Intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions

Abstract

The invention relates to the technical field of traditional Chinese medicine prescription recommendation, and discloses a traditional Chinese medicine prescription intelligent auxiliary recommendation management system, which comprises a recommendation module, a symptom database and a traditional Chinese medicine database, and is characterized in that the recommendation module comprises a data acquisition unit, a prescription generation unit, a first verification unit, a second verification unit and an evaluation unit; a data acquisition unit is arranged to acquire experience data of doctors and traditional Chinese medicine four-diagnosis information, a prescription recommendation model is established by taking the experience data of the doctors and the traditional Chinese medicine four-diagnosis information of patients as parameters, recommended prescriptions are obtained, and the prescription recommendation model is corrected according to feedback of the doctors and the patients on the recommended prescriptions. According to the invention, the system provides more accurate prescription recommendation through a machine learning algorithm in combination with historical data and specific conditions of patients, and the automatic and personalized recommendation improves the treatment effect, reduces the burden of doctors, and avoids subjective deviation.

Description

Technical Field

[0001] The present invention relates to the technical field of traditional Chinese medicine prescription recommendation, and particularly to an intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions. Background Art

[0002] In traditional Chinese medicine treatment, doctors use methods such as observation, auscultation and olfaction, interrogation, and palpation, combined with the symptoms and constitution of patients, to formulate treatment plans. Traditional Chinese medicine prescriptions are generally selected by doctors based on their rich experience and profound understanding of traditional Chinese medicine theory for medicinal materials and dosages. Although this process makes full use of doctors' professional knowledge, there are also problems such as strong subjectivity, difficulty in accurately considering individual differences, and heavy doctor workload. In addition, traditional methods cannot quickly synthesize a large amount of medical record data and historical prescription information, easily leading to the neglect of some important data, thus affecting the accuracy and personalization of prescriptions.

[0003] To make up for the deficiencies of traditional methods, some hospitals have begun to introduce auxiliary systems based on rule engines. However, these systems mostly rely on preset rule libraries, fail to fully utilize patients' individual characteristics for dynamic adjustment, and lack a flexible feedback mechanism.

[0004] Therefore, we propose an intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions. By combining doctors' experience data and patients' four diagnostic information, machine learning and feedback mechanisms are used to continuously optimize the prescription recommendation system, so as to achieve more accurate, safer, and more personalized traditional Chinese medicine prescription recommendations. Summary of the Invention

[0005] The present invention mainly solves the technical problems existing in the above-mentioned prior art, and provides an intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions.

[0006] To achieve the above object, the present invention adopts the following technical solution. An intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions includes a recommendation module, a symptom database, and a traditional Chinese medicine database. The recommendation module includes a data collection unit, a prescription generation unit, a first verification unit, a second verification unit, and an evaluation unit. By setting the data collection unit to collect doctors' experience data and the four diagnostic information of traditional Chinese medicine, a prescription recommendation model is established with doctors' experience data and patients' four diagnostic information of traditional Chinese medicine as parameters to obtain a recommended prescription. At the same time, the prescription recommendation model is corrected according to the feedback from doctors and patients on the recommended prescription, so that the prescription recommendation model is continuously improved, specifically including the following multiple steps:

[0007] Step 1: Data collection;

[0008] Step 2: Prescription generation;

[0009] Step 3: First verification unit;

[0010] Step 4: Second verification unit;

[0011] Step 5: Evaluation Unit;

[0012] Through data collection, prescription generation, dual verification, and evaluation mechanisms, the system ensures that the traditional Chinese medicine prescription recommendations not only conform to traditional Chinese medicine theory but also are personalized, accurate, and continuously improved in practical applications.

[0013] Preferably, for the data collection, doctor experience data and four diagnostic information are integrated to form a multi-dimensional patient data profile, and the weighted average method is used for information fusion;

[0014] where x f use d is the fused data, x i is the value of each data source, and w i is the weight of each data source.

