Method, system and readable medium for recommending electroacupuncture schemes based on big data

Abstract

The present application discloses a method, system and readable medium for recommending electroacupuncture schemes based on big data, which relate to the field of medical information processing technology. The method comprises: obtaining case big data, extracting data that affects treatment effect data as influencing factors, and fitting and constructing a relationship model for characterizing the data relationship between each influencing factor and the treatment effect data; obtaining case data of the current patient, extracting and analyzing the influencing factors and data relationships therein, and matching to obtain multiple reference relationship models; selecting the reference relationship model closest to the current case data, and directly obtaining the influencing factor values ​​corresponding to the treatment scheme data based on the influencing factors that already exist in the reference relationship model or calculating them according to the functional relationship between the influencing factors to generate treatment scheme data. The acupuncture scheme data obtained by the above method is more suitable for the current case needs, and can effectively improve the treatment effect after being output for the doctor's reference.

Description

Technical Field

[0001] The present application relates to the field of medical information processing technology, and in particular to a method, system and readable medium for recommending electroacupuncture schemes based on big data. Background Art

[0002] Currently, some patients are prone to postoperative complications, such as urinary incontinence caused by central or peripheral nerve damage during surgery. Doctors generally use acupuncture to alleviate these symptoms, and some doctors also use electroacupuncture to improve treatment effectiveness.

[0003] Electroacupuncture, a treatment method developed from acupuncture, combines electrical stimulation with traditional acupuncture techniques. An electroacupuncture device is attached to the needle handle, generating electrical signals of a specific frequency and intensity. These signals are then transmitted to the acupuncture points through the needles, simulating bioelectricity and enhancing the stimulating effect of acupuncture. This type of electroacupuncture therapy has shown significant therapeutic effects for some painful conditions and neurological disorders, such as urinary incontinence.

[0004] Before electroacupuncture treatment, optimal treatment results require the correct selection of treatment parameters, including the type of electrical pulse signal (waveform, frequency, pulse width, and current intensity), treatment duration, and electrode placement (e.g., the depth of acupuncture insertion). Incorrect treatment parameters can potentially harm the patient. In practice, each of these treatment parameters requires the physician to make informed decisions based on their experience and individual patient circumstances, placing significant demands on their experience and expertise.

[0005] In order to optimize treatment plans and improve treatment effects, some hospitals currently use data screening systems to provide doctors with electroacupuncture medical data similar to the current patient's symptoms for reference. For example, the data of patients of the same age group, gender, and symptoms as the current case are provided to doctors for reference. Doctors can then provide more accurate treatment plans by comparing the above reference data. However, the data provided to doctors for reference in the above-mentioned electroacupuncture scheme recommendation methods are often fixed. In special cases, even if the reference data are roughly the same, the final treatment effect will vary greatly due to a slight difference in the data. At the same time, each patient's own physiological parameters will also change during the treatment process, and the rules of the above changes vary from person to person. Obviously, the fixed treatment parameters suitable for other patients are likely not completely applicable to the current case.

[0006] In summary, how to generate the best electroacupuncture plan based on electroacupuncture medical big data information and combined with the individual conditions of the current patients for doctors to refer to in order to improve the treatment effect is an urgent problem that needs to be solved. Summary of the Invention

[0007] In view of the problem that in actual applications, electroacupuncture plans need to be generated by doctors based on their experience and judgment, which leads to omissions in electroacupuncture plans, and the reference plans generated by simple matching and screening based on big data are difficult to accurately adapt to different patients, the first purpose of this application is to provide an electroacupuncture plan recommendation method based on big data, which transforms the traditional data matching mode into data relationship matching, and at the same time improves the reference data through artificial intelligence or inquiring doctors and assigns weight values ​​and degrees of freedom to each reference data, and then uses the PCA algorithm to achieve dimensionality reduction and integration of the reference data, so that the assisted electroacupuncture plan can be more accurately adapted to the current patient, helping doctors determine the final treatment plan and improving the treatment effect. In order to realize the above-mentioned electroacupuncture plan recommendation method based on big data, the second purpose of this application is to provide an electroacupuncture plan auxiliary generation system based on big data, and finally propose a computer-readable storage medium to protect the computer program module used to implement the above-mentioned electroacupuncture plan recommendation method. The specific plan is as follows:

[0008] A method for recommending electroacupuncture protocols based on big data, including:

[0009] Obtaining symptom data, treatment plan data, and treatment effect data from each case data, converting each of the symptom data, treatment plan data, and treatment effect data into numerical quantities based on a specific quantification method, preliminarily determining and extracting influencing factors in the symptom data and treatment plan data that affect the treatment effect data, and associating and storing them in a basic database;

[0010] Obtain core factors and marginal factors in impact factors based on the basic database;

[0011] Based on the core factors, multiple relationship models are constructed to characterize the functional relationships between the influencing factors and the treatment effect data, and different weight values ​​are assigned to the influencing factors included in each relationship model, and the influencing factors are stored in association as a model database;

[0012] Obtaining the case data of the current patient, extracting and analyzing the core factors, marginal factors, and their data relationships therein, matching and obtaining multiple relationship models from the model database based on the aforementioned influencing factors and intrinsic data relationships, and taking the relationship model whose treatment effect value exceeds the set value as the reference relationship model;

[0013] Compare the current case data with the influencing factors included in the above reference relationship model to determine the completeness of the number of influencing factors in the current case data and the difference in their values:

[0014] If the number of influencing factors in the current case data is missing, a data request instruction is output to obtain the missing influencing factor data from the doctor side, and / or the missing influencing factor data is automatically generated based on an intelligent algorithm;

[0015] If the number of influencing factors in the current case data is complete, the difference between the influencing factor value in the current case data and the corresponding influencing factor value in the reference relationship model is calculated, and the weighted deviation value is calculated in combination with the weight value of each influencing factor in the reference relationship model;

[0016] Select the reference relationship model with the smallest weighted deviation value, and based on the influencing factors that already exist in the above reference relationship model to characterize the treatment plan data, or according to the functional relationship between the influencing factors, directly or by calculation obtain the influencing factor values ​​corresponding to the treatment plan data, and generate the treatment plan data after conversion and output.

