Knee osteoarthritis acupuncture curative effect prediction system based on deep learning

Abstract

The invention relates to the technical field of medical data processing, in particular to a knee osteoarthritis acupuncture curative effect prediction system based on deep learning. The system obtains acupuncture treatment parameters and acupuncture effect parameters of historical knee osteoarthritis patients; the historical knee osteoarthritis patients are divided into patient categories based on the acupuncture effect parameters, and the contribution degree is obtained according to the distribution of each acupuncture treatment parameter in each patient category and the distribution difference between the acupuncture treatment parameter and the acupuncture treatment parameter in other patient categories; acquiring a local curative effect feedback value according to the contribution degree and the acupuncture effect parameter, and further acquiring an evaluation score; and training the acupuncture curative effect prediction model through the acupuncture treatment parameters, the evaluation scores and the acupuncture effect parameters to obtain a trained acupuncture curative effect prediction model. According to the method, the evaluation score is dynamically obtained, the participation degree of each acupuncture treatment parameter of each historical knee osteoarthritis patient is accurately determined, and the accuracy of acupuncture treatment effect prediction model training is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing, and in particular to a system for predicting the efficacy of acupuncture and moxibustion for knee osteoarthritis based on deep learning. Background Art

[0002] Knee osteoarthritis is a chronic degenerative disease characterized by articular cartilage degeneration, bone hyperplasia, and synovial inflammation, primarily manifesting as knee pain, stiffness, and limited mobility. Acupuncture, a core method of traditional Chinese medicine, regulates local Qi and blood circulation and alleviates inflammatory responses by stimulating specific acupoints. It has been shown in numerous clinical studies to effectively improve pain levels and joint function in patients with knee osteoarthritis. For example, electroacupuncture stimulation of six acupoints on the knee (such as Dubi and Yanglingquan) can significantly reduce intra-articular pressure and promote cartilage repair. However, varying acupuncture treatment parameters during treatment can have varying effects on the therapeutic effect in patients with knee osteoarthritis. Therefore, in order to effectively treat patients with knee osteoarthritis, it is necessary to accurately predict the therapeutic effect using acupuncture treatment parameters.

[0003] In existing methods, an acupuncture efficacy prediction model is trained using a large number of historical acupuncture treatment parameters from patients with knee osteoarthritis, along with fixed weights corresponding to these parameters and acupuncture efficacy parameters that can characterize the treatment effects of these patients. This yields a trained acupuncture efficacy prediction model, which is then used to predict the acupuncture treatment effects of patients with knee osteoarthritis based on their various acupuncture treatment parameters. However, in reality, the relationship between various acupuncture treatment parameters and efficacy is not fixed. For example, electroacupuncture has a significant analgesic effect at a frequency of 2-4Hz, but exceeding 4Hz may result in a decreased effect due to neural adaptability. Therefore, the training of the acupuncture efficacy prediction model in existing methods is biased, which in turn affects the accurate prediction of acupuncture treatment effects in patients with knee osteoarthritis. Summary of the Invention

[0004] In order to solve the technical problem of deviation in the training of acupuncture efficacy prediction models in existing methods, the purpose of the present invention is to provide an acupuncture efficacy prediction system for knee osteoarthritis based on deep learning. The technical solutions adopted are as follows: An embodiment of the present invention provides a deep learning-based acupuncture efficacy prediction system for knee osteoarthritis, the system comprising the following steps: A data acquisition module is used to obtain various acupuncture treatment parameters and acupuncture effect parameters of each historical knee osteoarthritis patient; a contribution degree acquisition module, for dividing historical knee osteoarthritis patients into patient categories based on acupuncture effect parameters, and obtaining the contribution degree of each acupuncture treatment parameter in each patient category based on the distribution of each acupuncture treatment parameter in each patient category and the difference in distribution with other patient categories; An evaluation score acquisition module is used to obtain the efficacy contribution curve of each acupuncture treatment parameter based on the contribution degree and the acupuncture effect parameter in each patient category; obtain the local efficacy feedback value based on the changes in the efficacy contribution curve; and obtain the evaluation score of each acupuncture treatment parameter of each historical knee osteoarthritis patient based on each acupuncture treatment parameter and the corresponding local efficacy feedback value; The acupuncture efficacy prediction model acquisition module is used to train the acupuncture efficacy prediction model through each acupuncture treatment parameter, evaluation score and acupuncture effect parameter of each historical knee osteoarthritis patient to obtain a trained acupuncture efficacy prediction model.

[0005] Furthermore, the method for obtaining the contribution degree is: According to the distribution of each acupuncture treatment parameter in each patient category, the compactness of each acupuncture treatment parameter in each patient category is obtained; According to the compactness difference and size difference of each acupuncture treatment parameter in each patient category and other patient categories, the special degree of each acupuncture treatment parameter in each patient category is obtained; The product of the compactness and specificity of each acupuncture treatment parameter in each patient category is taken as the contribution of each acupuncture treatment parameter in each patient category.

[0006] Furthermore, the method for obtaining the compactness is: For any acupuncture treatment parameter and any patient category, the DBSCAN density clustering algorithm is used to cluster the historical knee osteoarthritis patients within the patient category based on the magnitude of the acupuncture treatment parameter for each patient within the patient category, and the local category corresponding to the acupuncture treatment parameter in the patient category is obtained; The number of patients with historical knee OA in each local category was obtained and used as the number of local patients; The ratio of the maximum number of local patients to the total number of all historical knee osteoarthritis patients in this patient category is used as the reference stability of this acupuncture treatment parameter in this patient category; The product of the reference stability level and the inverse of the number of local categories is used as the compactness of the acupuncture treatment parameter in the patient category.

