Personalized acupuncture treatment scheme recommendation method and system based on data analysis

Abstract

The invention discloses a personalized acupuncture treatment scheme recommendation method and system based on data analysis, and relates to the technical field of acupuncture. The method comprises the following steps: constructing a feature vector matrix according to multi-dimensional feature data of a target patient; taking the matrix as a constraint, retrieving similar historical cases in a pre-constructed acupuncture case database, and obtaining a sample case set and a similarity set; configuring a data selection strategy based on the similarity set, constructing a sample training set, and training to generate an acupuncture parameter evaluation plug-in; and defining an acupuncture parameter optimization space based on the sample case set, performing iterative optimization in the space by using the evaluation plug-in, and finally outputting an acupuncture treatment parameter scheme adaptive to the target patient. According to the method, the whole-process personalized recommendation of the acupuncture treatment scheme from case screening, model evaluation to parameter optimization is realized, and the accuracy, objectivity and clinical operability of treatment are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of acupuncture, in particular to an individualized acupuncture treatment scheme recommendation method and system based on data analysis. BACKGROUND

[0002] As an important treatment method of traditional Chinese medicine, the efficacy of acupuncture has been verified in the clinical practice of various diseases. However, the traditional acupuncture treatment scheme is highly dependent on the personal experience and subjective judgment of the physician. With the development of modern medical information, a large amount of patient feature data and historical treatment records have been generated in various diagnosis and treatment processes, which contain valuable information for guiding individualized treatment. How to use data analysis technology to mine the rules and build an objective and quantitative decision model to assist in generating accurate acupuncture treatment schemes has become a key problem to improve the scientific and standardized level of acupuncture clinical practice.

[0003] The existing technical solutions have obvious deficiencies in processing individualized recommendation of acupuncture treatment. Some studies attempt to apply simple data matching or rule-based reasoning, but it is difficult to effectively handle the complex nonlinear relationships and high-dimensional parameter combinations in acupuncture treatment. Another recommendation system based on general machine learning models often lacks deep integration of knowledge in the field of acupuncture, and cannot accurately model the complex relationship between specific acupuncture parameters and efficacy. More importantly, existing methods usually treat case retrieval and parameter optimization in isolation, and fail to build a complete closed-loop system from similar case discovery to individualized evaluation model construction and then to automatic optimization of treatment parameters, resulting in insufficient individualization of the recommended scheme, poor interpretability, and difficulty in dynamically adapting to the unique physiological and pathological characteristics of different patients, which cannot meet the urgent needs of clinical high-precision and high-adaptability acupuncture treatment schemes. SUMMARY

[0004] The present application provides an individualized acupuncture treatment scheme recommendation method and system based on data analysis to address the technical problems of excessive reliance on subjective experience of physicians, lack of objective data support, and difficulty in integrating knowledge in the field of acupuncture in existing data analysis methods.

[0005] The technical solution of the present application to solve the above technical problems is as follows: In a first aspect, the present application provides an individualized acupuncture treatment scheme recommendation method based on data analysis, comprising: constructing a feature vector matrix according to the patient feature data of the target patient; performing similar traversal in the pre-constructed acupuncture case database with the feature vector matrix as a constraint, selecting historical acupuncture cases with a similarity greater than a preset similarity threshold as sample acupuncture cases, and obtaining a sample acupuncture case set and a sample feature similarity set; setting a data selection strategy based on the sample feature similarity set, performing data selection in the sample acupuncture case set to obtain a sample training set, and constructing an acupuncture parameter evaluation plug-in using the sample training set; constructing an acupuncture parameter optimization space based on the sample acupuncture case set, performing iterative optimization of acupuncture parameters in the acupuncture parameter optimization space using the acupuncture parameter evaluation plug-in, and outputting adaptive acupuncture parameters as an adaptive acupuncture treatment scheme for the target patient.

[0006] In a second aspect, the present application provides a personalized acupuncture treatment scheme recommendation system based on data analysis, comprising: a feature vector construction module configured to construct a feature vector matrix based on patient feature data of a target patient; a similar case retrieval module configured to perform similar traversal in a pre-constructed acupuncture case database with the feature vector matrix as a constraint, select historical acupuncture cases with a similarity greater than a preset similarity threshold as sample acupuncture cases, and obtain a sample acupuncture case set and a sample feature similarity set; an evaluation plug-in construction module configured to set a data selection strategy based on the sample feature similarity set, perform data selection in the sample acupuncture case set to obtain a sample training set, and construct an acupuncture parameter evaluation plug-in using the sample training set; a treatment scheme generation module configured to construct an acupuncture parameter optimization space based on the sample acupuncture case set, perform iterative optimization of acupuncture parameters in the acupuncture parameter optimization space using the acupuncture parameter evaluation plug-in, and output adaptive acupuncture parameters as an adaptive acupuncture treatment scheme for the target patient.

[0007] The present application has the following advantages: Compared with the prior art, the present application first constructs a multi-dimensional patient feature vector matrix and performs intelligent historical case similarity retrieval based thereon, providing a highly relevant and personalized data basis for subsequent analysis. Secondly, a weighted training set is dynamically constructed based on the retrieved case similarity, and an acupuncture parameter evaluation plug-in is trained and generated, which can accurately quantify the expected efficacy of different treatment parameter combinations. Thirdly, an acupuncture parameter optimization space is constructed, and an iterative optimization algorithm is guided to search automatically in this space using the personalized evaluation plug-in, realizing the intelligent search for the optimal treatment scheme from a large number of parameter combinations. Finally, the case retrieval, model construction, and parameter optimization are organically integrated into a complete, data-driven closed-loop recommendation process, which not only improves the personalization and precision of the acupuncture treatment scheme, but also enhances the explainability and clinical credibility of the entire recommendation process, providing a systematic solution for intelligent auxiliary decision-making in acupuncture treatment. BRIEF DESCRIPTION OF DRAWINGS

[0008] Figure 1A flowchart of a personalized acupuncture treatment scheme recommendation method based on data analysis provided by the present application is shown in the figure. Figure 2 A structure diagram of a personalized acupuncture treatment scheme recommendation system based on data analysis provided by the present application is shown in the figure.