[0017] Through the above technical solution, all data information in the treatment process are quantified and correlation analysis is performed with the treatment effect data, which can effectively eliminate data interference; based on the above correlation analysis, various influencing factors affecting the treatment effect are obtained, and the functional relationship between the treatment effect and the above influencing factors, as well as the influencing factors, are stored as a relational model. In the subsequent process, the data relationship between the influencing factors in the current case data is directly used to find the corresponding relational model, which can avoid the interference of single data and make the final matched relational model more accurate and the matching speed faster; after determining the reference relational model matched by the current case data, the case data can be improved independently or the doctor can be reminded to improve the case data to improve the reliability of the later generated plan. Finally, the data relationship in the reference relational model is used to solve the treatment plan data or directly match it, so that the above treatment plan data is more suitable for the needs of the current case patients. After outputting it for the doctor's reference, it can effectively improve the later treatment effect.

[0018] Furthermore, the disease data includes physiological parameter data of the patient, as well as drug information data and external environment data that affect the above physiological parameter data;

[0019] The treatment plan data includes waveform data, frequency data, pulse width data, current intensity data, application sequence data, application duration data of the electric signal applied by the electroacupuncture instrument to each electric needle, as well as the acupuncture point name information and acupuncture depth data of the above-mentioned electric acupuncture points;

[0020] The treatment effect data include the urinary incontinence frequency, bladder capacity, residual urine volume, urethral pressure, pelvic floor electromyography signal strength, frequency of urinary leakage during specific movements, and recurrence rate of the patients after treatment.

[0021] Through the above technical solutions, we can comprehensively sort out the various factors that affect the treatment effect of incontinence, avoid data omissions, and lay the foundation for providing accurate auxiliary acupuncture plans in the later stage.

[0022] Furthermore, the factors affecting the treatment effect data in the disease data and treatment plan data are preliminarily determined and extracted, including:

[0023] The quantified symptom data and treatment plan data in each case data are used as independent variable data, and the treatment effect data is used as dependent variable data, and the data are associated and stored as a two-dimensional data table;

[0024] Based on mutual information or partial correlation analysis, the disease data and treatment plan data that are not related to the treatment effect data are obtained and excluded to determine the influencing factors.

[0025] Through the above technical solution, when there is enough case data, interference data that is not related to the treatment effect data can be quickly eliminated, which helps to improve the accuracy of subsequent data analysis.

[0026] Furthermore, the method further comprises:

[0027] Based on the theoretical correlation between various data or through the correlation coefficient method, the correlation data with linear relationship between the symptom data and the treatment plan data and the correlation data that has an impact on the treatment plan data are obtained to form a multidimensional data table;

[0028] Perform dimensionality reduction processing on the above-mentioned correlation data based on principal component analysis to generate composite data;

[0029] quantizing the composite data into a numerical value based on the specific quantization method;

[0030] The numerical values ​​corresponding to the composite data and the treatment effect data are subjected to correlation determination, and are marked as core factors or marginal factors according to the determination results and stored.

[0031] Through the above technical solution, some data with linear correlation can be simplified by dimensionality reduction, the complexity of the relationship model can be simplified without affecting the matching of subsequent case data, the speed of data matching can be improved, and the interference of abnormal data during data analysis can be reduced.

[0032] Furthermore, the weighted deviation value is calculated, which also includes:

[0033] According to the property that the influencing factors change over time or under the influence of the patient's environment, the influencing factors are divided into fixed factors and floating factors;

[0034] Obtain and store the correlation between each floating factor associated with the current case patient and time and set environmental factors;

[0035] Obtain and determine the time of electroacupuncture and the patient's environment during treatment;

[0036] Collect the current floating factor data and adjust the floating factor values ​​according to the above time and the patient's environment during treatment;

[0037] The difference between the floating factor value and the corresponding impact factor value in the reference relationship model is calculated, and the weighted deviation value is calculated in combination with the weight value of each impact factor in the reference relationship model.

[0038] Through the above technical solution, the influencing factors that change with time or the environment in which the current case patient is located are separately extracted as floating factors. As a result, the data of various treatment plans will be more accurate and applicable when implementing electroacupuncture surgery, which is conducive to improving the treatment effect of the electroacupuncture plan.

[0039] Furthermore, the method further comprises:

[0040] Obtain the impact factor data corresponding to each impact factor that cannot be exerted simultaneously, and establish a database of mutually exclusive relationships between impact factors;

[0041] Comparing the current case data with the influencing factors included in the reference relationship model also includes determining whether the influencing factors included in the reference relationship model have a mutually exclusive relationship with the influencing factors in the current case data:

[0042] If there are mutually exclusive influencing factors, another relationship model with a treatment effect value exceeding the set value is selected as the reference relationship model;

[0043] If the treatment effect value exceeds the set value and there is always an influencing factor in the relationship model corresponding to the influencing factor that is mutually exclusive with a certain influencing factor in the current case data, an alarm message will be output to the doctor.

[0044] Through the above technical solution, it can be ensured that incompatible influencing factors will not appear in the treatment plan at the same time, such as using auxiliary drugs to which the patient is allergic to treat the patient, thereby ensuring the safety and reliability of the treatment plan.