[0015] Preferably, for the prescription generation, through a rule engine, a preliminary prescription recommendation model is established according to traditional Chinese medicine theory, and a machine learning algorithm is used to further train the recommendation model.

[0016] Preferably, for the prescription generation, the decision tree algorithm is combined with information gain to construct a decision tree model applicable to traditional Chinese medicine prescription recommendation. The specific implementation steps include S201, target and input variables, and S202, decision tree model construction.

[0017] Preferably, for the prescription generation, in S201, the target is to generate a reasonable traditional Chinese medicine prescription based on the patient's symptoms and constitution. The input variables include symptom information, constitution information, patient history, and environmental factors. These input variables serve as the features of the decision tree algorithm and determine the patient's final traditional Chinese medicine prescription.

[0018] Preferably, for the prescription generation, in S202, the decision tree model construction process includes S21, selecting split features, and S22, recursive splitting process. Split features are selected according to the features of the input data. At each node, the system calculates the optimal split feature according to the information gain formula:

[0019]

[0020] The information gain calculates the degree of reduction in data uncertainty after splitting the data set by a certain feature, so as to select the feature with the largest information gain as the split feature of the current node. At each node, the information gain of each feature is calculated, and the optimal feature is selected for splitting until the stop condition is reached to generate a decision tree.

[0021] Preferably, the first verification unit verifies whether the recommended prescription conforms to the compatibility principles and safety standards of traditional Chinese medicine through a set medical rule base, and at the same time allows doctors to manually review the prescriptions recommended by the system to ensure that the recommended results are consistent with the patient's actual clinical situation.

[0022] Preferably, the second verification unit re-verifies the prescription according to the patient's individual characteristics, and at the same time verifies whether there are adverse drug interactions between the drugs in the recommended prescription to ensure that the synergistic effect of the drugs is maximized and the side effects are minimized.

[0023] Preferably, the evaluation unit conducts a follow-up evaluation based on the patient's treatment feedback information to judge the effect of the recommended prescription, and collects treatment effect data through multiple methods such as regular follow-up visits, follow-up data, and health monitoring devices.

[0024] Beneficial effects

[0025] The present invention provides an intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions, which has the following beneficial effects:

[0026] (1) The intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions can generate personalized traditional Chinese medicine prescriptions according to the patient's multi-dimensional information (such as symptoms, constitution, historical medical records, etc.) by combining the doctor's experience data and the four diagnostic information (inspection, auscultation and olfaction, interrogation, palpation). Traditional Chinese medicine prescriptions often rely on doctors' experience and judgment, while the system can provide more accurate prescription recommendations through machine learning algorithms, combined with historical data and the patient's specific situation. For example, when a patient shows symptoms of "excessive liver fire", the system will not only recommend herbs that conform to traditional Chinese medicine theory (such as bupleurum, gentian, etc.), but also adjust the dosage and compatibility of the herbs according to the patient's individual factors such as age, constitution, and allergy history. This automated and personalized recommendation process significantly improves the effect of traditional Chinese medicine treatment, while reducing the burden on doctors and avoiding the subjective bias that may exist in traditional methods.

[0027] (2) This intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions conducts multiple verifications through the first verification unit and the second verification unit to ensure that the recommended prescriptions comply with the principles of traditional Chinese medicine compatibility and the individual differences of patients, thereby improving the safety and effectiveness of the prescriptions. The first verification unit can automatically detect problems such as common drug incompatibilities and contraindications in the prescriptions. For example, the system will warn of situations such as "the combination of ginseng and schizandra may cause overexcitation" to help doctors avoid the adverse consequences of drug interactions. The second verification unit combines information such as the patient's allergy history and constitution to further ensure the personalization and safety of the prescriptions. If the herbs recommended by the system may cause allergies or adverse reactions, the system will automatically adjust the herb compatibility to ensure the safety of the prescriptions. The dual verification mechanism greatly reduces the risk of medication errors and ensures a safer and more effective treatment process for patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art.