[0009] In the figure, the components represented by the respective numbers are as follows: The feature vector construction module 11, the similar case retrieval module 12, the evaluation plug-in construction module 13, and the treatment scheme generation module 14. DETAILED DESCRIPTION

[0010] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0011] In the description of the present application, the terms "first", "second", etc. are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second", etc. can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0012] In the description of the present application, the term "for example" is used to indicate "as an example, illustration or explanation". Any embodiment described as "for example" in the present application is not necessarily interpreted as more preferred or more advantageous than other embodiments. The following description is given in order to enable any person skilled in the art to implement and use the present application. In the following description, details are listed for the purpose of explanation. It should be understood that those skilled in the art can recognize that the present application can be implemented without using these specific details. In other examples, well-known structures and processes will not be described in detail in order to avoid unnecessary details making the description of the present application obscure. Therefore, the present application is not intended to be limited to the shown embodiments, but is consistent with the broadest scope of principles and features disclosed.

[0013] Embodiment one, as shown in the present application, provides a personalized acupuncture treatment scheme recommendation method based on data analysis, which comprises: Figure 1 S10: constructing a feature vector matrix according to patient feature data of a target patient; S10: constructing a feature vector matrix according to patient feature data of a target patient; The patient characteristic data includes basic physical sign data, basic condition data, imaging data, biochemical index data, electrophysiological data, and acupuncture region body fat distribution.

[0014] Firstly, multi-dimensional diagnosis and treatment information of the target patient, i.e., patient characteristic data, needs to be comprehensively collected. The target patient refers to a specific patient individual who currently needs to be treated with acupuncture treatment scheme individualization recommendation and formulation. The patient characteristic data covers multiple categories, specifically including basic physical sign data reflecting physiological state, basic condition data describing disease condition, imaging data providing morphological basis, biochemical index data representing internal biochemical environment, electrophysiological data recording neural and muscular electrical activity law, and acupuncture region body fat distribution data characterizing soft tissue morphology of a specific acupuncture region.

[0015] The imaging data provides visualized information about internal structure and function of the body through medical imaging technology, and is used to assist in locating lesion site and judging its nature. The specific content of the imaging data includes but is not limited to: X-ray image for observing bone structure; computed tomography image providing clearer cross-sectional image of the body, especially suitable for observing fine structure of bone and internal organs; magnetic resonance imaging image showing excellent effect on muscle, ligament, intervertebral disc, cartilage and nerve; ultrasonic image commonly used for observing real-time dynamic image of soft tissues such as muscle, tendon and ligament; and infrared thermal image reflecting blood circulation condition by detecting surface temperature distribution, which can provide direct and objective basis for acupuncture point selection.

[0016] Biochemical index data reveals the metabolism, endocrine and immune state of the body from a microscopic level, and provides objective evidence for TCM internal pathogenesis theory. The specific content of the biochemical index data includes but is not limited to: blood routine index for understanding infection or anemia condition; index for judging whether there is inflammatory reaction in the body, such as C-reactive protein and erythrocyte sedimentation rate; hormone level data closely related to diseases related to endocrine disorders, such as thyroid hormone, sex hormone and cortisol; and blood glucose, blood lipid and liver and kidney function indexes reflecting body metabolism condition and related to TCM pathogenesis such as phlegm-dampness and blood stasis.

[0017] Electrophysiological data directly reflects the functional state of meridians and nervous system by measuring bioelectricity. The specific content of the electrophysiological data includes but is not limited to: acupoint electrical property detection data for providing quantitative reference for meridian differentiation and point selection by measuring resistance or conductivity and potential of specific acupoints through instruments; and electromyography data for evaluating nerve and muscle function and recording electrical activity of muscle at rest and contraction.

[0018] The acupuncture area body fat distribution data is specific morphological data for the local area of acupuncture treatment, which directly affects the accuracy of point selection and operation method. The acupuncture area body fat distribution data specifically includes: the subcutaneous fat layer thickness of a specific acupoint or treatment area measured by B-ultrasound or skinfold caliper; classification information representing the type of fat distribution; and local morphological characteristics such as the body surface contour and muscle fullness of the treatment area. The collection of acupuncture area body fat distribution data has direct guiding significance for determining the depth and angle of acupuncture, evaluating the conduction effect of needle sensation, and selecting appropriate stimulation methods.

[0019] Secondly, the various types of raw patient feature data collected above are standardized and preprocessed. The preprocessing process includes data cleaning to handle missing values and outliers, and normalization or standardization operations to eliminate differences in dimensions and orders of magnitude of different data indicators, ensuring fairness and effectiveness of subsequent calculations.

[0020] After preprocessing, the various types of standardized patient feature data are combined and vectorized according to the predetermined order and structure. Specifically, the patient feature data of each dimension is mapped to a specific dimension value in the feature vector. All dimensions are arranged in order and spliced together to form a feature vector matrix that can comprehensively and structurally represent the individual condition of the target patient. The final obtained feature vector matrix is a digital comprehensive representation carrier integrating multi-source heterogeneous medical information, which serves as a digital fingerprint of the patient's individual condition, providing a unified and standardized mathematical basis for subsequent case similarity measurement and data analysis.

[0021] S20: With the feature vector matrix as a constraint, similar traversal is performed in the pre-constructed acupuncture case database, and historical acupuncture cases with a similarity greater than a preset similarity threshold are selected as sample acupuncture cases to obtain a sample acupuncture case set and a sample feature similarity set; Among them, the acupuncture case database includes several historical acupuncture cases, wherein the historical acupuncture cases include historical patient feature vector matrices, historical acupuncture treatment parameters and historical treatment quality coefficients, wherein the acupuncture treatment parameters include acupoint distribution and acupuncture parameter groups, and the acupuncture parameter groups include lifting amplitude, lifting frequency, rotation angle, rotation speed, acupuncture duration and needle retention duration.

[0022] Among them, the historical treatment quality coefficient is determined based on the improvement rate of the specific disease functional scale score of the historical patient, the change rate of the historical digital pain score, and the number of historical treatments required to reach the minimum clinically important difference.

[0023] After the feature vector matrix is constructed, since the individual feature data of a single patient is insufficient to support the direct construction of a complex treatment scheme model, and directly modeling globally from a large number of historical cases will introduce a large amount of irrelevant data noise, affecting the accuracy and individualization degree of the scheme, therefore, the feature vector matrix needs to be used as a constraint condition to search for similar historical acupuncture cases in the pre-constructed acupuncture case database, and select historical acupuncture cases with a similarity greater than a pre-set similarity threshold as sample acupuncture cases, to obtain a sample acupuncture case set and a sample feature similarity set.