[0045] A big data-based electroacupuncture plan auxiliary generation system, including a doctor side and a data connection platform side;

[0046] The platform side includes:

[0047] A data acquisition unit configured to acquire symptom data, treatment plan data, and treatment effect data from each case data;

[0048] an influencing factor confirmation unit, configured to convert each of the symptom data, treatment plan data, and treatment effect data into numerical quantities based on a specific quantification algorithm, and determine and extract influencing factors that affect the treatment effect data from the symptom data and treatment plan data based on a built-in correlation analysis algorithm module;

[0049] An impact factor classification unit is configured to classify each impact factor based on a built-in principal component analysis algorithm module to determine the core factors and marginal factors in the impact factors;

[0050] a relationship model construction unit configured to construct, based on the core factors, a plurality of relationship models for characterizing the functional relationships between the influencing factors and the treatment effect data, and assign different weight values ​​to the influencing factors included in each relationship model;

[0051] A basic database configured to store symptom data, treatment plan data, and treatment effect data in each case data, and associated storage of factors affecting the treatment effect data in the symptom data and treatment plan data;

[0052] A model database configured to store the relationship model and the weight value of each influencing factor included in the relationship model in an associated manner;

[0053] A case data parsing unit is configured to obtain and parse the case data of the current patient, and extract and parse the core factors, edge factors and their data relationships therein;

[0054] a relationship model matching unit configured to obtain a plurality of relationship models from the model database according to the influencing factors and their inherent data relationships contained in the current patient case data, and to use the relationship model whose treatment effect value exceeds the set value as a reference relationship model;

[0055] The first data processing unit is configured to compare the current case data with the influencing factors included in the above-mentioned reference relationship model, and determine the completeness of the number of influencing factors in the current case data and the difference in their values: if the number of influencing factors in the current case data is missing, output a data request instruction to obtain the missing influencing factor data from the doctor side, and / or automatically generate the missing influencing factor data based on an intelligent algorithm; if the number of influencing factors in the current case data is complete, calculate the difference between the influencing factor value in the current case data and the corresponding influencing factor value in the reference relationship model, and calculate the weighted deviation value in combination with the weight value of each influencing factor in the reference relationship model;

[0056] The second data processing unit is configured to obtain the weighted deviation value corresponding to each reference relationship model, select the reference relationship model with the smallest weighted deviation value and calculate the influencing factor value corresponding to the treatment plan data based on the functional relationship between the influencing factors in the above reference relationship model, generate the treatment plan data after conversion and output it.

[0057] Furthermore, the system further comprises:

[0058] A floating factor storage unit is configured to be data-connected to the case data analysis unit, and is used to obtain and store the correlation relationship between each floating factor associated with the current case patient and the time and set environmental factors;

[0059] An environmental factor collection unit configured to collect and obtain the electroacupuncture time and the patient's environment during treatment;

[0060] A floating factor data optimization unit is configured to collect current floating factor data and adjust floating factor values ​​according to correlations between the floating factors and time and the patient's environment during treatment;

[0061] The weighted deviation value optimization unit is configured to use the above floating factor data to replace the corresponding impact factor data in the first data processing unit, and calculate and generate an updated weighted deviation value.

[0062] Furthermore, the system further comprises:

[0063] A mutually exclusive relationship database is configured to obtain influencing factor data corresponding to each influencing factor that cannot act simultaneously, and store the mutually exclusive influencing factors in an associated manner;

[0064] A mutually exclusive relationship verification unit is configured to connect with the case data parsing unit and the mutually exclusive relationship database data to obtain and compare whether there is a mutually exclusive relationship between the current case data and the influencing factors included in the reference relationship model;

[0065] A reference model replacement unit is configured to be data-connected to the mutually exclusive relationship verification unit. If the influencing factors in the current case data have a mutually exclusive relationship with the influencing factors in the current reference relationship model, another relationship model whose treatment effect value exceeds the set value is selected as the reference relationship model, and the mutually exclusive relationship determination step is repeated until a relationship model is found in which the influencing factors contained in the relationship model do not have a mutually exclusive relationship with the influencing factors in the current case data, and the above relationship model is output to the first data processing unit as a new reference relationship model, or all relationship models whose treatment effect values ​​exceed the set value are determined to be completed, and an alarm message is output to the alarm unit;

[0066] The alarm unit is configured to be connected to the doctor-side data and is used to output an alarm indicating that the influencing factors have a mutually exclusive relationship.

[0067] Finally, the present application proposes to protect a computer-readable storage medium loaded with a program module for implementing the aforementioned method for recommending electroacupuncture protocols based on big data.

[0068] The above technical solution facilitates the promotion and use of the recommended method for electroacupuncture scheme described in this application.

[0069] In summary, this application has the following beneficial technical effects:

[0070] (1) Quantifying all data information during the treatment process and performing correlation analysis with the treatment effect data can effectively eliminate data interference;

[0071] (2) Based on correlation analysis, various factors affecting the treatment effect are obtained, and the functional relationship between the treatment effect and the above-mentioned factors, as well as the functional relationship between the factors, is stored as a relational model. In the subsequent process, the corresponding relational model can be directly found from the database by directly using the data relationship between the influencing factors in the current case data. This can avoid the interference of a single data, making the final matched relational model more accurate and the matching speed faster;

[0072] (3) After determining the reference relationship model that matches the current case data, the system can automatically complete or remind doctors to complete the missing case data, thereby improving the reliability of the subsequent generated plan;

[0073] (4) The treatment plan data is obtained by solving the data relationship in the reference relationship model or directly matching the treatment plan data from the reference relationship model, so that the final treatment plan data is more suitable for the needs of the current case patients. After the output is provided for the doctor's reference, it can effectively improve the later treatment effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 This is a schematic diagram of the overall recommended method for the electroacupuncture program of this application;

[0075] Figure 2 It is a schematic diagram of a method for optimizing weighted deviation values;

[0076] Figure 3 This is a schematic diagram of the functional modules of the electroacupuncture scheme recommendation system of this application.

[0077] Figure numerals: 1. Data acquisition unit; 2. Influencing factor confirmation unit; 3. Influencing factor classification unit; 4. Relationship model construction unit; 5. Basic database; 6. Model database; 7. Case data parsing unit; 8. Relationship model matching unit; 9. First data processing unit; 10. Second data processing unit. DETAILED DESCRIPTION

[0078] Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings.

[0079] Throughout this specification, reference to the terms "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0080] The present application discloses a method for recommending electroacupuncture regimens based on big data. Figure 1 As shown, it mainly includes the following steps.

