thresholds (e.g., 80%) are set to filter association rules that meet the conditions. Support represents the frequency of a rule's occurrence in the dataset, and confidence represents the conditional probability that the conclusion is true under the given conditions. For example, the algorithm might discover the rule "moderate epigastric pain → white peony root dosage ≥20g". If the support of this rule is 35% and the confidence is 85%, it indicates that the rule is statistically significant.

[0065] Third, a team of renowned traditional Chinese medicine experts reviews the rules obtained from data mining, eliminating rules that do not conform to clinical logic or are merely coincidental, and revising the dosage ranges or applicable conditions in the rules, ultimately forming a formally effective prescription adjustment rule base. This rule base can be periodically updated as clinical data accumulates to maintain its timeliness and accuracy.

[0066] In a complete application example of this invention, 300 patients with gastric distension diagnosed by renowned traditional Chinese medicine doctors were selected as verification subjects. The symptom descriptions and objective test data of these patients were input into the system of this invention. The system automatically generated diagnostic results and compared them with the manual diagnostic results of the renowned traditional Chinese medicine doctors. The comparison results showed that the overall diagnostic accuracy of the system reached 85%, with the diagnostic accuracy for the spleen and stomach deficiency-cold type reaching 90%. The consistency between the system and manual diagnostics was 70%, significantly higher than the 55% consistency level of existing TCM auxiliary systems.

[0067] Regarding personalized prescriptions, feedback data from 1000 actual prescriptions from renowned traditional Chinese medicine (TCM) doctors was collected and input into the system's model iteration module to retrain the prescription recommendation model. Comparison before and after model iteration showed that the personalized matching degree between prescriptions and the experience of renowned TCM doctors improved from 60% to 70%, the accuracy of drug dosage recommendations increased from 75% to 83%, and the consistency of drug combinations improved from 65% to 75%. This indicates that the closed-loop iterative optimization mechanism adopted in this invention—that is, automatically capturing expert adjustments to recommended prescriptions and using these adjustments as training samples to re-optimize the model—enables the system performance to continuously improve with the accumulation of clinical data.

[0068] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0069] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0070] Figure 2 A schematic diagram of the device for recommending prescriptions for stomach distension based on traditional Chinese medicine experience, provided in an embodiment of the present invention, is shown. For ease of explanation, only the parts relevant to the embodiment of the present invention are shown, and are described in detail below: like Figure 2 As shown, the prescription recommendation device 2 for stomach discomfort based on traditional Chinese medicine experience includes: The symptom acquisition module 21 is used to acquire symptom information of the target object; wherein, the symptom information includes symptom type and dialectical quantification factor; The symptom quantification module 22 is used to assign a quantitative value to each symptom type based on a basic quantification value and at least one dialectical quantification factor, thereby determining the quantification value of that symptom type; wherein, the dialectical quantification factor includes disease course characteristic indicators, Western medicine objective indicators, and individual characteristic parameters; The comprehensive dialectical module 23 is used to aggregate the quantitative values ​​of each symptom belonging to the same syndrome according to the preset correspondence between syndrome types and symptoms, obtain the comprehensive score of each syndrome type, and determine the dialectical result of the target object based on each comprehensive score. The prescription recommendation module 24 is used to generate recommended prescriptions for the target object based on the dialectical results and at least one dialectical quantification factor, the mapping relationship between the syndrome type and the basic prescription, and a preset prescription adjustment rule base.

[0071] In one possible implementation, the comprehensive dialectical module 23 is specifically used for: The quantitative values ​​of each symptom and objective indicators from Western medicine are input into a trained Transformer classification model to obtain the probability values ​​of each syndrome type, which serve as the comprehensive score for each syndrome type. The dialectical result of the target object is the one with the highest probability value.