[0007] Furthermore, the method for obtaining the degree of specialness is: For any acupuncture treatment parameter and any patient category, obtain the mean value of the compactness of the acupuncture treatment parameter in all patient categories except the patient category as the first eigenvalue; The difference between the compactness of the acupuncture treatment parameter in the patient category and the first characteristic value is used as the first discrimination degree of the acupuncture treatment parameter in the patient category; Obtain the mean value of the acupuncture treatment parameter for all patients with a history of knee osteoarthritis in the patient category as the overall reference value of the acupuncture treatment parameter in the patient category; Obtaining the mean of the overall reference values of the acupuncture treatment parameter in all patient categories except the patient category as the second eigenvalue; The difference between the overall reference value of the acupuncture treatment parameter in the patient category and the second characteristic value is used as the second discrimination degree of the acupuncture treatment parameter in the patient category; The result of normalizing the sum of the first degree of distinction and the second degree of distinction is used as the special degree of the acupuncture treatment parameter in the patient category.

[0008] Furthermore, the method for obtaining the efficacy contribution curve is: The mean of the acupuncture effect parameters of all patients with historical knee OA in each patient category was used as the target effect parameter for each patient category; For any acupuncture treatment parameter, the contribution of the acupuncture treatment parameter in each patient category is taken as the y value in the two-dimensional coordinate, and the target effect parameter of each patient category is taken as the x value in the two-dimensional coordinate to obtain the reference coordinate of each patient category corresponding to the acupuncture treatment parameter; The points corresponding to the reference coordinates in the two-dimensional coordinate system are fitted into a curve as a contribution curve of the therapeutic effect of the acupuncture treatment parameters.

[0009] Furthermore, the method for obtaining the local therapeutic effect feedback value is: For any efficacy contribution curve of acupuncture treatment parameters, the extreme value points on the efficacy contribution curve of the acupuncture treatment parameters are all used as segmentation points, and the efficacy contribution curve of the acupuncture treatment parameters is divided into multiple local curve segments; For any local curve segment, the ratio of the contribution range of the local curve segment to the overall contribution range of the efficacy contribution curve of the acupuncture treatment parameter is obtained as the local feedback degree of the local curve segment; The product of the slope of the local curve segment and the local feedback degree is used as a local efficacy feedback value of the acupuncture treatment parameter.

[0010] Furthermore, the method for obtaining the evaluation score is: For any patient with a history of knee osteoarthritis and any acupuncture treatment parameter, the mean value of the acupuncture treatment parameter of all patients with a history of knee osteoarthritis in the patient category corresponding to the maximum local efficacy feedback value of the acupuncture treatment parameter is used as the target analysis value; The difference between the acupuncture treatment parameter of the patient with historical knee osteoarthritis and the target analysis value is used as the first important analysis value of the acupuncture treatment parameter of the patient with historical knee osteoarthritis; The difference between the local efficacy feedback value corresponding to the patient category of the acupuncture treatment parameter of the historical knee osteoarthritis patient and the maximum local efficacy feedback value of the acupuncture treatment parameter is used as the second important analysis value of the acupuncture treatment parameter of the historical knee osteoarthritis patient; The first important analysis value and the second important analysis value are added together, negatively correlated, and normalized to obtain the result, which is used as the evaluation score of the acupuncture treatment parameter for the patient with a history of knee osteoarthritis.

[0011] Furthermore, the method for obtaining the acupuncture efficacy prediction model is: Each acupuncture treatment parameter and the evaluation score of each acupuncture treatment parameter of each historical knee osteoarthritis patient are used as the input of the acupuncture efficacy prediction model, and the acupuncture effect parameter of each historical knee osteoarthritis patient is used as the output of the acupuncture efficacy prediction model. The acupuncture efficacy prediction model is trained to obtain a trained acupuncture efficacy prediction model.

[0012] Furthermore, the method for obtaining the patient category is: The DBSCAN density clustering algorithm was used to cluster the historical knee osteoarthritis patients according to the acupuncture effect parameters of each historical knee osteoarthritis patient to obtain the patient categories.

[0013] Furthermore, the acupuncture effect parameter is a score given by a doctor based on the treatment effect of each patient with historical knee osteoarthritis after acupuncture treatment.