[0081] S100, obtaining symptom data, treatment plan data, and treatment effect data from each case data, converting each of the symptom data, treatment plan data, and treatment effect data into numerical values ​​based on a specific quantification method, determining and extracting influencing factors in the symptom data and treatment plan data that affect the treatment effect data, and storing the influencing factors in a related manner in a basic database;

[0082] S200, obtains the core factors and marginal factors of impact factors based on the basic database;

[0083] S300, constructing a plurality of relationship models based on the core factors to characterize the functional relationships between the influencing factors and the treatment effect data, assigning different weight values ​​to the influencing factors included in each relationship model, and storing the associated values ​​in a model database;

[0084] S400, obtaining case data of the current patient, extracting and analyzing the core factors, marginal factors, and their data relationships therein, matching multiple relationship models from the model database based on the aforementioned influencing factors and intrinsic data relationships, and selecting the relationship model whose treatment effect value exceeds the set value as the reference relationship model;

[0085] S500: Compare the current case data with the influencing factors included in the reference relationship model to determine the completeness of the number of influencing factors in the current case data and the difference in their values:

[0086] S510, if the number of influencing factors in the current case data is missing, output a data request instruction to obtain the missing influencing factor data from the doctor side, and / or automatically generate the missing influencing factor data based on an intelligent algorithm;

[0087] S520, if the number of influencing factors in the current case data is complete, then calculating the difference between the influencing factor value in the current case data and the corresponding influencing factor value in the reference relationship model, and combining the weight value of each influencing factor in the reference relationship model to calculate a weighted deviation value;

[0088] S600, select the reference relationship model with the smallest weighted deviation value, and based on the influencing factors that already exist in the above reference relationship model to characterize the treatment plan data or according to the functional relationship between the influencing factors, directly or by calculation obtain the influencing factor values ​​corresponding to the treatment plan data, and generate the treatment plan data after conversion and output.

[0089] In step S100, the case data primarily includes treatment data collected and recorded by various medical institutions for the treatment of incontinence and other conditions using electroacupuncture. The treatment data primarily includes: patient identity information, medical institution information, date information, symptom data, treatment plan data, and treatment effect data. In the embodiments of the present application, during the acquisition of the case data, the data is first cleaned to remove data significantly unrelated to the treatment effect data, such as patient identity information, date information, and medical institution information. This eliminates interference and streamlines the resulting basic database.

[0090] The disease data in step S100 mainly include:

[0091] a. Physiological parameter data of the case patient, including: weight and body mass index (BMI), heart rate, respiratory rate, blood pressure, body temperature, skin resistance, nerve transmission speed, nerve sensitivity, etc.;

[0092] b. Drug information data and external environmental data that affect the above physiological parameter data, including: drug ingredients, ambient temperature, ambient oxygen concentration, and patient age and gender, etc.

[0093] Through the above-mentioned disease data, we can accurately portray the patients in the case data, so that the above-mentioned treatment data information can provide a reference for the treatment plan of subsequent patients.

[0094] Correspondingly, the treatment plan data includes the waveform data, frequency data, pulse width data, current intensity data, application sequence data, application duration data of the electric signal applied by the electroacupuncture instrument to each electric needle, as well as the acupuncture point name information and acupuncture depth data of the above-mentioned electroacupuncture.

[0095] The treatment effect data mainly include the frequency of urinary incontinence, bladder capacity, residual urine volume, urethral pressure, pelvic floor electromyography signal strength, frequency of urinary leakage during specific movements, and recurrence rate of the patients after treatment.

[0096] In step S100, each symptom data, treatment plan data, and treatment effect data are quantified, mainly for non-numerical data information, such as converting the good or bad treatment effect into a numerical value. In one embodiment, the above conversion method can be configured as follows:

[0097] The numerical values ​​of treatment effect data are obtained, such as incontinence frequency f1, bladder capacity L1, urine volume L2, urethral pressure P, pelvic floor electromyographic signal intensity I, frequency of urinary leakage during specific movements f2, and recurrence rate K.

[0098] Then, weights are assigned to the numerical values ​​of each of the above data based on the set criteria. For example, if only incontinence frequency f1 and urine volume L2 are considered when evaluating treatment efficacy, the treatment efficacy can be quantified as Ef = 0.5 * f1 + 0.05 * L2. In practical applications, preferably, the larger the numerical value of the treatment efficacy data, the smaller the weight assigned to it. For example, when urine volume is measured in liters (L), the weight assigned is large, while when urine volume is measured in milliliters (mL), the weight assigned is small. This prevents the quantified treatment efficacy data Ef from being affected by the measurement unit being too large or too small.

[0099] In step S100, factors influencing the treatment effect data from the disease data and treatment plan data are preliminarily determined and extracted, including:

[0100] S110, taking the quantified symptom data and treatment plan data in each case data as independent variable data, taking the treatment effect data as dependent variable data, and storing them in a two-dimensional data table;

[0101] S120, based on the mutual information (MI) or partial correlation analysis method, obtain and exclude disease data and treatment plan data that are not related to the treatment effect data, and determine the influencing factors. Among them, the mutual information method measures the mutual dependence between two variables by the ratio of the joint probability and the marginal probability, while the partial correlation analysis method controls and changes the values ​​of other variables except the two variables to be determined to evaluate whether there is a direct correlation between the two variables. In the embodiment of the present application, considering the different categories of each data, it is preferred to use the mutual information method to evaluate the data correlation.

[0102] In practical applications, there are a large number of influencing factors related to treatment effect data, which increases the difficulty of constructing and matching the subsequent relationship model. Therefore, in steps S100-S200 of the embodiment of the present application, the method further includes:

[0103] A100, based on the theoretical correlation between various data or through the correlation coefficient method, obtains the correlation data between the symptom data and the treatment plan data that has a linear relationship and has an impact on the treatment plan data, and forms a multidimensional data table.

[0104] A200 performs dimensionality reduction processing on the above-mentioned related data based on the principal component analysis method to generate composite data. For example, among the patient's physiological parameter information, the body mass index is a composite data, which represents the proportional relationship between weight and height squared. The body mass index can be used to integrate height and weight data into one data.