[0072] In one possible implementation, the comprehensive dialectical module 23 is specifically used for: Based on the inclusion relationship between syndrome types and symptom types, the quantitative values ​​of each symptom type are summed to obtain the comprehensive score of each syndrome type; The certificate type with the highest overall score is identified as the primary certificate type; If the ratio of the comprehensive score of the second-highest certificate type to the highest comprehensive score reaches a preset threshold, then the second-highest certificate type will be identified as a certificate type with both comprehensive and dual certifications.

[0073] In one possible implementation, the symptom quantification module 22 is specifically used for: Determine the seizure characteristic correction coefficient based on the duration of the disease in the target population; Based on the objective indicators corresponding to each symptom type in Western medicine, determine the objective indicator correction coefficient. For each symptom type, the symptom level is determined based on the frequency of onset, duration of each episode, and degree of impact on daily life, combined with corresponding objective indicators from Western medicine, and a corresponding symptom grading coefficient is determined. For each symptom type, an individual characteristic correction coefficient is determined based on the individual age of the target subject; For each symptom type, a comorbidity correction coefficient is determined based on the comorbidity relationship between other symptom types and that symptom type. For each symptom type, the base quantitative value, symptom grading coefficient, onset characteristic correction coefficient, objective indicator correction coefficient, individual characteristic correction coefficient, and comorbidity correction coefficient corresponding to that symptom type are multiplied together to obtain the quantitative value of that symptom type.

[0074] In one possible implementation, the symptom acquisition module 21 is specifically used for: The original symptom text of the target object is input into a large language model, and standardized terms are obtained through semantic mapping. The large language model is fine-tuned based on the correspondence between non-standardized symptoms and standardized terms.

[0075] In one possible implementation, the prescription recommendation module 24 is specifically used for: Based on the mapping relationship between syndrome types and basic prescriptions, the drug types and dosage ranges corresponding to the syndrome differentiation results are used as basic prescriptions. Based on at least one dialectical quantitative factor, the basic prescription is adjusted according to a preset prescription adjustment rule base to obtain a recommended prescription for the target subject; wherein, the prescription adjustment rule base includes adding or subtracting drug types according to concurrent symptoms and special symptoms, and adjusting drug dosages according to symptom grading, disease course or individual characteristics.

[0076] In one possible implementation, the prescription recommendation module 24 is also used for: Before generating a recommended prescription for the target object based on the dialectical results and at least one dialectical quantification factor, the mapping relationship between the syndrome type and the basic prescription, and the preset prescription adjustment rule base, prescription adjustment rules are extracted from multiple historical prescriptions to form an initial rule base. The Apriori association rule algorithm was used to mine the initial rule base, and prescription adjustment rules with support greater than the first preset value and confidence greater than the second preset value were selected to form the prescription adjustment rule base.

[0077] This invention introduces a symptom quantification mechanism, incorporating symptom type, disease duration, objective Western medicine indicators, and individual characteristics into the quantification process. This allows the quantified value of each symptom to comprehensively reflect its severity and the patient's overall background. Symptom quantification values ​​are aggregated by syndrome type to generate a comprehensive score, which is then ranked to determine the diagnostic result. This transforms the previously vague, experience-dependent diagnostic process into a calculable and reproducible numerical comparison. Based on the diagnostic result and the quantified factors, a recommended prescription is generated by combining the syndrome type-basic prescription mapping relationship and a prescription adjustment rule base. This shifts from a fixed prescription library to a dynamic adjustment of prescriptions based on individual multidimensional characteristics, enabling recommended prescriptions to adapt to patients with different disease durations, objective indicators, and individual characteristics. Compared to existing technologies, this invention transforms the diagnostic and treatment experience of renowned traditional Chinese medicine practitioners in treating gastric distension into a quantifiable and reusable computational model. This improves the consistency of diagnosis and the matching degree between prescriptions and expert experience, thereby solving the technical problems of difficulty in inheriting TCM experience in treating gastric distension, low consistency in diagnosis and treatment, and insufficient personalized prescriptions.