[0024] The pre-constructed acupuncture case database is a digital knowledge base composed of long-term accumulated structured clinical case data, and a large number of historical acupuncture cases are pre-stored in the acupuncture case database, each historical acupuncture case being a structured record unit, including three key components: first, a historical patient feature vector matrix of a historical patient before treatment, which is consistent with the search reference structure in this step; second, historical acupuncture treatment parameters received by the historical patient, which specifically define the specific treatment operation, including the selected acupoint distribution and acupuncture parameter group, which is further refined into lifting amplitude, lifting frequency, rotation angle, rotation speed, acupuncture duration and needle retention duration; and third, a historical treatment quality coefficient representing the final treatment effect.

[0025] Specifically, the historical treatment quality coefficient is a key indicator for comprehensive quantitative evaluation of treatment effect. Its specific calculation process is based on the changes of a plurality of objective clinical indicators recorded by the historical patient before and after treatment, and is determined by comprehensive evaluation. The clinical indicators used for evaluation mainly include three categories: the first category is the improvement rate of the specific disease function scale score, which aims to measure the improvement degree of the actual life ability and function of the patient after treatment; for example, the score change of the patient in the Oswestry Disability Index and other standardized scales before and after treatment can effectively reflect the depth of functional recovery. The second category is the change rate of digital pain score, which is the most direct and classical quantitative method for evaluating the relief degree of core pain symptoms; the patient scores within the numerical range from no pain to the most severe pain according to his own feeling, and the change rate provides objective data of the symptom relief intensity. The third category of indicators is the number of historical treatments required to reach the minimum clinically important difference; the minimum clinically important difference represents the change threshold corresponding to the minimum improvement degree that the patient can perceive; the fewer the number of treatments required to reach the threshold, the faster the corresponding treatment scheme takes effect and the higher the treatment efficiency, thereby introducing the time and efficiency dimensions of the efficacy evaluation.

[0026] By integrating the above multiple dimension indicators respectively representing the depth, intensity and efficiency of the treatment effect, and using a weighted fusion method, a comprehensive, objective and quantitative historical treatment quality coefficient can be calculated. Among them, each weight coefficient is pre-set according to expert clinical experience consensus or based on statistical analysis results of a large number of historical treatment data, for example, in the evaluation of the treatment effect on chronic musculoskeletal pain, the improvement rate of the specific disease function scale score can be given a relatively high weight of 0.6 to emphasize the fundamental and long-term benefits of functional recovery; the digital pain score change rate, as a direct reflection of the core symptoms, can be given a weight of 0.3 to represent its intensity importance; and the number of historical treatments required to reach the minimum clinically important difference is given a weight of 0.1 as an indicator of treatment efficiency to reflect its time cost consideration in treatment plan selection. The historical treatment quality coefficient calculated can effectively evaluate the comprehensive advantages and disadvantages of the historical treatment plan, and provide high-quality supervision signals for subsequent construction of acupuncture parameter evaluation plug-ins.

[0027] Further, when performing similarity traversal, the similarity between the feature vector matrix of the target patient and the historical patient feature vector matrix of each historical case in the acupuncture case database needs to be calculated. Specifically, the similarity calculation can be implemented using cosine similarity, Euclidean distance, Mahalanobis distance, etc.

[0028] For example, the direction consistency of two vectors in a multi-dimensional space can be measured by calculating the cosine similarity between the two vectors. The specific calculation process is as follows: the feature vector matrix of the target patient and the historical patient feature vector matrix of any historical case are respectively regarded as two multi-dimensional vectors, the dot product of the two vectors is calculated, and then the dot product result is divided by the product of the respective lengths of the two vectors. The ratio is the cosine similarity, which takes a value between -1 and 1. When the directions of the two vectors are completely the same, the value is 1, when the directions are opposite, the value is -1, and when the directions are perpendicular, the value is 0. This cosine similarity algorithm can effectively quantify the closeness of two cases in the patient feature dimension. The larger the cosine similarity obtained by calculation, the more consistent the directions of the two cases in the multi-dimensional feature space, the higher the similarity of the patient features, indicating that they have higher comparability and reference value in physiological state, disease manifestation and other dimensions; on the contrary, the smaller the cosine similarity, the greater the difference in the direction of the two cases in the feature space, the more dissimilar the composition mode of the patient features, indicating that they have lower commonality or reference value in individual condition.

[0029] After the similarity calculation of all historical cases is completed, each calculated similarity value is compared with a preset similarity threshold. The preset similarity threshold is an empirical parameter that can be adjusted based on specific clinical needs and data distribution characteristics, and its core function is to define the minimum similarity boundary required in the feature space between the effective reference case and the target patient. The setting of the similarity threshold can be obtained according to the experience of field experts or historical data analysis. For example, in the application scenario of chronic pain management, the similarity threshold can be initially set to 0.7 to ensure that the retrieved historical cases have strong consistency with the target patient in key features.

[0030] Specifically, only the historical cases with a similarity value strictly greater than the preset similarity threshold are determined to be effective samples with sufficient reference value and are formally selected. All selected effective historical cases collectively constitute a sample acupuncture case set, and their respective corresponding similarity values constitute a sample feature similarity set. This screening mechanism can ensure that the data basis provided for subsequent model construction and parameter optimization is derived from effective historical treatment records that are highly relevant to the individual condition of the target patient, thereby laying a precise and reliable data premise for constructing an evaluation model with strong individualization ability and subsequent treatment scheme optimization process.

[0031] S30: setting a data selection strategy based on the sample feature similarity set, performing data selection in the sample acupuncture case set to obtain a sample training set, and constructing an acupuncture parameter evaluation plug-in using the sample training set; Specifically, based on the sample feature similarity set, a data selection strategy is set, and data selection is performed in the sample acupuncture case set to obtain a sample training set, including: The ratio of the sample feature similarity of each sample acupuncture case to the mean of the sample feature similarity set is set as a data selection compensation coefficient, the product of the data selection compensation coefficient and 10 is rounded to obtain a data selection frequency, and the data selection frequency set is calculated based on the sample feature similarity set; According to the data selection frequency set, data is repeatedly selected in the sample acupuncture case set to construct a sample training set.

[0032] Further, a deep learning model is trained using the sample training set with the historical patient feature vector matrix and historical acupuncture treatment parameters as input and the historical treatment quality coefficient as supervision until convergence, generating an acupuncture parameter evaluation plug-in.