[0105] A300, quantizing the composite data into a numerical value based on the specific quantization method;

[0106] A400, performing correlation determination on the numerical values ​​corresponding to the composite data and the treatment effect data, marking them as core factors or marginal factors according to the determination results and storing them.

[0107] Based on the above technical solution, some data with linear correlation can be simplified by dimensionality reduction, which can simplify the complexity of the relationship model without affecting the subsequent case data matching, improve the speed of data matching, and reduce the interference of abnormal data during data analysis.

[0108] In step S200, the division of core factors and marginal factors can be completed by manual designation, such as taking the acupoint number, the wavelength of the electric pulse signal and the application time as the core factors, and the patient's height and weight data as the marginal factors. In actual applications, the core factors and marginal factors can also be obtained through the system processor through principal component analysis.

[0109] In step S300, by analyzing the big data of the treatment effect data and the relevant influencing factor data in the basic database, the correlation between the influencing factors and the treatment effect data can be fitted. The above functional relationship is very clear after fitting through big data. At the same time, the relationship between the influencing factors in step S300 includes the ratio relationship between the numerical values ​​of the influencing factors. For example, the patient's body mass index is a ratio relationship between height and weight. Similarly, the ratio relationship between the numerical values ​​of multiple influencing factors can be counted and generated. When there are enough such ratio relationships, a data profile of the case patient can be made. Even if an influencing factor is missing in the case data, it will not interfere with the subsequent data matching.

[0110] According to the functional relationship between treatment effect data and influencing factors, the treatment effect data can be represented by a functional model, for example:

[0111] Ef=h(α·A,β·B,...,θ·C)+k;

[0112] Where Ef is the quantified value of the treatment effect data, h(A, B, C) is the current relationship model function, A and B are core factors, C is the marginal factor, α, β, and θ are the weights corresponding to these core and marginal factors, and k is the fitting coefficient. In practical applications, the parameters in this relationship model include multiple influencing factors and composite data obtained through dimensionality reduction using the PCA algorithm.

[0113] In step S400, the matching relationship model not only matches based on a single influencing factor, but also matches based on the proportional relationship between the numerical values ​​of multiple influencing factors. This not only improves the matching speed, but also improves the matching accuracy and eliminates interference to a certain extent. It should be noted that the above matching process does not require the numerical values ​​of the influencing factors to be exactly the same. Different influencing factors are configured with different redundancies. For example, when height is used as an influencing factor, the numerical value of a height of 170 cm and the numerical value of a height of 173 cm can be considered matched. However, because the patient's neurosensitivity has a significant impact on the treatment effect, the numerical value of the quantified neurosensitivity is not configured with matching redundancy.

[0114] In actual application, some factors that affect the final therapeutic effect of electroacupuncture will change with time and the changes in the patient's environment. For example, the nerve sensitivity of some patients will change with the change of the temperature of their own environment, and the metabolic rate of some patients will change with time. Obviously, if the corresponding relationship model is directly matched to the model database based on the patient's current symptom data, the obtained relationship model may not be suitable for the patient during the electroacupuncture procedure. In view of the above situation, corresponding to step S520 of this application, as shown in FIG. Figure 2 As shown, calculating the weighted deviation value also includes the following optimization steps:

[0115] B100, based on the property that influencing factors change over time or the patient's environment, the influencing factors are divided into fixed factors and floating factors. The above-mentioned fixed factors are defined as influencing factors that do not change with time and the patient's environment, while floating factors are defined as influencing factors that do change with time and the patient's environment.

[0116] B200 obtains and stores the correlation between each floating factor associated with the current patient, time, and set environmental factors, such as the correlation between the patient's skin resistance and ambient temperature, the correlation between the patient's metabolic rate and time, etc. The above correlation can be a correlation between two floating factors, or a correlation between multiple floating factors and time.

[0117] B300, obtain and determine the electroacupuncture time and the patient's environment during treatment.

[0118] B400 collects the current floating factor data and adjusts the above floating factor values ​​according to the above time and the patient's environment during treatment.

[0119] B500, calculating the difference between the floating factor value and the corresponding impact factor value in the reference relationship model, and combining the weight value of each impact factor in the reference relationship model to calculate the weighted deviation value.

[0120] In practice, factors such as electroacupuncture time and the ambient temperature during treatment are known in advance. Therefore, the symptom data during surgery can be estimated based on the patient's real-time symptom data. This is automatically calculated by the system using the corresponding algorithm module. The doctor only needs to enter parameters such as the surgery time, the ambient temperature during treatment, and the patient's current symptom data into the doctor interface.

[0121] The method for recommending an electroacupuncture regimen according to the embodiment of the present application further includes the following steps:

[0122] C100, obtain the impact factor data corresponding to each impact factor that cannot be exerted simultaneously, and establish a database of mutually exclusive relationships between impact factors;

[0123] C200 compares the influencing factors contained in the current case data and the above-mentioned reference relationship model, and also includes determining whether the influencing factors contained in the reference relationship model are mutually exclusive with the influencing factors in the current case data. For example, if the patient has an allergic reaction to a certain type of drug, the matching reference relationship model should not contain the influencing factors corresponding to the above-mentioned drugs.

[0124] C210, if there are mutually exclusive influencing factors, then select another relationship model whose treatment effect value exceeds the set value as the reference relationship model, and continue to determine whether the newly selected reference relationship model contains influencing factors that are mutually exclusive with the influencing factors contained in the current case data, until a suitable reference relationship model is found;

[0125] C220: If the treatment effect value exceeds the set value, and there is always an influencing factor in the relationship model corresponding to the influencing factor that is mutually exclusive with a certain influencing factor in the current case data, it means that there is no electroacupuncture plan suitable for the current patient in the big data. An alarm message is output to the doctor, who will formulate a more personalized acupuncture plan based on the specific situation.