[0078] Figure 3 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. For example... Figure 3 As shown, the electronic device 3 of this embodiment includes a processor 30 and a memory 31. The memory 31 stores a computer program 32. When the processor 30 executes the computer program 32, it implements the steps in the various method embodiments described above. Alternatively, when the processor 30 executes the computer program 32, it implements the functions of each module / unit in the various device embodiments described above.

[0079] For example, computer program 32 may be divided into one or more modules / units, which are stored in memory 31 and executed by processor 30 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 32 in electronic device 3.

[0080] Electronic device 3 may include, but is not limited to, processor 30 and memory 31. Those skilled in the art will understand that... Figure 3 This is merely an example of electronic device 3 and does not constitute a limitation on electronic device 3. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device 3 may also include input / output devices, network access devices, buses, etc.

[0081] The processor 30 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0082] The memory 31 can be an internal storage unit of the electronic device 3, such as a hard disk or memory of the electronic device 3. The memory 31 can also be an external storage device of the electronic device 3, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 3. Furthermore, the memory 31 can include both internal and external storage units of the electronic device 3. The memory 31 is used to store the computer program 32 and other programs and data required by the electronic device 3. The memory 31 can also be used to temporarily store data that has been output or will be output.

[0083] For the sake of simplicity and clarity, only the above-described functional modules / units are used as examples. In practical applications, the functions described above can be assigned to different functional modules / units as needed. These modules / units can be implemented in hardware, software, or a combination of both.

[0084] This invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the methods described in the above-described method embodiments.

[0085] This invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the methods described in the above-described method embodiments.

[0086] Computer programs include computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. Computer-readable media can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0087] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0088] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for recommending a prescription for Weipi disease by applying traditional Chinese medicine experience, characterized in that, include: Obtain symptom information of the target object; wherein, the symptom information includes symptom type and dialectical quantification factor; For each symptom type, a quantitative value is assigned based on a baseline quantitative value and at least one dialectical quantitative factor to determine the quantitative value of that symptom type; among which, the dialectical quantitative factor includes disease course characteristic indicators, Western medicine objective indicators, and individual characteristic parameters; According to the pre-defined correspondence between syndrome types and symptoms, the quantitative values ​​of each symptom belonging to the same syndrome type are aggregated to obtain a comprehensive score for each syndrome type, and the dialectical result of the target object is determined based on each comprehensive score. Based on the dialectical results and at least one dialectical quantification factor, a recommended prescription for the target object is generated according to the mapping relationship between syndrome type and basic prescription, as well as a preset prescription adjustment rule base.

2. The method for recommending prescriptions for stomach fullness based on traditional Chinese medicine experience according to claim 1, characterized in that, The process involves aggregating the quantitative values ​​of symptoms belonging to the same syndrome type according to a pre-defined correspondence between syndrome types and symptoms, obtaining a comprehensive score for each syndrome type, and determining the dialectical result of the target object based on each comprehensive score, including: The quantitative values ​​of each symptom and objective indicators from Western medicine are input into a trained Transformer classification model to obtain the probability values ​​of each syndrome type, which serve as the comprehensive score for each syndrome type. The dialectical result of the target object is the one with the highest probability value.

3. The method for recommending prescriptions for stomach fullness based on traditional Chinese medicine experience according to claim 1, characterized in that, The process involves aggregating the quantitative values ​​of symptoms belonging to the same syndrome type according to a pre-defined correspondence between syndrome types and symptoms, obtaining a comprehensive score for each syndrome type, and determining the dialectical result of the target object based on each comprehensive score, including: Based on the inclusion relationship between syndrome types and symptom types, the quantitative values ​​of each symptom type are summed to obtain the comprehensive score of each syndrome type; The certificate type with the highest overall score is identified as the primary certificate type; If the ratio of the comprehensive score of the second-highest certificate type to the highest comprehensive score reaches a preset threshold, then the second-highest certificate type is determined as a dual certificate type.