[0033] Specifically, based on the sample feature similarity set, a data selection strategy is set, a sample training set is obtained by data selection in the sample acupuncture case set, and an acupuncture parameter evaluation plug-in is constructed. The step aims to further construct a sample training set that can effectively train the prediction model and emphasizes individualized features from the screened sample acupuncture case set. The core logic is to give historical cases with higher similarity to the target patient greater influence, so that the training process focuses more on the treatment experience that is highly adapted to the current patient.

[0034] First, the arithmetic mean of the sample feature similarity set is calculated as a benchmark reference. For each case in the sample acupuncture case set, the sample feature similarity corresponding to the case is compared with the average value, the ratio between the two is calculated, and the ratio is defined as the data selection compensation coefficient of the sample acupuncture case. The data selection compensation coefficient directly reflects the deviation of the sample acupuncture case from the average similarity level. If the data selection compensation coefficient is greater than 1, it means that the similarity of the sample acupuncture case is higher than the average level, and the sample acupuncture case has stronger representativeness and reference value in reflecting the individual characteristics of the target patient. In subsequent model training, it should be given higher weight.

[0035] Second, the data selection compensation coefficient is multiplied by a preset amplification base, i.e. 10, and the product is rounded down to get the final data selection times for the sample acupuncture case. The final data selection times is an integer value not less than 1, and its fundamental purpose is to map the continuous data selection compensation coefficient to discrete repeated sampling times that can be directly used to construct the training set, so as to realize the differential weighting of different reference value cases at the data level. The setting of the preset amplification base 10 aims to ensure that even the cases with slightly higher similarity than the average level can also obtain more than 1 selection times through moderate numerical amplification, so as to effectively narrow the weight gradient of different similarity cases in the training set, and enhance the discrimination and representativeness of the training data distribution.

[0036] For all sample acupuncture cases, repeat this calculation process to get a data selection times set corresponding to each sample acupuncture case. The higher the sample feature similarity of the sample acupuncture case, the greater the data selection compensation coefficient, and the more data selection times calculated, which ensures that the most similar case experience to the target patient will play a more dominant role in the model training process.

[0037] Further, after obtaining the data selection frequency set, a sample training set is constructed. Specifically, according to the frequency specified by the data selection frequency set, repeated selection with replacement is performed in the sample acupuncture case set. For example, if the data selection frequency of a certain sample acupuncture case is 3, the sample acupuncture case will appear 3 times in the sample training set. In this way, historical acupuncture cases with high similarity to the target patient characteristics and higher reference value will have a higher frequency and weight in the final training set. The data distribution of the sample training set thus constructed is consciously adjusted to strengthen the influence of the treatment mode represented by high similarity cases on model training.

[0038] Further, the sample training set constructed above is used to train the acupuncture parameter evaluation plug-in. The specific training process takes the combination of the historical patient feature vector matrix and the historical acupuncture treatment parameters as the model input. The historical patient feature vector matrix comprehensively describes the individual physiological and pathological conditions of the patient, and the historical acupuncture treatment parameters accurately describe the adopted acupoint distribution and needle operation parameter group. The historical treatment quality coefficient is used as the supervision signal of the model, that is, the target output pursued by the training. A deep learning model architecture, such as a multi-layer perceptron or a hybrid structure combining a convolutional neural network and a fully connected layer, is used to iteratively optimize and train the model parameters based on the sample training set.

[0039] The training process will continue until the model converges. The convergence condition can be set according to actual needs. For example, when the prediction error of the model on the training set, such as mean square error or mean absolute error, no longer significantly decreases in consecutive training cycles and is lower than a preset convergence threshold, such as 0.01, the model is determined to have converged. At this time, the trained deep learning model is the acupuncture parameter evaluation plug-in.

[0040] Specifically, the acupuncture parameter evaluation plug-in can predict the corresponding quantitative treatment quality coefficient for any given patient feature data that is completely consistent with the structure of the target patient feature vector matrix and any set of acupuncture treatment parameters within the defined range. This treatment quality coefficient represents the expected comprehensive efficacy that can be achieved by using a specific treatment plan in the context of a specific patient characteristic, thereby enabling rapid, objective, and quantitative evaluation of the potential efficacy of different candidate treatment plans, providing a crucial data-based evaluation basis and guidance signal for subsequent iterative optimization search in the acupuncture parameter optimization space.

[0041] For example, due to the highly nonlinear and complex mapping relationship between the patient multi-dimensional feature vector matrix, the complex acupuncture treatment parameters, and the final treatment quality coefficient, and the significant advantages of deep learning models in integrating multi-source heterogeneous data and mining deep nonlinear relationships, a deep neural network model can be selected to construct the acupuncture parameter evaluation plug-in.

[0042] Specifically, the acupuncture parameter evaluation plug-in mainly consists of a feature fusion input layer, a multi-level feature abstraction layer, and a treatment quality prediction output layer. The feature fusion input layer receives the historical patient feature vector matrix after standardization processing and the structured coded historical acupuncture treatment parameters, and splices and fuses the two to form a combined feature vector. The multi-level feature abstraction layer adopts a multi-layer fully connected neural network structure, the number of hidden layer neurons is adaptively configured according to the total dimension of the combined feature vector, each layer of neural network adopts a ReLU activation function to introduce a nonlinear modeling capability, and a Dropout layer is embedded between network layers with a dropout rate set between 0.2 and 0.4 to effectively suppress model overfitting and improve its generalization performance. The output layer adopts a linear activation function to map the deeply abstracted features to a continuous numerical value as the predicted treatment quality coefficient.

[0043] During the training process, the key hyperparameters include a learning rate set to 0.001, a training round number set to 200, and a batch size set to 32. The learning rate is set based on the balance between training stability and convergence speed, the training round number is set to ensure that the model fully learns the complex mapping relationship of historical cases, and the batch size is selected to balance the training efficiency and the stability of gradient update. Specifically, the training method of supervised learning is adopted, the historical patient feature vector matrix and the historical acupuncture treatment parameters in the sample training set are taken as the input sample set, and the corresponding historical treatment quality coefficient is obtained to form the sample label set. The input sample set and the corresponding label sample set are divided into a training subset, a validation subset, and a test subset in a ratio of 8:1:1 to ensure the effectiveness of model training and evaluation.