[0014] The present invention has the following beneficial effects: The present invention first divides historical knee osteoarthritis patients into patient categories based on acupuncture effect parameters, which is conducive to the subsequent efficient analysis of the therapeutic effect of each acupuncture treatment parameter on knee osteoarthritis; in order to analyze the influence of each acupuncture treatment parameter on the acupuncture treatment effect, the contribution degree of each acupuncture treatment parameter in each patient category is obtained according to the distribution of each acupuncture treatment parameter in each patient category and the distribution difference with other patient categories, and accurately reflects the influence degree of each acupuncture treatment parameter range corresponding to each patient category on the acupuncture treatment effect; and then, according to the contribution degree and the acupuncture effect parameter in each patient category, the efficacy contribution curve of each acupuncture treatment parameter is obtained, and the feedback of each acupuncture treatment parameter on the acupuncture treatment effect is accurately reflected, so that the reference degree corresponding to each acupuncture treatment parameter of each historical knee osteoarthritis patient is accurately analyzed in the subsequent analysis; and then, based on the change of the efficacy contribution curve, the local efficacy feedback value is obtained. Accurately reflect the impact of each acupuncture treatment parameter on each stage of the acupuncture treatment effect, and then dynamically obtain the evaluation score of each acupuncture treatment parameter of each historical knee osteoarthritis patient according to each acupuncture treatment parameter and the corresponding local efficacy feedback value of each historical knee osteoarthritis patient, accurately reflecting the effective weight of each acupuncture treatment parameter of each historical knee osteoarthritis patient for obtaining the acupuncture effect parameter; therefore, through each acupuncture treatment parameter, evaluation score and acupuncture effect parameter of each historical knee osteoarthritis patient, the acupuncture efficacy prediction model is trained, and the trained acupuncture efficacy prediction model is accurately obtained, which effectively improves the accuracy of the acupuncture efficacy prediction model in predicting the acupuncture treatment effect of knee osteoarthritis patients, which is conducive to helping doctors further optimize the acupuncture treatment plan for knee osteoarthritis patients, improve the pertinence and effectiveness of acupuncture treatment, so as to achieve better acupuncture treatment effect and improve the treatment satisfaction of knee osteoarthritis patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 This is a structural block diagram of a deep learning-based acupuncture efficacy prediction system for knee osteoarthritis provided by one embodiment of the present invention; Figure 2 A flow chart of a method for obtaining contribution degree provided by one embodiment of the present invention; Figure 3 A schematic diagram of a computer device provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0017] To further illustrate the technical means and effectiveness of the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of a deep learning-based acupuncture efficacy prediction system for knee osteoarthritis proposed by the present invention. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0019] The following describes in detail a specific scheme of a deep learning-based acupuncture efficacy prediction system for knee osteoarthritis provided by the present invention in conjunction with the accompanying drawings.

[0020] Example 1: This paper proposes a deep learning-based prediction system for acupuncture efficacy of knee osteoarthritis. Figure 1 , which shows a structural block diagram of a deep learning-based knee osteoarthritis acupuncture efficacy prediction system provided by an embodiment of the present invention. The system includes: a data acquisition module 10, a contribution degree acquisition module 20, an evaluation score acquisition module 30 and an acupuncture efficacy prediction model acquisition module 40.

[0021] The data acquisition module 10 is used to obtain various acupuncture treatment parameters and acupuncture effect parameters of each historical knee osteoarthritis patient.

[0022] Specifically, information on a preset number of historical knee osteoarthritis patients is obtained through the hospital's database, including various acupuncture treatment parameters and acupuncture effect parameters for each historical knee osteoarthritis patient. The acupuncture effect parameters are scores assigned by doctors based on the treatment effect of each historical knee osteoarthritis patient after acupuncture treatment. Acupuncture treatment parameters include various information, such as electroacupuncture frequency and electroacupuncture depth, as well as characteristic information of the historical knee osteoarthritis patient that affects the acupuncture efficacy, such as age (age affects the acupuncture treatment effect). In this embodiment, the preset number is set to 10,000, and the range of the acupuncture effect parameters is from 0 to 10. The implementer can set the preset number and the range of the acupuncture effect parameters based on actual conditions, and they are not limited here.

[0023] It should be noted that non-numeric acupuncture treatment parameters need to be converted into numerical format through coding to ensure that various acupuncture treatment parameters that appear later are all in numerical format. Among them, the method of converting non-numeric values into numerical format is a well-known technology and will not be described in detail.

[0024] The contribution degree acquisition module 20 is used to divide historical knee osteoarthritis patients into patient categories based on acupuncture effect parameters, and obtain the contribution degree of each acupuncture treatment parameter in each patient category according to the distribution of each acupuncture treatment parameter in each patient category and the difference in distribution with other patient categories.

[0025] Specifically, since the various acupuncture treatment parameters of different historical knee osteoarthritis patients may be different, the acupuncture treatment effects of different historical knee osteoarthritis patients may be different. In order to better analyze the impact of each acupuncture treatment parameter on the acupuncture treatment effect, this embodiment first divides historical knee osteoarthritis patients into patient categories based on the acupuncture effect parameters. It should be noted that the acupuncture treatment effects of historical knee osteoarthritis patients in the same patient category are similar. In this embodiment, the DBSCAN density clustering algorithm is used to cluster the historical knee osteoarthritis patients according to the acupuncture effect parameters of each historical knee osteoarthritis patient to obtain patient categories. Among them, the DBSCAN density clustering algorithm is a well-known technology and will not be described in detail.

[0026] When the distribution of a certain acupuncture treatment parameter for patients with historical knee osteoarthritis in a certain patient category is more concentrated, it indicates that the acupuncture treatment parameter should have a greater impact on the treatment effect of patients with historical knee osteoarthritis in this patient category. When the distribution of acupuncture treatment parameter for patients with historical knee osteoarthritis in this patient category is more dispersed, it can be reflected that the acupuncture treatment parameter has a smaller impact on the treatment effect of patients with historical knee osteoarthritis in this patient category. On the other hand, when the distribution of a certain acupuncture treatment parameter in a certain patient type is significantly different from the distribution in other patient categories, it indicates that the acupuncture treatment parameter is more specific in this patient category, which indirectly indicates that the acupuncture treatment parameter may have a greater impact on the treatment effect of patients with historical knee osteoarthritis in this patient category. Furthermore, this embodiment obtains the contribution degree of each acupuncture treatment parameter in each patient category based on the distribution of each acupuncture treatment parameter in each patient category and the difference in distribution with other patient categories. The greater the contribution degree, the greater the impact of the corresponding acupuncture treatment parameter on the treatment effect of patients with historical knee osteoarthritis in the corresponding patient category.