[0126] In step S500, since there is redundancy in the matching of certain influencing factors, in order to find the relationship model closest to the current case data, a weighted deviation value is introduced as a judgment criterion. The smaller the weighted deviation value, the more suitable the relationship model is for the current case data.

[0127] In step S600, since the numerical values ​​corresponding to the influencing factors are obtained after quantification processing, after obtaining the influencing factors used to characterize the treatment plan data, such as the wavelength and duration of the electric pulse signal, it is also necessary to reversely convert the numerical values ​​corresponding to the above-mentioned influencing shadows into actual treatment plan data, and finally output them to the doctor for reference.

[0128] The above technical solution can ensure that incompatible influencing factors will not appear in the treatment plan at the same time, such as using auxiliary drugs to which the patient is allergic to treat the patient, thereby ensuring the safety and reliability of the treatment plan.

[0129] To implement the aforementioned electroacupuncture protocol recommendation method, embodiments of the present application also disclose a big data-based electroacupuncture protocol-assisted generation system, comprising a physician client and a data-connected platform client. The physician client can be configured as a smartphone, tablet, or PC with data communication and human-computer interaction capabilities. The platform client is preferably deployed in a cloud server to facilitate the processing and storage of big data using the computing power and storage modules of the data center, and to facilitate the collection of electroacupuncture treatment data from various medical institutions.

[0130] Detailed, such as Figure 3 As shown, according to the functional requirements, the platform includes: a data acquisition unit 1, an influencing factor confirmation unit 2, an influencing factor classification unit 3, a relationship model construction unit 4, a basic database 5, a model database 6, a case data analysis unit 7, a relationship model matching unit 8, a first data processing unit 9 and a second data processing unit 10.

[0131] The data acquisition unit 1 is configured to acquire symptom data, treatment plan data, and treatment effect data from each case data. In practice, a server configured in the cloud collects electroacupuncture treatment data uploaded by various medical institutions through a data interface. Preferably, a data cleaning module is configured in the data acquisition unit 1 to remove data such as patient identity information from the treatment data.

[0132] The influencing factor confirmation unit 2 is configured to convert each of the symptom data, treatment plan data and treatment effect data into numerical quantities based on a specific quantification algorithm, and then determine and extract the influencing factors that affect the treatment effect data in the symptom data and treatment plan data based on the built-in correlation analysis algorithm module.

[0133] The impact factor classification unit 3 is configured to classify the impact factors based on the built-in principal component analysis algorithm module or the pre-set classification rules, and determine the core factors and marginal factors in the impact factors.

[0134] The basic database 5 is configured to store the symptom data, treatment plan data and treatment effect data in each case data, and store the influencing factors affecting the treatment effect data in the symptom data and treatment plan data in an associated manner.

[0135] The relationship model construction unit 4 has a built-in data fitting algorithm module, which is configured to connect with the influencing factor classification unit 3 and the basic database 5 data, and to fit and construct multiple relationship models based on the core factors to characterize the functional relationship between the influencing factors and between the influencing factors and the treatment effect data, and automatically or according to the feedback from the doctor's side, assign different weight values ​​to the influencing factors contained in each relationship model.

[0136] The model database 6 is data-connected to the relationship model building unit 4 and is configured to store the relationship model and the weight value of each influencing factor included in the relationship model in an associated manner.

[0137] The case data parsing unit 7 is configured to obtain and parse the case data of the current patient, extract and parse the core factors, edge factors and their data relationships therein, and its specific implementation methods include but are not limited to: using text or OCR image recognition technology to identify and extract specific text or values ​​in the case data, such as identifying the patient's height, weight and other data in the case data, and calculating the data relationship between the above data, such as the ratio between weight and height mentioned above.

[0138] The relationship model matching unit 8 is configured to be data-connected to the case data analysis unit 7, and receives and matches multiple relationship models from the model database 6 based on the influencing factors and their inherent data relationships contained in the current patient case data. The relationship model whose treatment effect value exceeds a set value is used as a reference relationship model. The set value is a comparison threshold used to indicate the quality of the treatment effect.

[0139] The first data processing unit 9 is implemented using a data processing module in a cloud server and is configured to compare the current case data with the influencing factors contained in the above-mentioned reference relationship model based on a built-in program algorithm module to determine the completeness of the number of influencing factors in the current case data and the difference in the values: if the number of influencing factors in the current case data is missing, a data request instruction is output to obtain the missing influencing factor data from the doctor side, and / or the missing influencing factor data is automatically generated based on an intelligent algorithm, such as automatically generating the patient's weight data by combining the body mass index with the patient's height data. If the number of influencing factors in the current case data is complete, the difference between the influencing factor values ​​in the current case data and the corresponding influencing factor values ​​in the reference relationship model is calculated, and the weighted deviation value is calculated in combination with the weight value of each influencing factor in the reference relationship model.

[0140] The second data processing unit 10 is data-connected to the first data processing unit 9 and is configured to obtain the weighted deviation values ​​corresponding to each reference relationship model, select the reference relationship model with the smallest weighted deviation value, and calculate the influencing factor values ​​corresponding to the treatment plan data based on the functional relationship between the influencing factors in the above reference relationship model, generate the treatment plan data after conversion, and output it.

[0141] In detail, in order to adapt to the influencing factors that change with time and the patient's environment, the electroacupuncture plan auxiliary generation system described in the embodiment of the present application also includes: a floating factor storage unit, an environmental factor acquisition unit, a floating factor data optimization unit and a weighted deviation value optimization unit.

[0142] The floating factor storage unit is configured to be data-connected to the case data analysis unit 7 and is used to obtain and store the associations between various floating factors associated with the current patient case and time and set environmental factors. Such associations include functional relationships between the numerical values ​​of certain influencing factors and time or environmental factors. The environmental factor acquisition unit is configured to collect and obtain electroacupuncture time and the patient's environment during treatment. The environmental factor acquisition unit is configured to be data-connected to the doctor's terminal, and the doctor inputs such data information from the doctor's terminal.