[0044] Further, the sample combined feature vector in the training subset is taken as the input, and the corresponding historical treatment quality coefficient is taken as the supervision signal. Through the back propagation algorithm, the network weight parameters are iteratively optimized with the Adam optimizer. The mean square error loss function is used to measure the deviation between the predicted treatment quality coefficient and the historical treatment quality coefficient, and the training process is monitored through the validation subset. When the validation subset loss function value no longer significantly decreases in continuous multiple training periods, and the mean square error is lower than a preset convergence standard threshold such as 0.01, it is determined that the model converges, the training is terminated, and the converged acupuncture parameter evaluation plug-in is finally obtained.

[0045] S40: Based on the sample acupuncture case set, an acupuncture parameter optimization space is constructed, the acupuncture parameter evaluation plug-in is used to iteratively optimize the acupuncture parameters in the acupuncture parameter optimization space, and the adaptive acupuncture parameters are output as the adaptive acupuncture treatment scheme for the target patient.

[0046] Specifically, based on the sample acupuncture case set, the acupuncture parameter optimization space is constructed, including: screening historical acupuncture treatment parameters in the sample acupuncture case set to obtain a historical acupuncture treatment parameter set; taking the union set of the acupoints in the historical acupuncture treatment parameter set as the acupoint threshold for acupoint selection; taking the upper and lower limits of the acupuncture parameters in the historical acupuncture treatment parameter set as the acupuncture parameter threshold; constructing an acupuncture parameter optimization space based on the acupoint threshold and the acupuncture parameter threshold.

[0047] Firstly, all historical acupuncture treatment parameters used by all historical cases in the sample acupuncture case set are extracted to form a historical acupuncture treatment parameter set. This historical acupuncture treatment parameter set covers all specific operation parameters used in the historical treatment experience highly related to the target patient.

[0048] Secondly, based on the historical acupuncture treatment parameter set, the boundaries of the acupuncture parameter optimization space are defined. This definition process includes two key aspects. On the one hand, the acupoint selection range is defined. The specific method is to analyze and merge the acupoint distribution recorded by all historical cases in the historical acupuncture treatment parameter set, and all acupoints contained in the acupoint distribution are combined into a union set. This acupoint union set is set as the acupoint threshold for acupoint selection in the optimization process. This threshold limits the range of acupoints allowed to be selected in the subsequent optimization search process, ensuring that the core acupoint selection comes from the acupoint set that has been verified by similar historical cases, thereby ensuring the clinical relevance of the optimized scheme in acupoint compatibility.

[0049] On the other hand, the numerical range of acupuncture operation parameters is defined. The specific acupuncture operation parameters include lifting and inserting amplitude, lifting and inserting frequency, twisting angle, twisting speed, needle insertion time and needle retention time. Through comprehensive statistical analysis of the historical acupuncture treatment parameter set, the minimum value and maximum value of each acupuncture parameter appearing in all historical case records are determined. The closed interval formed by the minimum value as the lower limit and the maximum value as the upper limit is defined as the acupuncture parameter threshold of this parameter. The obtained acupuncture parameter thresholds together constitute the legal numerical boundaries of each operation parameter in the optimization space.

[0050] Finally, the acupoint threshold and each acupuncture parameter threshold are combined to define the acupuncture parameter optimization space. The acupuncture parameter optimization space is a discrete candidate acupoint set in the acupoint selection dimension, and a continuous numerical interval in each acupuncture operation parameter dimension. The purpose of constructing this acupuncture parameter optimization space is to constrain the subsequent optimization search process within the parameter range derived from the similar historical successful treatment experience of the target patient, thereby greatly reducing the search range, improving the optimization efficiency, and ensuring that the generated candidate treatment scheme has a solid clinical practice foundation and safety guarantee.

[0051] Further, the acupuncture parameter evaluation plug-in is used to iteratively optimize the acupuncture parameters in the acupuncture parameter optimization space, and output adaptive acupuncture parameters, including: randomly selecting a plurality of initial acupuncture parameters in the acupuncture parameter optimization space; using the acupuncture parameter evaluation plug-in to evaluate a plurality of treatment quality coefficients of the plurality of initial acupuncture parameters; based on the acupuncture parameter optimization space, iteratively optimizing the acupuncture parameters according to the plurality of treatment quality coefficients and the plurality of initial acupuncture parameters, and output adaptive acupuncture parameters.

[0052] First, in the acupuncture parameter optimization space, a plurality of initial acupuncture parameter combinations are randomly selected according to the defined acupoint threshold and acupuncture parameter threshold. The plurality of initial acupuncture parameter combinations serve as the starting point of the optimization search process, aiming to cover a certain diversity in part of the optimization space.

[0053] Secondly, each initial acupuncture parameter combination is combined with the feature vector matrix of the target patient and input into the trained acupuncture parameter evaluation plug-in. The acupuncture parameter evaluation plug-in performs forward calculation on each input and outputs a corresponding predicted treatment quality coefficient, thereby obtaining a plurality of treatment quality coefficients corresponding to the plurality of initial acupuncture parameters, and quantitatively evaluating the potential efficacy of the initial solution set for the first time.

[0054] Further, based on the boundary constraints set in the acupuncture parameter optimization space, the plurality of treatment quality coefficients obtained are used as evaluation criteria, and the plurality of initial acupuncture parameters are used as the initial population to start the iterative optimization algorithm. The iterative optimization algorithm intelligently adjusts, updates, eliminates and generates new acupuncture parameter combinations in the acupuncture parameter optimization space according to the level of treatment quality coefficients. In each iteration, the new acupuncture parameter combination needs to be predicted for efficacy by the acupuncture parameter evaluation plug-in, and the prediction result is used as a feedback signal to guide the next iteration update. This cycle continues, so that the acupuncture parameter combination continuously evolves in the direction of higher predicted treatment quality coefficients in the acupuncture parameter optimization space. When the optimization process meets the preset termination condition, the algorithm terminates, and the acupuncture parameter combination with the highest predicted treatment quality coefficient in the current iteration is output as the final adaptive acupuncture parameter. The adaptive acupuncture parameter is the best treatment scheme core operation parameter customized for the target patient and optimized by data driving.