[0027] Preferably, in one possible implementation of this embodiment, the method for obtaining the contribution degree is as follows: Figure 2 , which shows a flow chart of a method for obtaining contribution degree provided by this embodiment, the method comprising the following steps: Step S201: according to the distribution of each acupuncture treatment parameter in each patient category, obtain the compactness of each acupuncture treatment parameter in each patient category.

[0028] The more similar acupuncture treatment parameters are among patients with a history of knee osteoarthritis within a particular patient category, the more compact the distribution of that acupuncture treatment parameter is within that patient category. This embodiment then determines the degree of compactness of each acupuncture treatment parameter within each patient category based on the distribution of that acupuncture treatment parameter within each patient category. The greater the degree of compactness, the greater the potential impact of that acupuncture treatment parameter on the treatment outcome of patients with a history of knee osteoarthritis within that patient category.

[0029] In one possible implementation of this embodiment, the method for obtaining the degree of compactness is as follows: for any acupuncture treatment parameter and any patient category, cluster the historical knee osteoarthritis patients within the patient category using the DBSCAN density clustering algorithm based on the size of the acupuncture treatment parameter of each historical knee osteoarthritis patient within the patient category, and obtain the local category corresponding to the acupuncture treatment parameter in the patient category; when the number of historical knee osteoarthritis patients in the local category is greater and the number of local categories is smaller, it means that the distribution of the acupuncture treatment parameter in the patient category is more compact. Therefore, this embodiment obtains the number of historical knee osteoarthritis patients in each local category as the number of local patients; then the ratio of the maximum number of local patients to the number of all historical knee osteoarthritis patients in the patient category is used as the reference stability of the acupuncture treatment parameter in the patient category; finally, the product of the reference stability and the inverse of the number of local categories is used as the compactness of the acupuncture treatment parameter in the patient category.

[0030] At this point, the compactness of each acupuncture treatment parameter in each patient category is obtained.

[0031] Step S202: Obtain the specificity of each acupuncture treatment parameter in each patient category based on the compactness difference and size difference of each acupuncture treatment parameter in each patient category and other patient categories.

[0032] When the compactness of a certain acupuncture treatment parameter in a certain patient category is significantly different from the compactness of the same acupuncture treatment parameter in other patient categories, and when the acupuncture treatment parameter of patients with a history of knee osteoarthritis in the patient category is significantly different from the acupuncture treatment parameter of patients with a history of knee osteoarthritis in other patient categories, it indicates that the acupuncture treatment parameter is more special in the patient category. In this embodiment, the specialness of each acupuncture treatment parameter in each patient category is obtained based on the difference in compactness and size between each patient category and other patient categories. The greater the specialness, the greater the treatment effect of the corresponding acupuncture treatment parameter on patients with a history of knee osteoarthritis in the corresponding patient category.

[0033] In one possible implementation of this embodiment, the method for obtaining the degree of distinctiveness is as follows: for any acupuncture treatment parameter and any patient category, the mean value of the compactness of the acupuncture treatment parameter in all patient categories except the patient category is obtained as the first eigenvalue; then the absolute value of the difference between the compactness of the acupuncture treatment parameter in the patient category and the first eigenvalue is used as the first degree of distinctiveness of the acupuncture treatment parameter in the patient category; further, the mean value of the acupuncture treatment parameter of all patients with a history of knee osteoarthritis in the patient category is obtained as the first degree of distinctiveness of the acupuncture treatment parameter in the patient category. ; obtain the mean of the overall reference values of the acupuncture treatment parameter in all patient categories except the patient category as the second eigenvalue; then use the absolute value of the difference between the overall reference value of the acupuncture treatment parameter in the patient category and the second eigenvalue as the second degree of distinction of the acupuncture treatment parameter in the patient category; when both the first degree of distinction and the second degree of distinction are larger, it means that the acupuncture treatment parameter is more special in the patient category, and then add the first degree of distinction and the second degree of distinction and perform normalization as the degree of specialness of the acupuncture treatment parameter in the patient category. In this embodiment, the sum of the first degree of distinction and the second degree of distinction is normalized using the norm normalization function.

[0034] Thus, the specificity of each acupuncture treatment parameter in each patient category is obtained.

[0035] Step S203: The product of the compactness and the specificity of each acupuncture treatment parameter in each patient category is used as the contribution degree of each acupuncture treatment parameter in each patient category.

[0036] It is known that a greater degree of compactness and a greater degree of specificity indicate that a corresponding acupuncture treatment parameter has a greater therapeutic impact on patients with a history of knee osteoarthritis in the corresponding patient category. Therefore, in this embodiment, the product of the degree of compactness and the degree of specificity of each acupuncture treatment parameter in each patient category is used as the contribution of each acupuncture treatment parameter in each patient category.

[0037] At this point, the contribution of each acupuncture treatment parameter in each patient category is obtained.

[0038] The evaluation score acquisition module 30 is used to obtain the efficacy contribution curve of each acupuncture treatment parameter based on the contribution degree and the acupuncture effect parameters in each patient category; obtain the local efficacy feedback value based on the change of the efficacy contribution curve, and obtain the evaluation score of each acupuncture treatment parameter of each historical knee osteoarthritis patient based on each acupuncture treatment parameter and the corresponding local efficacy feedback value.