[0143] The floating factor data optimization unit is configured to collect current floating factor data and adjust the floating factor value based on the correlation between the floating factor and time and the patient's environment during treatment. The current floating factor data can be collected by the doctor and then input to the platform via the doctor's end, such as collecting the patient's current heart rate and blood pressure data. The weighted deviation value optimization unit is configured to be data-connected to the above-mentioned floating factor data optimization unit and the first data processing unit 9, obtain and use the above-mentioned floating factor data to replace the corresponding influencing factor data in the first data processing unit 9, and then calculate and generate an updated weighted deviation value and output it.

[0144] In order to prevent two mutually exclusive influencing factors on treatment efficacy from appearing in the reference relationship model, the system further includes: a mutually exclusive relationship database, a mutually exclusive relationship verification unit, a reference model replacement unit, and an alarm unit.

[0145] The mutually exclusive relationship database is configured to obtain the influencing factor data corresponding to each influencing factor that cannot act simultaneously, and store the mutually exclusive influencing factors in an associated manner. In actual applications, the mutually exclusive influencing factor data needs to be manually input for storage.

[0146] The mutually exclusive relationship verification unit is configured to be connected to the case data analysis unit 7 and the mutually exclusive relationship database data, and obtain and compare whether there is a mutually exclusive relationship between the current case data and the influencing factors contained in the above-mentioned reference relationship model. The reference model replacement unit is configured to be connected to the data of the mutually exclusive relationship verification unit. If the influencing factors in the current case data and the influencing factors in the current reference relationship model have a mutually exclusive relationship, then another relationship model with a treatment effect value exceeding the set value is selected as the reference relationship model, and the mutually exclusive relationship determination step is repeated until a relationship model is found in which the influencing factors contained in the relationship model do not have a mutually exclusive relationship with the influencing factors in the current case data, and the above-mentioned relationship model is output to the first data processing unit 9 as a new reference relationship model. If all relationship models with treatment effect values ​​exceeding the set value are determined to be completed, an alarm message is output to the alarm unit. The alarm unit is configured to be connected to the doctor's end data, and after receiving the alarm message that the above-mentioned influencing factors have a mutually exclusive relationship, it is output to the doctor's end.

[0147] Finally, the present application proposes to protect a computer-readable storage medium loaded with a program module for implementing the aforementioned method for recommending electroacupuncture protocols based on big data.

[0148] The above-mentioned computer-readable storage media include but are not limited to disk storage, CD-ROM, optical storage, etc., which facilitate the promotion and use of the electroacupuncture scheme recommendation method described in this application.

[0149] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for recommending electroacupuncture regimens based on big data, characterized in that: include: Acquire and store the symptom data, treatment plan data, and treatment effect data in each case data, convert each of the symptom data, treatment plan data, and treatment effect data into numerical values ​​based on a specific quantification method, preliminarily determine and extract the influencing factors that affect the treatment effect data in the symptom data and treatment plan data, and associate and store them into a basic database (5); Obtain the core factors and marginal factors of impact factors based on the basic database (5); Based on the core factors, multiple relationship models are constructed to characterize the functional relationships between the influencing factors and the treatment effect data, and different weight values ​​are assigned to the influencing factors included in each relationship model, and the associated values ​​are stored as a model database (6); Obtain the case data of the current patient, extract and analyze the core factors, marginal factors and their data relationships, match and obtain multiple relationship models from the model database (6) based on the above-mentioned influencing factors and internal data relationships, and use the relationship model with a treatment effect value exceeding the set value as the reference relationship model; Compare the current case data with the influencing factors included in the above reference relationship model to determine the completeness of the number of influencing factors in the current case data and the difference in their values: If the number of influencing factors in the current case data is missing, a data request instruction is output to obtain the missing influencing factor data from the doctor side, and / or the missing influencing factor data is automatically generated based on an intelligent algorithm; If the number of influencing factors in the current case data is complete, the difference between the influencing factor value in the current case data and the corresponding influencing factor value in the reference relationship model is calculated, and the weighted deviation value is calculated in combination with the weight value of each influencing factor in the reference relationship model; Selecting a reference relationship model with the smallest weighted deviation value, and directly or by calculation obtaining the influencing factor values ​​corresponding to the treatment plan data based on the influencing factors already existing in the reference relationship model for characterizing the treatment plan data, or based on the functional relationship between the influencing factors, converting the values ​​to generate the treatment plan data and outputting the data; The weighted deviation value is calculated, and also includes: According to the property that the influencing factors change over time or under the influence of the patient's environment, the influencing factors are divided into fixed factors and floating factors; Obtain and store the correlation between each floating factor associated with the current case patient and time and set environmental factors; Obtain and determine the time of electroacupuncture and the patient's environment during treatment; Collect the current floating factor data and adjust the floating factor values ​​according to the above time and the patient's environment during treatment; The difference between the floating factor value and the corresponding impact factor value in the reference relationship model is calculated, and the weighted deviation value is calculated in combination with the weight value of each impact factor in the reference relationship model.

2. The method for recommending electroacupuncture regimens based on big data according to claim 1, characterized in that: The disease data includes physiological parameter data of the patient, as well as drug information data and external environment data that affect the above physiological parameter data; The treatment plan data includes waveform data, frequency data, pulse width data, current intensity data, application sequence data, application duration data of the electric signal applied by the electroacupuncture instrument to each electric needle, as well as the acupuncture point name information and acupuncture depth data of the above-mentioned electric acupuncture points; The treatment effect data include the frequency of urinary incontinence, bladder capacity, residual urine volume, urethral pressure, pelvic floor electromyography signal strength, frequency of urinary leakage during specific movements, and recurrence rate of the patients after treatment.

3. The method for recommending electroacupuncture regimens based on big data according to claim 1, characterized in that: Preliminary determination and extraction of factors affecting treatment effect data from disease data and treatment plan data, including: The quantified symptom data and treatment plan data in each case data are used as independent variable data, and the treatment effect data is used as dependent variable data, and the data are associated and stored as a two-dimensional data table; Based on mutual information or partial correlation analysis, the disease data and treatment plan data that are not related to the treatment effect data are obtained and excluded to determine the influencing factors.