[0029] Figure 1 It is a view of the system module structure of the present invention;

[0030] Figure 2 It is a view of the data collection and processing flow of the present invention;

[0031] Figure 3 It is a view of the prescription generation process of the present invention;

[0032] Figure 4 It is a view of the construction of the decision tree model of the present invention;

[0033] Figure 5 It is a flowchart of verification and evaluation of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0035] An intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions, including a recommendation module, a symptom database, and a traditional Chinese medicine database. The recommendation module includes a data collection unit, a prescription generation unit, a first verification unit, a second verification unit, and an evaluation unit. In the present invention, the data collection unit is set to collect the experience data of doctors and the four diagnostic information of traditional Chinese medicine. A prescription recommendation model is established with the experience data of doctors and the four diagnostic information of patients as parameters to obtain a recommended prescription. At the same time, the prescription recommendation model is corrected according to the feedback of doctors and patients on the recommended prescription, so that the prescription recommendation model is continuously improved, thus better meeting the actual needs of doctors and patients, and being able to better realize the auxiliary recommendation of prescriptions. The specific operation is realized through the following multiple steps;

[0036] Step 1: Data collection

[0037] The electronic medical record data recorded by doctors in their daily work, including the diagnosis information, treatment plan, drug dosage, etc. of patients, is directly extracted from the hospital information system. Intelligent diagnostic devices are used to collect the health data of patients in real time, such as intelligent bracelets, thermometers, pulse sensors, etc.

[0038] The collected data is de-duplicated, de-noised, and missing value processed to ensure the quality and integrity of the data. The data from different sources is unified in format. For example, data such as pulse and blood pressure is converted into standardized numerical values, and tongue image is converted into feature values.

[0039] The doctor experience data and the four diagnostic information are fused to form a multi-dimensional patient data file for subsequent model training and prescription recommendation. The weighted average method is used for information fusion:

[0040] where, x f use d is the fused data, x i is the value of each data source, and w i is the weight of each data source.

[0041] Step 2: Prescription generation

[0042] Through a rule engine, a preliminary prescription recommendation model is established according to traditional Chinese medicine theory (such as syndrome differentiation and treatment, herbal compatibility principle). According to different input conditions such as symptoms, constitution, medical history, etc., the system automatically selects appropriate herbs, dosages, and taking methods through traditional Chinese medicine theory.

[0043] A preliminary prescription is generated according to the symptom information of the patient. In this process, through the association of the symptom database and the traditional Chinese medicine database, herbs that conform to traditional Chinese medicine theory are selected and recommended. Through the rule engine, a preliminary prescription recommendation is quickly provided based on traditional Chinese medicine theory and preset rules, providing a basic reference framework for the system;

[0044] Query in the symptom database according to the symptoms input by the patient (e.g., "excessive liver fire"). The symptom database contains various common symptoms and their corresponding traditional Chinese medicine syndrome differentiation types and treatment principles. For example, the typical symptoms of "excessive liver fire" will be defined in the symptom database, such as "headache, dizziness, restlessness, dry mouth", etc. After finding the symptom type (e.g., "excessive liver fire"), the system makes recommendations based on the herbal compatibility information in the traditional Chinese medicine database. The traditional Chinese medicine database contains data such as the efficacy of each herb, compatibility principles, and dosage recommendations. For the symptom of "excessive liver fire", traditional Chinese medicine theory recommends using herbs with the effect of clearing liver fire, such as bupleurum and gentiana;

[0045] Recommend appropriate dosages for each herb according to the patient's specific situation (such as age, constitution, allergy history, etc.). The system adjusts the dosage according to traditional Chinese medicine theory and historical data (such as doctor experience, historical prescriptions).