4. The method for recommending prescriptions for stomach fullness based on traditional Chinese medicine experience according to claim 1, characterized in that, The process of assigning a quantitative value to each symptom type based on a base quantitative value and at least one dialectical quantitative factor to determine the quantitative value of that symptom type includes: The seizure characteristic correction coefficient is determined based on the duration of the disease in the target group; Based on the objective indicators corresponding to each symptom type in Western medicine, determine the objective indicator correction coefficient. For each symptom type, the symptom level is determined based on the frequency of onset, duration of each episode, and degree of impact on daily life, combined with corresponding objective indicators from Western medicine, and a corresponding symptom grading coefficient is determined. For each symptom type, an individual characteristic correction coefficient is determined based on the individual age of the target subject; For each symptom type, a comorbidity correction coefficient is determined based on the comorbidity relationship between other symptom types and that symptom type. For each symptom type, the base quantitative value, symptom grading coefficient, onset characteristic correction coefficient, objective indicator correction coefficient, individual characteristic correction coefficient, and comorbidity correction coefficient corresponding to that symptom type are multiplied together to obtain the quantitative value of that symptom type.

5. The method for recommending prescriptions for stomach discomfort based on traditional Chinese medicine experience according to claim 1, characterized in that, The acquisition of symptom information of the target object includes: The original symptom text of the target object is input into a large language model, and standardized terms are obtained through semantic mapping; wherein, the large language model is fine-tuned based on the correspondence between non-standardized symptoms and standardized terms.

6. The method for recommending prescriptions for gastric distension based on traditional Chinese medicine experience according to any one of claims 1 to 5, characterized in that, The step of generating a recommended prescription for the target object based on the dialectical result and at least one dialectical quantification factor, the mapping relationship between syndrome type and basic prescription, and a preset prescription adjustment rule base includes: Based on the mapping relationship between syndrome types and basic prescriptions, the drug types and dosage ranges corresponding to the syndrome differentiation results are used as basic prescriptions. Based on at least one dialectical quantification factor, the basic prescription is adjusted according to a preset prescription adjustment rule base to obtain a recommended prescription for the target object; wherein, the prescription adjustment rule base includes adding or subtracting drug types based on concurrent symptoms and special symptoms, and adjusting drug dosages based on symptom grading, disease course, or individual characteristics.

7. The method for recommending prescriptions for stomach fullness based on traditional Chinese medicine experience according to claim 6, characterized in that, Before generating the recommended prescription for the target object based on the dialectical result and at least one dialectical quantification factor, the mapping relationship between syndrome type and basic prescription, and a preset prescription adjustment rule base, the method further includes: Extract prescription adjustment rules from multiple historical prescriptions to form an initial rule base; The Apriori association rule algorithm was used to mine the initial rule base, and prescription adjustment rules with support greater than the first preset value and confidence greater than the second preset value were selected to form the prescription adjustment rule base.

8. A prescription recommendation device for stomach discomfort based on traditional Chinese medicine experience, characterized in that, include: The symptom acquisition module is used to acquire symptom information of the target object; wherein, the symptom information includes symptom type and dialectical quantification factor; The symptom quantification module is used to assign a quantitative value to each symptom type based on a base quantification value and at least one dialectical quantification factor, thereby determining the quantification value of that symptom type; among which, the dialectical quantification factor includes disease course characteristic indicators, Western medicine objective indicators, and individual characteristic parameters; The comprehensive dialectical module is used to aggregate the quantitative values ​​of each symptom belonging to the same syndrome according to the preset correspondence between syndrome types and symptoms, to obtain a comprehensive score for each syndrome type, and to determine the dialectical result of the target object based on each comprehensive score. The prescription recommendation module is used to generate a recommended prescription for the target object based on the dialectical results and at least one dialectical quantification factor, the mapping relationship between the syndrome type and the basic prescription, and a preset prescription adjustment rule base.

9. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images