4. The method for recommending electroacupuncture regimens based on big data according to claim 3, characterized in that: The method further comprises: Based on the theoretical correlation between various data or through the correlation coefficient method, the correlation data with linear relationship between the symptom data and the treatment plan data and the correlation data that has an impact on the treatment plan data are obtained to form a multidimensional data table; Perform dimensionality reduction processing on the above-mentioned correlation data based on principal component analysis to generate composite data; quantizing the composite data into a numerical value based on the specific quantization method; The numerical values ​​corresponding to the composite data and the treatment effect data are subjected to correlation determination, and are marked as core factors or marginal factors according to the determination results and stored.

5. The method for recommending electroacupuncture regimens based on big data according to claim 1, characterized in that: The method further comprises: Obtain the impact factor data corresponding to each impact factor that cannot be exerted simultaneously, and establish a database of mutually exclusive relationships between impact factors; Comparing the current case data with the influencing factors included in the reference relationship model also includes determining whether the influencing factors included in the reference relationship model have a mutually exclusive relationship with the influencing factors in the current case data: If there are mutually exclusive influencing factors, another relationship model with a treatment effect value exceeding the set value is selected as the reference relationship model to re-determine the mutually exclusive relationship; If the treatment effect value exceeds the set value and there is always an influencing factor in the relationship model corresponding to the influencing factor that is mutually exclusive with a certain influencing factor in the current case data, an alarm message will be output to the doctor.

6. A big data-based electroacupuncture program auxiliary generation system, including a doctor side and a data connection platform side, characterized in that: The platform side includes: A data acquisition unit (1) is configured to acquire symptom data, treatment plan data, and treatment effect data from each case data; An influencing factor confirmation unit (2) is configured to convert the symptom data, treatment plan data, and treatment effect data into numerical values ​​based on a specific quantification algorithm, and determine and extract influencing factors that affect the treatment effect data from the symptom data and treatment plan data based on a built-in correlation analysis algorithm module; An impact factor classification unit (3) is configured to classify each impact factor based on a built-in principal component analysis algorithm module, and determine core factors and marginal factors in the impact factors; A relationship model construction unit (4) is configured to construct a plurality of relationship models based on the core factors to characterize the functional relationships between the influencing factors and the treatment effect data, and to assign different weight values ​​to the influencing factors included in each relationship model; A basic database (5) is configured to store symptom data, treatment plan data, and treatment effect data in each case data, and associated storage of factors affecting the treatment effect data in the symptom data and treatment plan data; A model database (6) configured to store the relationship model and the weight value of each influencing factor contained in the relationship model in an associated manner; A case data parsing unit (7) is configured to obtain and parse the case data of the current patient, extract and parse the core factors, edge factors and their data relationships therein; A relationship model matching unit (8) is configured to obtain a plurality of relationship models from the model database (6) according to the influencing factors and their inherent data relationships contained in the current patient case data, and to use the relationship model whose treatment effect value exceeds the set value as a reference relationship model; The first data processing unit (9) is configured to compare the influencing factors included in the current case data and the reference relationship model, and determine the completeness of the number of influencing factors in the current case data and the difference in the values: if the number of influencing factors in the current case data is missing, output a data request instruction to obtain the missing influencing factor data from the doctor end, and / or automatically generate the missing influencing factor data based on an intelligent algorithm; if the number of influencing factors in the current case data is complete, calculate the difference between the value of the influencing factor in the current case data and the value of the corresponding influencing factor in the reference relationship model, and calculate the weighted deviation value in combination with the weight value of each influencing factor in the reference relationship model; The second data processing unit (10) is configured to obtain weighted deviation values ​​corresponding to each reference relationship model, select the reference relationship model with the smallest weighted deviation value, and calculate the influencing factor values ​​corresponding to the treatment plan data based on the functional relationship between the influencing factors in the reference relationship model, generate the treatment plan data after conversion, and output the data; the system also includes: A floating factor storage unit is configured to be data-connected to the case data analysis unit (7) and is used to obtain and store the correlation relationship between each floating factor associated with the current case patient and time and set environmental factors; An environmental factor collection unit configured to collect and obtain the electroacupuncture time and the patient's environment during treatment; A floating factor data optimization unit is configured to collect current floating factor data and adjust floating factor values ​​according to correlations between the floating factors and time and the patient's environment during treatment; The weighted deviation value optimization unit is configured to use the floating factor data to replace the corresponding impact factor data in the first data processing unit (9), and calculate and generate an updated weighted deviation value.

7. The electroacupuncture plan auxiliary generation system based on big data according to claim 6 is characterized in that: The system further comprises: A mutually exclusive relationship database is configured to obtain influencing factor data corresponding to each influencing factor that cannot act simultaneously, and store the mutually exclusive influencing factors in an associated manner; a mutually exclusive relationship verification unit configured to connect with the case data parsing unit (7) and the mutually exclusive relationship database data, obtain and compare whether there is a mutually exclusive relationship between the current case data and the influencing factors included in the reference relationship model; The reference model replacement unit is configured to be connected to the data of the mutually exclusive relationship verification unit. If the influencing factors in the current case data and the influencing factors in the current reference relationship model have a mutually exclusive relationship, another relationship model whose treatment effect value exceeds the set value is selected as the reference relationship model, and the mutually exclusive relationship determination step is repeated until a relationship model is found in which the influencing factors contained in the relationship model and the influencing factors in the current case data do not have a mutually exclusive relationship. The above relationship model is output to the first data processing unit (9) as a new reference relationship model, or all relationship models whose treatment effect values ​​exceed the set value are determined to be completed, and an alarm message is output to the alarm unit; The alarm unit is configured to be connected to the doctor-side data and is used to output an alarm indicating that the influencing factors have a mutually exclusive relationship.

8. A computer-readable storage medium, characterized in that A program module for implementing the big data-based electroacupuncture scheme recommendation method as described in any one of claims 1 to 5 is loaded thereon.

Images