A method and system for recommending traditional Chinese medicine rehabilitation programs

Abstract

The present invention relates to the field of TCM rehabilitation program recommendation, and specifically to a TCM rehabilitation program recommendation method and system. The method first obtains the examination reports of the patient to be tested and each historical patient, and obtains the rehabilitation program of each historical patient, performs natural language processing on each examination report and rehabilitation program, obtains the word vector of the phrases in the examination report and the rehabilitation program, obtains the symptom similarity of each historical patient based on the word vector difference of the phrases in the examination report between the patient to be tested and the historical patient, and screens out similar patients of the patient to be tested, obtains the program difference index of the patient to be tested based on the word vector difference of the phrases in the rehabilitation program between each similar patient, and recommends TCM rehabilitation program to the patient to be tested in different ways based on the program difference index. The present invention can accurately recommend TCM rehabilitation programs for patients and improve the accuracy of the TCM rehabilitation program recommendation system.

Description

Technical Field

[0001] The present invention relates to the field of TCM rehabilitation program recommendation, and in particular to a TCM rehabilitation program recommendation method and system. Background Art

[0002] The TCM rehabilitation program is based on TCM theory and practice, and aims to treat diseases and restore health by adjusting the body's yin and yang balance, qi and blood circulation, and organ functions. It is formulated according to each person's specific situation and can better meet the needs of patients. Therefore, the correct recommendation of the TCM rehabilitation program can provide patients with a more comprehensive, personalized, and integrated treatment plan, which is of great significance in the field of TCM.

[0003] In the related art, the patient's condition is usually analyzed, and the rehabilitation plans of other patients with similar symptoms are recommended to the patient. However, since Chinese medicine pays more attention to individual differences among patients, there are also large differences in the rehabilitation plans required by patients with similar symptoms. As a result, existing methods are unable to accurately recommend rehabilitation plans for specific patients, thereby reducing the accuracy of the Chinese medicine rehabilitation plan recommendation system. Summary of the invention

[0004] In order to solve the technical problem that the existing methods cannot accurately recommend rehabilitation programs for specific patients, thereby reducing the accuracy of the TCM rehabilitation program recommendation system, the purpose of the present invention is to provide a TCM rehabilitation program recommendation method and system, and the technical solutions adopted are as follows:

[0005] The present invention proposes a method for recommending a traditional Chinese medicine rehabilitation program, the method comprising:

[0006] Obtain the examination reports of the patient to be tested and each historical patient, and obtain the rehabilitation plan of each historical patient, perform natural language processing on each examination report and rehabilitation plan, and obtain the word vector of each phrase in the examination report and rehabilitation plan;

[0007] According to the difference in the word vectors of the phrases in the examination reports between the patient to be tested and each historical patient, the symptom similarity of each historical patient is obtained; based on the symptom similarity, similar patients of the patient to be tested are screened out from all historical patients; according to the difference in the word vectors of the phrases in the rehabilitation plans between the similar patients, the plan difference index of the patient to be tested is obtained;

[0008] If the plan difference index of the patient to be tested is less than the preset difference threshold, the recommended rehabilitation plan for the patient to be tested is selected from the rehabilitation plans of all similar patients according to the symptom similarity of each similar patient; if the plan difference index of the patient to be tested is not less than the preset difference threshold, any similar patient is taken as the target similar patient, and the priority recommendation coefficient of the target similar patient is obtained according to the difference in word vectors of phrases in the rehabilitation plans between the target similar patient and other similar patients except the target similar patient, and the symptom similarity of the target similar patient; and based on the priority recommendation coefficient of each similar patient, the recommended rehabilitation plan for the patient to be tested is selected from the rehabilitation plans of all similar patients.

[0009] Furthermore, obtaining the symptom similarity of each historical patient includes:

[0010] Taking any historical patient as the target historical patient, selecting any phrase from the examination report of the patient to be tested and the examination report of the target historical patient, respectively, and combining the two selected phrases as a first phrase;

[0011] The absolute value of the cosine similarity between the word vectors of the two word groups in each first word group combination is used as the similarity degree of each first word group combination;

[0012] The average value of the similarity levels of all first word group combinations is taken as the symptom similarity of the target historical patient.