[0039] Specifically, during acupuncture treatment of patients with knee osteoarthritis, each acupuncture treatment parameter has a different impact on knee osteoarthritis. To accurately analyze the impact of each acupuncture treatment parameter on knee osteoarthritis treatment, each acupuncture treatment parameter was analyzed separately. First, based on the contribution of each acupuncture treatment parameter in each patient category and the acupuncture effect parameter in each patient category, a therapeutic effect contribution curve was obtained for each acupuncture treatment parameter, accurately reflecting the impact of each acupuncture treatment parameter on the acupuncture treatment effect. The therapeutic effect contribution curve was obtained by taking the mean of the acupuncture effect parameters of all historical knee osteoarthritis patients in each patient category as the target effect parameter for each patient category. For each acupuncture treatment parameter, the contribution of that acupuncture treatment parameter in each patient category was used as the y-value in a two-dimensional coordinate system, and the target effect parameter for each patient category was used as the x-value in the two-dimensional coordinate system to obtain the reference coordinates for each patient category. Then, a curve was fitted to the points corresponding to the reference coordinates in the two-dimensional coordinate system, which served as the therapeutic effect contribution curve for that acupuncture treatment parameter.

[0040] In order to analyze the impact of each acupuncture treatment parameter on the acupuncture treatment effect, the local efficacy feedback value of each acupuncture treatment parameter is obtained based on the changes in the efficacy contribution curve of each acupuncture treatment parameter. Among them, there can be multiple local efficacy feedback values for one acupuncture treatment parameter.

[0041] In one possible implementation of this embodiment, the method for obtaining the local efficacy feedback value is: for the efficacy contribution curve of any acupuncture treatment parameter, the extreme points on the efficacy contribution curve of the acupuncture treatment parameter are used as segmentation points, and the efficacy contribution curve of the acupuncture treatment parameter is divided into local curve segments; it should be noted that this embodiment uses the segmentation point as the end point of the local curve segment. For any local curve segment, the larger the contribution range corresponding to the local curve segment, the greater the effect of the acupuncture treatment parameter within the acupuncture effect parameter range corresponding to the local curve segment. The larger the slope of the local curve segment, the better the effect of the acupuncture treatment parameter within the acupuncture effect parameter range corresponding to the local curve segment on the acupuncture treatment effect. Therefore, this embodiment obtains the ratio of the contribution range of the local curve segment to the overall contribution range of the efficacy contribution curve of the acupuncture treatment parameter as the local feedback degree of the local curve segment. The product of the slope of the local curve segment and the local feedback degree is then used as a local efficacy feedback value for the acupuncture treatment parameter, thereby inferring a local efficacy feedback value for the acupuncture treatment parameter corresponding to the efficacy contribution curve of the acupuncture treatment parameter. It should be noted that the slope of the local curve segment is the slope of the straight line between the first and last points of the local curve segment. Furthermore, if the contribution range of the local curve segment is a single value, the local feedback degree of the local curve segment is defaulted to 0.

[0042] At this point, the local efficacy feedback value of each acupuncture treatment parameter is obtained. When the local efficacy feedback value is greater than 0 and the larger it is, the greater the positive influence of the corresponding acupuncture treatment parameter within the acupuncture effect parameter range corresponding to the corresponding local curve segment on the treatment effect. When the local efficacy feedback value is 0, the corresponding acupuncture treatment parameter within the acupuncture effect parameter range corresponding to the corresponding local curve segment has no influence on the treatment effect; when the local efficacy feedback value is less than 0 and the smaller it is, the greater the adverse influence of the corresponding acupuncture treatment parameter within the acupuncture effect parameter range corresponding to the corresponding local curve segment on the treatment effect. The local efficacy feedback value of each acupuncture treatment parameter accurately reflects the treatment feedback of each acupuncture treatment parameter on knee osteoarthritis.

[0043] Existing methods train an acupuncture efficacy prediction model using various acupuncture treatment parameters of historical knee osteoarthritis patients, the fixed weights corresponding to these parameters, and acupuncture effect parameters representing the treatment effects of these patients. This yields a trained acupuncture efficacy prediction model. The trained acupuncture efficacy prediction model is then used to predict the acupuncture treatment effects of these patients based on their various acupuncture treatment parameters. However, in reality, each patient with a history of knee osteoarthritis may have different acupuncture treatment parameters. When the magnitude of a particular acupuncture treatment parameter varies, the resulting acupuncture treatment effects may also vary. Therefore, during the training of the acupuncture efficacy prediction model, this embodiment dynamically obtains an evaluation score, or reference weight, for each acupuncture treatment parameter of each patient with a history of knee osteoarthritis, based on each acupuncture treatment parameter and the corresponding local efficacy feedback value. The larger the evaluation score, the better the acupuncture treatment effect produced by the corresponding acupuncture treatment parameter for the patient with a history of knee osteoarthritis. This facilitates subsequent accurate training of the acupuncture efficacy prediction model and improves the accuracy of the acupuncture efficacy prediction model.