[0055] Specifically, based on the acupuncture parameter optimization space, iteratively optimizing the acupuncture parameters according to the plurality of treatment quality coefficients and the plurality of initial acupuncture parameters, and output adaptive acupuncture parameters, including: The initial acupuncture parameters are set as initial solutions, and the initial acupuncture parameters are arranged in descending order of treatment quality coefficients to generate an initial solution sequence; The first solution of the initial solution sequence is selected as an optimal solution, and the remaining initial solutions in the initial solution sequence except the optimal solution are set as inferior solutions to obtain a plurality of inferior solutions; The optimal solution is taken as an optimization direction, and the plurality of inferior solutions are adjusted according to a preset optimization step to obtain a plurality of updated inferior solutions, The optimal solution and the plurality of updated inferior solutions are reordered in descending order of treatment quality coefficients to generate an initial updated solution sequence; The last Q solutions in the initial updated solution sequence are eliminated, Q initial solutions not appearing in the acupuncture parameter optimization space during the optimization process are supplemented, and an updated solution sequence is reordered, wherein Q is an integer, and Q decreases as the number of optimization iterations increases; Based on the updated solution sequence, acupuncture parameter iterative optimization is continued until a preset optimization convergence number is reached, and the optimal solution of the current updated solution sequence is output as an adaptive acupuncture parameter.

[0056] First, a plurality of randomly selected initial acupuncture parameters are set as initial solutions, and the plurality of initial acupuncture parameters are sorted in descending order of corresponding treatment quality coefficients calculated by the acupuncture parameter evaluation plug-in to generate an ordered initial solution sequence.

[0057] Secondly, from the initial solution sequence, the first ranked solution is set as the current optimal solution, that is, the optimal solution. At the same time, all the remaining solutions in the initial solution sequence except the optimal solution are classified as inferior solutions, thereby obtaining an optimal solution and a plurality of inferior solutions.

[0058] Further, the acupuncture parameter vector represented by the current optimal solution is taken as the dominant direction of search and evolution, and each inferior solution is systematically adjusted according to a preset optimization step. The adjustment process follows a clear strategy: for discrete parameters such as acupoint selection, a probability-based vector approximation operation can be used; for continuous numerical parameters such as lifting amplitude and frequency, linear interpolation or adding bounded random disturbance subject to a specific distribution to the optimal solution corresponding parameter value is used. In this way, the acupuncture parameter values of the inferior solutions are within the constraint range of the optimization space, and are targeted to move closer to the optimal solution or explore in its neighborhood, thereby generating a batch of updated inferior solutions.

[0059] The preset optimization step length is a key scalar or vector parameter for controlling the adjustment range of the acupuncture parameters. The specific value is dynamically set according to the scale of the optimization space, the data type of the parameters, and the iteration optimization stage. For example, in the early stage of iteration, in order to quickly explore, a relatively large initial step length can be set for continuous numerical parameters, such as taking 10% of the allowed variation range of the acupuncture parameters as the step length reference; and in the later stage of iteration, in order to fine-tune, the step length can be exponentially decayed to 1% or less of the range. The setting of the preset optimization step length aims to balance the global exploration and local development capabilities of the algorithm.

[0060] Further, the current optimal solution is combined with the plurality of updated inferior solutions, and the treatment quality coefficient corresponding to all solutions is calculated again by using the acupuncture parameter evaluation plug-in. All solutions are reordered according to the new treatment quality coefficient, thereby generating a new, ordered initial update solution sequence.

[0061] Secondly, the Q solutions with the lowest treatment quality coefficient are eliminated from the end of the initial update solution sequence. At the same time, in order to maintain the diversity of the population and avoid falling into local optimization, Q new acupuncture parameter combinations that have never appeared in this iteration and previous iterations are randomly generated in the acupuncture parameter optimization space as supplementary solutions. The remaining old solutions are combined with the Q new supplementary solutions, and are reordered to form an update solution sequence for the next round of iteration. Q is a positive integer, and its value is dynamically reduced as the number of optimization iterations increases, ensuring that in the early stage of search, a larger elimination and supplement ratio is needed to extensively explore the optimization space; as the iteration proceeds, the overall quality of the solution group improves and gradually converges, at which time the disturbance should be reduced and focused on fine development of the high-quality solution area.

[0062] Finally, the newly generated update solution sequence is taken as the input of the next round of iteration, and the complete process from determining the optimal solution to generating the update solution sequence is repeatedly executed. This cycle continues until the preset optimization convergence condition is met. When the number of iterations reaches the preset optimization convergence number, the algorithm automatically terminates, and the optimal solution ranked first in the current round of update solution sequence is output as the final determined adaptive acupuncture parameter.

[0063] The preset optimization convergence number is a key parameter for controlling the total amount of calculation in the optimization process. The specific value is set comprehensively according to the complexity of the problem, the scale of the optimization space, and the actual constraints of the computing resources. For example, for a moderately complex treatment parameter optimization problem, the convergence number can be set to 200 times based on preliminary experiments; for a higher-dimensional parameter space or more complex correlation, the convergence number can be appropriately increased to 500 times or more to ensure that the algorithm has sufficient opportunity to approach the global optimal region. The adaptive acupuncture parameter is the optimal treatment scheme core operation parameter obtained through the above data-driven iterative search process.

[0064] In summary, the embodiments of this application have at least the following technical effects: Compared to existing technologies, this application first constructs a multi-dimensional patient feature vector matrix and performs precise historical case retrieval based on a similarity threshold. This effectively filters out treatment experience data highly relevant to the target patient, laying a highly personalized data foundation for subsequent analysis and overcoming the shortcomings of traditional methods, such as coarse data utilization and weak targeting. Secondly, a training set is dynamically constructed based on case similarity, and a dedicated acupuncture parameter evaluation plugin is generated using this set. This plugin can accurately quantify the expected efficacy of different acupuncture treatment parameters under specific patient characteristics, realizing a shift from experience-based judgment to data prediction and improving the objectivity and accuracy of treatment plan evaluation.

[0065] Furthermore, this application innovatively constructs an acupuncture parameter optimization space bounded by treatment parameters from similar cases. Within this space, an evaluation plugin guides an intelligent optimization algorithm for iterative searching. This not only significantly narrows the search range and improves optimization efficiency but also ensures that the recommended solutions always remain within the safe and effective range validated by clinical experience. Finally, the three key steps of case retrieval, model construction, and parameter optimization are organically integrated into a complete, closed-loop data-driven process. This achieves fully automated recommendation from patient feature input to personalized treatment plan generation, improving the precision and personalization of acupuncture treatment while enhancing the reliability, interpretability, and clinical operability of treatment plans. This provides a practical and feasible technical path for the modernization and intelligent development of acupuncture treatment.