[0046] On this basis, using the doctor's experience data (historical prescriptions) and the patient's four diagnostic information, a machine learning algorithm is used to further train the recommendation model, which enables the prescription recommendation to be not only based on theory but also to comprehensively consider the actually effective prescription combinations in historical data.

[0047] For example, if a certain type of patient often uses certain herb combinations and has good effects, the system will dynamically adjust the prescription generation strategy according to this historical data.

[0048] According to the prescription recommendation model, the system generates a preliminary prescription recommendation, including herbs, dosages, usage methods, etc. For example: "The patient's symptom is headache. It is recommended to use a combination of chrysanthemum and mint, with dosages of 10 grams and 5 grams, to be taken in 2 doses."

[0049] The system makes personalized adjustments according to the patient's specific situation (such as age, gender, constitution, weight, allergy history, etc.) to ensure that the prescription meets the actual needs of the patient.

[0050] For example, if the patient has a weak constitution, some tonifying herbs, such as ginseng and astragalus, may be added to the prescription.

[0051] The system generates multiple alternative plans, provides different treatment strategies according to different symptom manifestations, drug interactions, etc., for doctors to refer to and choose. For example, if the patient is allergic to certain herbs, the system automatically adjusts the herb components.

[0052] In the intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions, based on the preliminary recommendation results of the rule engine, the decision tree algorithm further optimizes by using the historical data of doctors and the individual information of patients (such as age, gender, constitution, allergy history, etc.), adjusts the dosage of herbs, selects drug combinations, etc., to generate personalized traditional Chinese medicine prescription recommendations. Combining the decision tree algorithm with information gain, a decision tree model suitable for traditional Chinese medicine prescription recommendation is constructed. The specific implementation steps are as follows:

[0053] S201. Goals and Input Variables

[0054] The goal is to generate a reasonable traditional Chinese medicine prescription based on the symptoms and constitution of the patient. The input variables include the following:

[0055] Symptom information: such as "headache", "fever", etc.;

[0056] Constitution information: such as "yang deficiency", "yin deficiency", etc.;

[0057] Patient history: such as "previous diseases", "allergy history", etc.;

[0058] Environmental factors: such as "season", "climate", etc.;

[0059] The above input variables will be used as the features of the decision tree algorithm to determine the final traditional Chinese medicine prescription for the patient;

[0060] S202. Decision Tree Model Construction

[0061] The decision tree constructs a tree structure according to the input features in a recursive manner. Each node represents a feature or condition, and each branch represents a certain value of the feature. The final leaf node is a prescription recommendation. The following is the construction process of the decision tree;

[0062] S21. Select Split Feature

[0063] Select the split feature according to the features of the input data (such as symptoms, constitution, etc.). At each node, the system will calculate the optimal split feature according to the information gain formula:

[0064]

[0065] D is the current data set (including the symptoms, constitution, etc. information of all patients);

[0066] A is the currently selected feature (such as "symptom" or "constitution");

[0067] Values(A) are all the values of feature A (for example, "headache" or "fever");

[0068] D vA data subset when the value of feature A is V (for example, the data of all patients with headache);

[0069] Entropy(D) is the entropy of the dataset D, representing the degree of chaos of the current data.

[0070] The information gain calculates the degree of reduction in the uncertainty of the data after splitting the dataset by a certain feature, and thus selects the feature with the largest information gain as the splitting feature of the current node.

[0071] S22. Recursive splitting process

[0072] At each node, calculate the information gain of each feature (such as symptoms, constitution, etc.), select the optimal feature for splitting, until the stopping condition is reached, and generate a decision tree.

[0073] By applying the decision tree algorithm to the intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions, the system can construct a tree model based on information such as the patient's symptoms, constitution, and historical data, and automatically generate personalized prescription recommendations. Each node in the model is split according to the input features (such as symptoms, constitution, drug compatibility, etc.), and finally a traditional Chinese medicine prescription suitable for the patient is generated, ensuring that the prescription recommendation conforms to traditional Chinese medicine theory and is optimized according to the specific needs of the patient.