[0044] Preferably, in one achievable manner of this embodiment, the method for obtaining the evaluation score is: for any patient with historical knee osteoarthritis and any acupuncture treatment parameter, the closer the acupuncture treatment parameter of the patient with historical knee osteoarthritis is to the acupuncture effect parameter range corresponding to the maximum local efficacy feedback value of the acupuncture treatment parameter, the better the treatment effect of the acupuncture treatment parameter of the patient with historical knee osteoarthritis on the patient with historical knee osteoarthritis, and the greater the reference significance of the acupuncture treatment parameter of the patient with historical knee osteoarthritis when analyzing the acupuncture treatment effect of the patient with historical knee osteoarthritis. Furthermore, in this embodiment, the mean value of the acupuncture treatment parameter of all historical knee osteoarthritis patients in the patient category corresponding to the maximum local efficacy feedback value of the acupuncture treatment parameter is used as the target analysis value; then, the absolute value of the difference between the acupuncture treatment parameter of the historical knee osteoarthritis patient and the target analysis value is used as the first important analysis value of the acupuncture treatment parameter of the historical knee osteoarthritis patient; the smaller the first important analysis value, the more likely the acupuncture treatment parameter of the historical knee osteoarthritis patient is to be within the range that has a favorable impact on the acupuncture treatment effect, indirectly reflecting that the acupuncture treatment parameter of the historical knee osteoarthritis patient has greater reference significance when analyzing the acupuncture treatment effect of the historical knee osteoarthritis patient; In order to more accurately analyze the influence of the acupuncture treatment parameter of the historical knee osteoarthritis patient on the acupuncture treatment effect of the historical knee osteoarthritis patient, the absolute value of the difference between the local efficacy feedback value corresponding to the patient category of the acupuncture treatment parameter of the historical knee osteoarthritis patient and the maximum local efficacy feedback value of the acupuncture treatment parameter is further used as the second important analysis value of the acupuncture treatment parameter of the historical knee osteoarthritis patient; the smaller the second important analysis value, the better the influence of the acupuncture treatment parameter of the historical knee osteoarthritis patient on the treatment effect, and the greater the reference degree of the acupuncture treatment parameter of the historical knee osteoarthritis patient should be; it should be noted that the essence of the change of the acupuncture treatment effect is caused by the change of the acupuncture treatment parameter. Therefore, the range of the acupuncture treatment parameter in the acupuncture effect parameter range corresponding to a certain local efficacy feedback value of the acupuncture treatment parameter is unique and fixed, which indirectly indicates that the acupuncture treatment parameter of the historical knee osteoarthritis patient will only correspond to one local efficacy feedback value; In order to accurately obtain the reference of the acupuncture treatment parameters of the patient with historical knee osteoarthritis, the first important analysis value and the second important analysis value are added together, negatively correlated and normalized, and the result is used as the evaluation score of the acupuncture treatment parameters of the patient with historical knee osteoarthritis. Negatively correlate and normalize the sum of the first important analysis value and the second important analysis value, where exp is an exponential function with a natural constant as the base; Refers to the sum of the first and second important analysis values.

[0045] At this point, obtaining the evaluation scores of various acupuncture treatment parameters for each historical knee osteoarthritis patient will facilitate the subsequent accurate training of the acupuncture efficacy prediction model.

[0046] The acupuncture efficacy prediction model acquisition module 40 is used to train the acupuncture efficacy prediction model based on each acupuncture treatment parameter, evaluation score and acupuncture effect parameter of each historical knee osteoarthritis patient to obtain a trained acupuncture efficacy prediction model.

[0047] Specifically, to train the acupuncture efficacy prediction model, this embodiment sets the training ratio of historical knee osteoarthritis patients to 7:3. The implementer can adjust the training ratio based on actual conditions and this ratio is not limited here. The loss function of the acupuncture efficacy prediction model is also set to mean square error. During the training process, each acupuncture treatment parameter and the evaluation score for each acupuncture treatment parameter for each historical knee osteoarthritis patient are used as inputs to the acupuncture efficacy prediction model. The acupuncture efficacy parameter for each historical knee osteoarthritis patient is used as the output of the acupuncture efficacy prediction model. The acupuncture efficacy prediction model is trained to obtain a trained acupuncture efficacy prediction model.

[0048] For any current knee osteoarthritis patient who needs acupuncture treatment, an acupuncture treatment plan for the knee osteoarthritis patient is set, and then various acupuncture treatment parameters of the knee osteoarthritis patient can be obtained, and then an evaluation score of each acupuncture treatment parameter of the knee osteoarthritis patient can be obtained. It should be noted that, in the process of obtaining the evaluation score of each acupuncture treatment parameter of the knee osteoarthritis patient, when a certain acupuncture treatment parameter of the knee osteoarthritis patient exceeds the maximum range of the acupuncture treatment parameter, the local efficacy feedback value corresponding to the maximum acupuncture treatment parameter is used as the local efficacy feedback value of the acupuncture treatment parameter of the knee osteoarthritis patient; when a certain acupuncture treatment parameter of the knee osteoarthritis patient is not within the range corresponding to the local efficacy feedback value of the acupuncture treatment parameter, but is between the ranges corresponding to two adjacent local efficacy feedback values, the local efficacy feedback value corresponding to the acupuncture treatment parameter of the knee osteoarthritis patient is calculated by interpolation according to the distance ratio. For example, taking electroacupuncture frequency as an example, assuming that the electroacupuncture frequency of the knee osteoarthritis patient is 4.5Hz, which is between the two adjacent electroacupuncture frequency ranges of 2Hz-4Hz and 5Hz-6Hz, wherein the local efficacy feedback value corresponding to 2Hz-4Hz is 0.9, and the local efficacy feedback value corresponding to 5Hz-6Hz is 0.5, then the local efficacy feedback value of the electroacupuncture frequency of the knee osteoarthritis patient is .