[0066] Example 2, as Figure 2 As shown, based on the same inventive concept as the method for recommending personalized acupuncture treatment plans based on data analysis provided in Embodiment 1, this embodiment of the invention also provides a personalized acupuncture treatment plan recommendation system based on data analysis, including: Feature vector construction module 11 is used to construct a feature vector matrix based on the patient feature data of the target patient; The similar case retrieval module 12 is used to perform similarity traversal in the pre-constructed acupuncture case database with the feature vector matrix as a constraint, select historical acupuncture cases with similarity greater than a preset similarity threshold as sample acupuncture cases, and obtain a sample acupuncture case set and a sample feature similarity set. The evaluation plugin construction module 13 is used to set a data selection strategy based on the sample feature similarity set, select data from the sample acupuncture case set to obtain a sample training set, and use the sample training set to construct an acupuncture parameter evaluation plugin. The treatment scheme generation module 14 is configured to construct an acupuncture parameter optimization space based on the sample acupuncture case set, perform iterative optimization of acupuncture parameters in the acupuncture parameter optimization space by using the acupuncture parameter evaluation plug-in, and output adaptive acupuncture parameters as an adaptive acupuncture treatment scheme for the target patient.

[0067] The feature vector construction module 11 is configured to: construct a feature vector matrix according to patient feature data of the target patient.

[0068] Specifically, the patient feature data includes basic physical sign data, basic disease condition data, imaging data, biochemical indicator data, electrophysiological data, and acupuncture region body fat distribution.

[0069] The similar case retrieval module 12 is configured to: perform similar traversal in a pre-constructed acupuncture case database with the feature vector matrix as a constraint, select historical acupuncture cases with a similarity greater than a preset similarity threshold as sample acupuncture cases, and obtain a sample acupuncture case set and a sample feature similarity set.

[0070] Specifically, the acupuncture case database includes a plurality of historical acupuncture cases, wherein the historical acupuncture cases include historical patient feature vector matrices, historical acupuncture treatment parameters, and historical treatment quality coefficients, the acupuncture treatment parameters include acupoint distribution and acupuncture parameter groups, and the acupuncture parameter groups include lifting and inserting amplitude, lifting and inserting frequency, rotation angle, rotation speed, acupuncture duration, and needle retention duration.

[0071] In addition, the historical treatment quality coefficient is determined based on a specific disease functional scale score improvement rate of a historical patient, a historical digital pain score change rate, and a number of historical treatments required to reach a minimum clinically important difference.

[0072] The evaluation plug-in construction module 13 is configured to: set a data selection strategy based on the sample feature similarity set, perform data selection in the sample acupuncture case set to obtain a sample training set, including: set a ratio of a sample feature similarity of each sample acupuncture case to a mean value of the sample feature similarity set as a data selection compensation coefficient, take an integer of a product of the data selection compensation coefficient and 10 as a data selection number, calculate a data selection number set based on the sample feature similarity set; perform repeated data selection in the sample acupuncture case set according to the data selection number set to construct the sample training set.

[0073] Specifically, a deep learning model is trained by taking the historical patient feature vector matrix and the historical acupuncture treatment parameters as input and taking the historical treatment quality coefficients as supervision until convergence, and an acupuncture parameter evaluation plug-in is generated.

[0074] The treatment scheme generation module 14 is specifically configured to: Based on the sample acupuncture case set, an acupuncture parameter optimization space is constructed, including: In the sample acupuncture case set, historical acupuncture treatment parameters are screened to obtain a historical acupuncture treatment parameter set; The union of the multiple acupoint distributions in the historical acupuncture treatment parameter set is taken as an acupoint threshold for taking acupoints; The upper and lower limits of the acupuncture parameters in the historical acupuncture treatment parameter set are selected as acupuncture parameter thresholds; Based on the acupoint threshold for taking acupoints and the acupuncture parameter threshold, an acupuncture parameter optimization space is constructed.

[0075] Further, the acupuncture parameter evaluation plug-in is used to perform iterative optimization of acupuncture parameters in the acupuncture parameter optimization space, and to output adaptive acupuncture parameters, including: Randomly select several initial acupuncture parameters in the acupuncture parameter optimization space; Use the acupuncture parameter evaluation plug-in to evaluate to obtain several treatment quality coefficients of the several initial acupuncture parameters; Based on the acupuncture parameter optimization space, the several treatment quality coefficients and the several initial acupuncture parameters are used to perform iterative optimization of acupuncture parameters, and adaptive acupuncture parameters are output.

[0076] Specifically, based on the acupuncture parameter optimization space, the several treatment quality coefficients and the several initial acupuncture parameters are used to perform iterative optimization of acupuncture parameters, and adaptive acupuncture parameters are output, including: The initial acupuncture parameters are set as initial solutions, the several initial acupuncture parameters are arranged in descending order of treatment quality coefficients, and an initial solution sequence is generated; The first solution of the initial solution sequence is selected as a superior solution, and the remaining initial solutions in the initial solution sequence except for the superior solution are set as inferior solutions, and a plurality of inferior solutions are obtained; The superior solution is taken as an optimization direction, and the plurality of inferior solutions are adjusted according to a preset optimization step size to obtain a plurality of updated inferior solutions, The superior solution and the plurality of updated inferior solutions are reordered in descending order of treatment quality coefficients to generate an initial updated solution sequence; Eliminate the last Q solutions in the initial update solution sequence, supplement the initial solutions that do not appear in the Q optimization process in the acupuncture parameter optimization space, and reorder to generate an updated solution sequence, wherein Q is an integer, and Q decreases with the increase of the number of optimization iterations; Based on the updated solution sequence, continue to perform acupuncture parameter iterative optimization until a preset optimization convergence number is reached, and output the optimal solution of the current updated solution sequence as the adaptive acupuncture parameter.

[0077] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And the above describes a specific embodiment of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0078] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0079] The specification and drawings of the present application are only exemplary descriptions of the present application, and are considered to cover any and all modifications, changes, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the present application and its equivalent technology, the present application intends to include these modifications and changes.