[0074] Step 3: The first verification unit

[0075] Verify whether the recommended prescription conforms to the compatibility principles and safety standards of traditional Chinese medicine through the set medical rule base. For example, check whether there are drug incompatibilities, taboos between medicinal materials, etc.

[0076] For example, "using ginseng and schizandra together may cause overexcitement", and the system needs to detect such drug taboos and provide warnings.

[0077] In step 3, doctors are allowed to manually review the prescriptions recommended by the system to ensure that the recommended results are consistent with the patient's clinical actual situation. If the prescription recommended by the system is inappropriate, the doctor can adjust or modify the recommendation.

[0078] Step 4: The second verification unit

[0079] In the second verification session, verify the prescription again according to the patient's individual characteristics (such as constitution, allergy history, disease history, etc.).

[0080] For example, for allergic patients, the system will automatically exclude the herbs that may cause allergic reactions to ensure the safety of the prescription.

[0081] Verify whether there are adverse drug interactions between the drugs in the recommended prescription, ensure that the synergistic effect of the drugs is maximized and the side effects are minimized.

[0082] Step 5: Evaluation Unit

[0083] The system conducts tracking and evaluation based on the patient's treatment feedback information (such as symptom improvement, the patient's subjective feelings, etc.), judges the effectiveness of the recommended prescription, and collects treatment effect data through multiple methods such as regular follow-up visits, follow-up data, and health monitoring devices.

[0084] By combining the doctor's experience data and the four diagnostic methods information (inspection, auscultation and olfaction, interrogation, and palpation), the intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions can generate personalized traditional Chinese medicine prescriptions based on the patient's multi-dimensional information (such as symptoms, constitution, historical medical records, etc.). Traditional Chinese medicine prescriptions often rely on the doctor's experience and judgment, while the system, through machine learning algorithms, combines historical data and the patient's specific situation, can provide more accurate prescription recommendations. At the same time, according to personalized factors such as the patient's age, constitution, and allergy history, the dosage and compatibility of herbs are adjusted. This automated and personalized recommendation process significantly improves the effect of traditional Chinese medicine treatment, while reducing the burden on doctors and avoiding the possible subjective biases in traditional methods.

[0085] After generating the prescription recommendation, multiple verifications are carried out through the first verification unit and the second verification unit to ensure that the recommended prescription conforms to the compatibility principles of traditional Chinese medicine and the patient's individual differences, improving the safety and effectiveness of the prescription. The first verification unit can automatically detect whether there are common drug incompatibilities, contraindications, etc. in the prescription, and the second verification unit combines information such as the patient's allergy history and constitution to further ensure the personalization and safety of the prescription. If the herbs recommended by the system may cause allergies or adverse reactions, the system will automatically adjust the herb compatibility to ensure the safety of the prescription.

[0086] The data collection unit collects the doctor's experience and the patient's four diagnostic methods information, combines traditional Chinese medicine theory and machine learning algorithms to generate personalized prescription recommendations. The system preliminarily screens herbs through a rule engine and uses decision tree algorithms to adjust the prescription according to the patient's symptoms, constitution, allergy history, etc., realizing automated and precise recommendations. At the same time, the system ensures the safety and effectiveness of the prescription through multiple-layer verification and continuously optimizes the recommendation model based on the feedback from doctors and patients, thereby improving the effect of traditional Chinese medicine treatment and reducing the burden on doctors.

[0087] The technical means disclosed in the solution of the present invention are not limited to the technical means disclosed above, but also include technical solutions composed of equivalent replacements of the above technical features. Matters not covered by the present invention belong to the common general knowledge of those skilled in the art.