[0049] The various acupuncture treatment parameters and evaluation scores of the knee osteoarthritis patient are used as inputs of the trained acupuncture efficacy prediction model, and then the acupuncture effect parameters of the knee osteoarthritis patient are outputted through the acupuncture efficacy prediction model; when the acupuncture effect parameters of the knee osteoarthritis patient are less than the preset acupuncture effect threshold, the doctor needs to further optimize the various acupuncture treatment parameters of the knee osteoarthritis patient until the predicted acupuncture effect parameters of the knee osteoarthritis patient are greater than or equal to the preset acupuncture effect threshold, and determine the acupuncture treatment plan for the knee osteoarthritis patient. In this embodiment, the preset acupuncture effect threshold is set to 0.7, and the implementer can set the size of the preset acupuncture effect threshold according to actual conditions, which is not limited here. The efficacy prediction based on the acupuncture efficacy prediction model helps doctors predict the efficacy of the acupuncture treatment plan for knee osteoarthritis patients before acupuncture treatment, provides a basis for formulating personalized acupuncture treatment plans for knee osteoarthritis patients, and is conducive to improving the pertinence and effectiveness of acupuncture treatment, so as to achieve better acupuncture treatment effects and improve the satisfaction of knee osteoarthritis patients.

[0050] In summary, this embodiment obtains acupuncture treatment parameters and acupuncture effect parameters of historical knee osteoarthritis patients; divides historical knee osteoarthritis patients into patient categories based on the acupuncture effect parameters, and obtains the contribution degree according to the distribution of each acupuncture treatment parameter in each patient category and the difference in distribution with other patient categories; obtains a local efficacy feedback value based on the contribution degree and the acupuncture effect parameter, and then obtains an evaluation score; trains the acupuncture efficacy prediction model using the acupuncture treatment parameters, evaluation scores, and acupuncture effect parameters to obtain a trained acupuncture efficacy prediction model. The present invention accurately determines the participation degree of each acupuncture treatment parameter for each historical knee osteoarthritis patient by dynamically obtaining the evaluation score, effectively improving the accuracy of the acupuncture efficacy prediction model training.

[0051] Example 2: The present invention also proposes a deep learning-based acupuncture efficacy prediction device for knee osteoarthritis. The device includes a memory and a processor. The memory stores executable program code, and the processor is configured to call and execute the executable program code to implement the deep learning-based acupuncture efficacy prediction system for knee osteoarthritis provided in the embodiments of this application. The device can be a chip, component, or module. The chip may include a processor and memory connected together. The memory is configured to store instructions. When the processor calls and executes the instructions, the chip executes the deep learning-based acupuncture efficacy prediction system for knee osteoarthritis provided in the embodiments described above.

[0052] In addition, the present application also protects a computer device, see Figure 3 The computer device includes a memory 401, a processor 402, and a computer program 403 stored in the memory 401 and running on the processor 402, wherein when the processor 402 executes the computer program 403, the computer device can execute any one of the deep learning-based knee osteoarthritis acupuncture efficacy prediction systems introduced above.

[0053] Example 3: The present invention also provides a computer-readable storage medium, which stores computer program code. When the computer program code is run on a computer, the computer executes the above-mentioned related method steps to implement a deep learning-based knee osteoarthritis acupuncture efficacy prediction system provided in the above embodiment.

[0054] Example 4: The present invention also provides a computer program product. When the computer program product is run on a computer, it enables the computer to execute the above-mentioned related steps to implement a deep learning-based knee osteoarthritis acupuncture efficacy prediction system provided in the above embodiment.

[0055] Among them, the device, computer-readable storage medium, computer program product or chip provided in this embodiment are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0056] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0057] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A deep learning-based acupuncture efficacy prediction system for knee osteoarthritis, characterized by: The system includes the following steps: A data acquisition module is used to obtain various acupuncture treatment parameters and acupuncture effect parameters of each historical knee osteoarthritis patient; a contribution degree acquisition module, for dividing historical knee osteoarthritis patients into patient categories based on acupuncture effect parameters, and obtaining the contribution degree of each acupuncture treatment parameter in each patient category based on the distribution of each acupuncture treatment parameter in each patient category and the difference in distribution with other patient categories; An evaluation score acquisition module is used to obtain the efficacy contribution curve of each acupuncture treatment parameter based on the contribution degree and the acupuncture effect parameter in each patient category; obtain the local efficacy feedback value based on the changes in the efficacy contribution curve; and obtain the evaluation score of each acupuncture treatment parameter of each historical knee osteoarthritis patient based on each acupuncture treatment parameter and the corresponding local efficacy feedback value; The acupuncture efficacy prediction model acquisition module is used to train the acupuncture efficacy prediction model through each acupuncture treatment parameter, evaluation score and acupuncture effect parameter of each historical knee osteoarthritis patient to obtain a trained acupuncture efficacy prediction model.

2. The deep learning-based acupuncture efficacy prediction system for knee osteoarthritis according to claim 1, characterized in that: The method for obtaining the contribution degree is: According to the distribution of each acupuncture treatment parameter in each patient category, the compactness of each acupuncture treatment parameter in each patient category is obtained; According to the compactness difference and size difference of each acupuncture treatment parameter in each patient category and other patient categories, the special degree of each acupuncture treatment parameter in each patient category is obtained; The product of the compactness and specificity of each acupuncture treatment parameter in each patient category is taken as the contribution of each acupuncture treatment parameter in each patient category.