Claims

1. A method for recommending a personalized acupuncture treatment plan based on data analysis, characterized in that, The method comprises: constructing a feature vector matrix according to patient feature data of a target patient; performing similar traversal in a pre-constructed acupuncture case database with the feature vector matrix as a constraint, selecting historical acupuncture cases with a similarity greater than a preset similarity threshold as sample acupuncture cases, obtaining a sample acupuncture case set and a sample feature similarity set; setting a data selection strategy based on the sample feature similarity set, performing data selection in the sample acupuncture case set to obtain a sample training set, and constructing an acupuncture parameter evaluation plug-in using the sample training set; constructing an acupuncture parameter optimization space based on the sample acupuncture case set, performing iterative optimization of acupuncture parameters in the acupuncture parameter optimization space using the acupuncture parameter evaluation plug-in, and outputting adaptive acupuncture parameters as an adaptive acupuncture treatment plan for the target patient.

2. The data analysis based personalized acupuncture treatment protocol recommendation method of claim 1, wherein, The patient feature data includes basic physical sign data, basic disease condition data, imaging data, biochemical indicator data, electrophysiological data, and acupuncture region body fat distribution. 3.The data analysis based personalized acupuncture treatment protocol recommendation method of claim 1, wherein, The acupuncture case database includes a plurality of historical acupuncture cases, wherein the historical acupuncture cases include historical patient feature vector matrices, historical acupuncture treatment parameters, and historical treatment quality coefficients, wherein the acupuncture treatment parameters include acupoint distribution and acupuncture parameter groups, and the acupuncture parameter groups include lifting and inserting amplitude, lifting and inserting frequency, rotation angle, rotation speed, acupuncture duration, and needle retention duration. 4.The data analysis based personalized acupuncture treatment protocol recommendation method of claim 3, wherein, The historical treatment quality coefficient is determined based on a specific disease functional scale score improvement rate of a historical patient, a historical digital pain score change rate, and a number of historical treatments required to achieve a minimum clinically important difference. 5.The data analysis based personalized acupuncture treatment protocol recommendation method of claim 1, wherein, Setting a data selection strategy based on the sample feature similarity set, performing data selection in the sample acupuncture case set to obtain a sample training set, comprises: setting a ratio of the sample feature similarity of each sample acupuncture case to a mean value of the sample feature similarity set as a data selection compensation coefficient, taking an integer of a product of the data selection compensation coefficient and 10 as a data selection number, and calculating a data selection number set based on the sample feature similarity set; performing repeated data selection in the sample acupuncture case set according to the data selection number set to construct a sample training set. 6.The data analysis based personalized acupuncture treatment protocol recommendation method of claim 3, wherein, Using the historical patient feature vector matrix and historical acupuncture treatment parameters as input, using the historical treatment quality coefficient as supervision, training a deep learning model using the sample training set until convergence, and generating an acupuncture parameter evaluation plug-in.

7. The data analysis based personalized acupuncture treatment protocol recommendation method of claim 3, wherein, Constructing an acupuncture parameter optimization space based on the sample acupuncture case set comprises: filtering historical acupuncture treatment parameters in the sample acupuncture case set to obtain a historical acupuncture treatment parameter set; taking a union set of acupoints in multiple acupoint distributions in the historical acupuncture treatment parameter set as an acupoint threshold for acupoint selection; selecting upper and lower limits of acupuncture parameters in the historical acupuncture treatment parameter set as acupuncture parameter thresholds; constructing an acupuncture parameter optimization space based on the acupoint threshold for acupoint selection and the acupuncture parameter thresholds. 8.The data analysis based personalized acupuncture treatment protocol recommendation method of claim 1, wherein, Performing iterative optimization of acupuncture parameters in the acupuncture parameter optimization space using the acupuncture parameter evaluation plug-in and outputting adaptive acupuncture parameters comprises: randomly selecting a plurality of initial acupuncture parameters in the acupuncture parameter optimization space; The acupuncture parameter evaluation plug-in is used to evaluate a plurality of initial acupuncture parameters to obtain a plurality of treatment quality coefficients; Based on the acupuncture parameter optimization space, the acupuncture parameter iterative optimization is performed according to the plurality of treatment quality coefficients and the plurality of initial acupuncture parameters, and the adaptive acupuncture parameters are output. 9.The data analysis based personalized acupuncture treatment protocol recommendation method of claim 8, wherein, Based on the acupuncture parameter optimization space, the acupuncture parameter iterative optimization is performed according to the plurality of treatment quality coefficients and the plurality of initial acupuncture parameters, and the adaptive acupuncture parameters are output, including: The initial acupuncture parameters are set as initial solutions, the plurality of initial acupuncture parameters are arranged in descending order of treatment quality coefficients, and an initial solution sequence is generated; The first solution of the initial solution sequence is set as a superior solution, and the remaining initial solutions in the initial solution sequence except the superior solution are set as inferior solutions, to obtain a plurality of inferior solutions; The superior solution is taken as an optimization direction, and the plurality of inferior solutions are adjusted according to a preset optimization step, to obtain a plurality of updated inferior solutions, The superior solution and the plurality of updated inferior solutions are reordered in descending order of treatment quality coefficients, to generate an initial updated solution sequence; The last Q solutions in the initial updated solution sequence are eliminated, Q initial solutions not appearing in the optimization process are supplemented in the acupuncture parameter optimization space, and an updated solution sequence is reordered, wherein Q is an integer, and Q decreases with an increase in the number of optimization iterations; Based on the updated solution sequence, the acupuncture parameter iterative optimization is continuously performed until a preset optimization convergence number is reached, and the superior solution of the current updated solution sequence is output as the adaptive acupuncture parameters.

10. A personalized acupuncture treatment protocol recommendation system based on data analysis, characterized by, The method for recommending an individualized acupuncture treatment plan based on data analysis according to any one of claims 1-9, comprising: a feature vector construction module configured to construct a feature vector matrix based on patient feature data of a target patient; a similar case retrieval module configured to perform similar traversal in a pre-constructed acupuncture case database with the feature vector matrix as a constraint, select historical acupuncture cases with a similarity greater than a preset similarity threshold as sample acupuncture cases, and obtain a sample acupuncture case set and a sample feature similarity set; an evaluation plug-in construction module configured to set a data selection strategy based on the sample feature similarity set, perform data selection in the sample acupuncture case set to obtain a sample training set, and construct an acupuncture parameter evaluation plug-in using the sample training set; a treatment plan generation module configured to construct an acupuncture parameter optimization space based on the sample acupuncture case set, perform acupuncture parameter iterative optimization in the acupuncture parameter optimization space using the acupuncture parameter evaluation plug-in, and output adaptive acupuncture parameters as an adaptive acupuncture treatment plan for the target patient.

Images