Claims

1. An intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions, characterized in that: It includes a recommendation module, a symptom database, and a traditional Chinese medicine database. The recommendation module includes a data collection unit, a prescription generation unit, a first verification unit, a second verification unit, and an evaluation unit. By setting up the data collection unit to collect doctors' experience data and the four diagnostic information of traditional Chinese medicine, a prescription recommendation model is established with doctors' experience data and patients' four diagnostic information of traditional Chinese medicine as parameters to obtain a recommended prescription. At the same time, the prescription recommendation model is corrected according to the feedback from doctors and patients on the recommended prescription, so that the prescription recommendation model is continuously improved. It specifically includes the following multiple steps: Step 1: Data collection; Step 2: Prescription generation; Step 3: First verification unit; Step 4: Second verification unit; Step 5: Evaluation unit; Through data collection, prescription generation, double verification, and evaluation mechanisms, the system ensures that the recommended traditional Chinese medicine prescriptions not only conform to traditional Chinese medicine theory but also are personalized and accurate, and are continuously improved in practical applications.

2. The intelligent auxiliary recommendation management system for a traditional Chinese medicine prescription according to claim 1, characterized in that: For the data collection, doctors' experience data and four diagnostic information are fused to form a multi-dimensional patient data file, and the weighted average method is used for information fusion; Among them, x f use d is the fused data, x i is the value of each data source, w i is the weight of each data source.

3. The intelligent auxiliary recommendation management system for a traditional Chinese medicine prescription according to claim 1, wherein: For the prescription generation, through a rule engine, a preliminary prescription recommendation model is established according to traditional Chinese medicine theory, and a machine learning algorithm is used to further train the recommendation model.

4. The intelligent auxiliary recommendation management system for a traditional Chinese medicine prescription according to claim 1, wherein: For the prescription generation, the decision tree algorithm is combined with information gain to construct a decision tree model suitable for traditional Chinese medicine prescription recommendation. The specific implementation steps include S201, target and input variables, and S202, decision tree model construction.

5. The intelligent auxiliary recommendation management system for a traditional Chinese medicine prescription according to claim 1, wherein: For the prescription generation, in S201, the target is to generate a reasonable traditional Chinese medicine prescription according to the patient's symptoms and constitution. The input variables include symptom information, constitution information, patient history, and environmental factors. These input variables are used as the features of the decision tree algorithm to determine the patient's final traditional Chinese medicine prescription.

6. The intelligent auxiliary recommendation management system for a traditional Chinese medicine prescription according to claim 1, wherein: For the prescription generation, in S202, the decision tree model construction process includes S21, selecting splitting features, and S22, recursive splitting process. According to the features of the input data, splitting features are selected. At each node, the system calculates the optimal splitting feature according to the information gain formula: The information gain calculation measures the degree of reduction in data uncertainty after splitting the data set by a certain feature, so as to select the feature with the largest information gain as the splitting feature of the current node. At each node, the information gain of each feature is calculated, and the optimal feature is selected for splitting until the stopping condition is reached to generate a decision tree.

7. The intelligent auxiliary recommendation management system for traditional Chinese medicine prescriptions according to claim 1, wherein: The first verification unit verifies whether the recommended prescription conforms to the compatibility principles and safety standards of traditional Chinese medicine through a set medical rule library, and at the same time allows doctors to manually review the prescriptions recommended by the system to ensure that the recommended results are consistent with the patient's clinical actual situation.

8. The intelligent auxiliary recommendation management system for a traditional Chinese medicine prescription according to claim 1, characterized in that: The second verification unit re-verifies the prescription according to the patient's individual characteristics, and at the same time verifies whether there are adverse drug interactions between the drugs in the recommended prescription to ensure that the synergistic effect of the drugs is maximized and the side effects are minimized.

9. The intelligent auxiliary recommendation management system for a traditional Chinese medicine prescription according to claim 1, wherein: The evaluation unit conducts a follow-up evaluation according to the patient's treatment feedback information to judge the effect of the recommended prescription, and collects treatment effect data through multiple methods such as regular follow-up visits, follow-up data, and health monitoring devices.

Images