3. The deep learning-based acupuncture efficacy prediction system for knee osteoarthritis according to claim 2, characterized in that: The method for obtaining the compactness is: For any acupuncture treatment parameter and any patient category, the DBSCAN density clustering algorithm is used to cluster the historical knee osteoarthritis patients within the patient category based on the magnitude of the acupuncture treatment parameter for each patient within the patient category, and the local category corresponding to the acupuncture treatment parameter in the patient category is obtained; The number of patients with historical knee OA in each local category was obtained and used as the number of local patients; The ratio of the maximum number of local patients to the total number of all historical knee osteoarthritis patients in this patient category is used as the reference stability of this acupuncture treatment parameter in this patient category; The product of the reference stability level and the inverse of the number of local categories is used as the compactness of the acupuncture treatment parameter in the patient category.

4. The deep learning-based acupuncture efficacy prediction system for knee osteoarthritis according to claim 2, characterized in that: The method for obtaining the special degree is: For any acupuncture treatment parameter and any patient category, obtain the mean value of the compactness of the acupuncture treatment parameter in all patient categories except the patient category as the first eigenvalue; The difference between the compactness of the acupuncture treatment parameter in the patient category and the first characteristic value is used as the first discrimination degree of the acupuncture treatment parameter in the patient category; Obtain the mean value of the acupuncture treatment parameter for all patients with a history of knee osteoarthritis in the patient category as the overall reference value of the acupuncture treatment parameter in the patient category; Obtaining the mean of the overall reference values of the acupuncture treatment parameter in all patient categories except the patient category as the second eigenvalue; The difference between the overall reference value of the acupuncture treatment parameter in the patient category and the second characteristic value is used as the second discrimination degree of the acupuncture treatment parameter in the patient category; The result of normalizing the sum of the first degree of distinction and the second degree of distinction is used as the special degree of the acupuncture treatment parameter in the patient category.

5. The deep learning-based acupuncture efficacy prediction system for knee osteoarthritis according to claim 1, characterized in that: The method for obtaining the efficacy contribution curve is: The mean of the acupuncture effect parameters of all patients with historical knee OA in each patient category was used as the target effect parameter for each patient category; For any acupuncture treatment parameter, the contribution of the acupuncture treatment parameter in each patient category is taken as the y value in the two-dimensional coordinate, and the target effect parameter of each patient category is taken as the x value in the two-dimensional coordinate to obtain the reference coordinate of each patient category corresponding to the acupuncture treatment parameter; The points corresponding to the reference coordinates in the two-dimensional coordinate system are fitted into a curve as a contribution curve of the therapeutic effect of the acupuncture treatment parameters.

6. The deep learning-based acupuncture efficacy prediction system for knee osteoarthritis according to claim 1, characterized in that: The method for obtaining the local therapeutic effect feedback value is: For any efficacy contribution curve of acupuncture treatment parameters, the extreme value points on the efficacy contribution curve of the acupuncture treatment parameters are all used as segmentation points, and the efficacy contribution curve of the acupuncture treatment parameters is divided into multiple local curve segments; For any local curve segment, the ratio of the contribution range of the local curve segment to the overall contribution range of the efficacy contribution curve of the acupuncture treatment parameter is obtained as the local feedback degree of the local curve segment; The product of the slope of the local curve segment and the local feedback degree is used as a local efficacy feedback value of the acupuncture treatment parameter.

7. The deep learning-based acupuncture efficacy prediction system for knee osteoarthritis according to claim 6, characterized in that: The method for obtaining the evaluation score is: For any patient with a history of knee osteoarthritis and any acupuncture treatment parameter, the mean value of the acupuncture treatment parameter of all patients with a history of knee osteoarthritis in the patient category corresponding to the maximum local efficacy feedback value of the acupuncture treatment parameter is used as the target analysis value; The difference between the acupuncture treatment parameter of the patient with historical knee osteoarthritis and the target analysis value is used as the first important analysis value of the acupuncture treatment parameter of the patient with historical knee osteoarthritis; The difference between the local efficacy feedback value corresponding to the patient category of the acupuncture treatment parameter of the historical knee osteoarthritis patient and the maximum local efficacy feedback value of the acupuncture treatment parameter is used as the second important analysis value of the acupuncture treatment parameter of the historical knee osteoarthritis patient; The first important analysis value and the second important analysis value are added together, negatively correlated, and normalized to obtain the result, which is used as the evaluation score of the acupuncture treatment parameter for the patient with a history of knee osteoarthritis.

8. The deep learning-based acupuncture efficacy prediction system for knee osteoarthritis according to claim 1, characterized in that: The method for obtaining the acupuncture efficacy prediction model is: Each acupuncture treatment parameter and the evaluation score of each acupuncture treatment parameter of each historical knee osteoarthritis patient are used as the input of the acupuncture efficacy prediction model, and the acupuncture effect parameter of each historical knee osteoarthritis patient is used as the output of the acupuncture efficacy prediction model. The acupuncture efficacy prediction model is trained to obtain a trained acupuncture efficacy prediction model.

9. The deep learning-based acupuncture efficacy prediction system for knee osteoarthritis according to claim 1, characterized in that: The method for obtaining the patient category is: The DBSCAN density clustering algorithm was used to cluster the historical knee osteoarthritis patients according to the acupuncture effect parameters of each historical knee osteoarthritis patient to obtain the patient categories.

10. The deep learning-based acupuncture efficacy prediction system for knee osteoarthritis according to claim 1, characterized in that: The acupuncture effect parameter is a score given by doctors based on the treatment effect of each patient with a history of knee osteoarthritis after acupuncture treatment.

Images