[0013] Furthermore, based on the symptom similarity, similar patients of the patient to be tested are screened from all historical patients, including:

[0014] According to the similarity of the symptoms of each historical patient, all historical patients are clustered to obtain multiple clusters;

[0015] The average of the symptom similarities of all historical patients in each cluster is taken as the cluster center value of each cluster;

[0016] The historical patients in the cluster corresponding to the maximum value of the cluster center value are taken as similar patients to the patient to be tested.

[0017] Furthermore, the step of obtaining the regimen difference index of the patient to be tested includes:

[0018] Taking any two similar patients as a first patient group, selecting any one phrase from the rehabilitation programs of the two similar patients in each first patient group, and taking the selected two phrases as a second phrase combination of each first patient group;

[0019] Performing negative correlation mapping on the Euclidean distance between the word vectors of two word groups in each second word group combination to obtain a similarity parameter of each second word group combination, and normalizing the average value of the similarity parameters of all second word group combinations to obtain the scheme similarity of each first patient group;

[0020] According to the distribution of the regimen similarities of all the first patient groups, a regimen difference index of the patient to be tested is obtained.

[0021] Furthermore, obtaining the regimen difference index of the patient to be tested according to the distribution of the regimen similarities of all the first patient groups includes:

[0022] Performing negative correlation mapping on the average values ​​of the scheme similarities of all first patient groups to obtain a first difference parameter of the patient to be tested;

[0023] Using the standard deviation of the similarity of the regimens of all the first patient groups as the second difference parameter of the patients to be tested;

[0024] The first difference parameter and the second difference parameter are integrated and normalized to obtain a regimen difference index for the patient to be tested.

[0025] Furthermore, selecting a recommended rehabilitation program for the patient to be tested from the rehabilitation programs of all similar patients according to the symptom similarity of each similar patient includes:

[0026] The rehabilitation plans of similar patients corresponding to a preset number of greatest symptom similarities are selected as the recommended rehabilitation plans for the patients to be tested.

[0027] Furthermore, obtaining the priority recommendation coefficient of the target similar patient includes:

[0028] The target similar patient and any other similar patient except the target similar patient are regarded as a second patient group, and the scheme similarity coefficient of each second patient group is obtained by using the calculation method of the scheme similarity of the first patient group;

[0029] Performing negative correlation mapping on the average values ​​of the scheme similarity coefficients of all second patient groups to obtain the scheme independence degree of target similar patients;

[0030] The regimen independence degree and symptom similarity of target similar patients are integrated and normalized to obtain the priority recommendation coefficient of target similar patients.

[0031] Furthermore, the step of selecting a recommended rehabilitation program for the patient to be tested from the rehabilitation programs of all similar patients based on the priority recommendation coefficient of each similar patient includes:

[0032] The priority recommendation coefficient of each similar patient is used as a parameter used by the ant colony algorithm, and the rehabilitation plans of all similar patients are input into the ant colony algorithm, which then outputs the recommended rehabilitation plan for the patient to be tested.

[0033] Furthermore, obtaining the word vector of each word group in the examination report and the rehabilitation plan includes:

[0034] The jieba word segmentation tool is used to segment each examination report and rehabilitation plan to obtain the phrases in each examination report and rehabilitation plan. The Word2Vec algorithm is used to process the phrases in the examination report and rehabilitation plan to obtain the word vectors of the phrases in each examination report and rehabilitation plan.

[0035] The present invention also proposes a TCM rehabilitation program recommendation system, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any one of the steps of a TCM rehabilitation program recommendation method when executing the computer program.

[0036] The present invention has the following beneficial effects:

[0037] The present invention takes into account that the existing methods cannot accurately recommend rehabilitation plans for specific patients, thereby reducing the accuracy of the TCM rehabilitation plan recommendation system. Therefore, the present invention first obtains the examination reports of the patient to be tested and each historical patient, and obtains the rehabilitation plan of each historical patient, and performs natural language processing on them to achieve quantification of the examination reports and rehabilitation plans, which is convenient for subsequent data analysis. Taking into account the different degrees of difference in the physical conditions of the patient to be tested and the historical patients, in the rehabilitation plan recommendation system, the rehabilitation plans of historical patients with similar physical conditions to the patient to be tested should be recommended to the patient to be tested first. Therefore, the similarity of the physical conditions between each historical patient and the patient to be tested can be reflected through the similarity of symptoms, and then similar patients to the patient to be tested can be screened out, so as to facilitate the subsequent selection of the best rehabilitation plan from the rehabilitation plans of similar patients and recommend it to the patient to be tested, thereby improving the accuracy of the TCM rehabilitation plan recommendation system. Since TCM pays more attention to patients Due to the differences in individual conditions and the overall treatment of patients, there are also large differences in the rehabilitation programs used even between patients with similar medical conditions. Therefore, the obtained program difference index reflects the degree of difference between the rehabilitation programs used by similar patients of the patient to be tested, so as to facilitate the subsequent recommendation of different types of rehabilitation programs for the patient to be tested based on the program difference index. When the program difference index is less than the preset difference threshold, it means that the rehabilitation programs of similar patients are basically suitable for the patient to be tested. At this time, the rehabilitation programs of similar patients with higher similarity in physical conditions with the patient to be tested can be recommended to the patient to be tested first. Then, according to the similarity of symptoms of each similar patient, accurate rehabilitation programs can be recommended for the patient to be tested. When the program difference index is not less than the preset difference threshold, it means that the rehabilitation programs of similar patients are quite different. At this time, the priority recommendation coefficient reflects the degree to which the rehabilitation programs of similar patients are given priority to the patient to be tested, thereby achieving accurate recommendation of traditional Chinese medicine rehabilitation programs for the patient to be tested. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. 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 creative work.

[0039] Figure 1 A flow chart of a method for recommending a TCM rehabilitation program provided by one embodiment of the present invention;

[0040] Figure 2 A flow chart of a method for obtaining a regimen difference index for a patient to be tested provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a method and system for recommending a TCM rehabilitation program proposed by the present invention, its specific implementation method, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0042] 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.

[0043] The specific scheme of a method and system for recommending a traditional Chinese medicine rehabilitation program provided by the present invention is described in detail below with reference to the accompanying drawings.

[0044] See also Figure 1 , which shows a flow chart of a method for recommending a TCM rehabilitation program provided by an embodiment of the present invention, the method comprising:

[0045] Step S1: Obtain the examination reports of the patient to be tested and each historical patient, and obtain the rehabilitation plan of each historical patient, perform natural language processing on each examination report and rehabilitation plan, and obtain the word vector of the phrase in each examination report and rehabilitation plan.

[0046] When conducting a TCM examination on a patient, medical staff will test the physical condition of each patient, such as pulse, heart rate and tongue condition, to form an examination report for the patient and save it in the medical system database. At the same time, medical staff will provide the patient with a corresponding rehabilitation plan based on the patient's examination report and combined with the patient's personal living habits, physical condition, medical history and other information, and save it in the medical system database along with the examination report.

[0047] Therefore, the embodiment of the present invention first conducts an on-site examination of the patient to be tested to obtain the examination report of the patient to be tested, and then collects the examination report and rehabilitation plan of each historical patient from the medical system database, wherein the historical patient refers to the patient who has been examined and treated, and the examination report contains the physical condition information of the patient to be tested and the historical patient regarding the traditional Chinese medicine examination, and the rehabilitation plan contains the specific method of traditional Chinese medicine treatment for the historical patient. Since the information in the examination report and the rehabilitation plan is usually presented in text form, it is not conducive to subsequent data processing. Therefore, the embodiment of the present invention needs to perform natural language processing on each examination report and rehabilitation plan to obtain the word vector of each phrase in the examination report and rehabilitation plan, so as to realize the quantification of the examination report and rehabilitation plan. Subsequently, based on the difference in the word vector of the phrase between the examination report or the rehabilitation plan, the differences in the patient's examination report and the rehabilitation plan can be accurately analyzed, so as to accurately recommend the best rehabilitation plan for the patient to be tested.

[0048] Preferably, in one embodiment of the present invention, word segmentation tools such as jieba or spaCy are used to perform word segmentation processing on each examination report and rehabilitation plan, respectively, to obtain the phrases in each examination report and rehabilitation plan, and the Word2Vec algorithm model is used to process the phrases in the examination report and rehabilitation plan to generate word vectors for each phrase, thereby obtaining the word vectors for the phrases in each examination report and rehabilitation plan, wherein each examination report contains multiple phrases, each phrase corresponds to a word vector, and the same is true for the rehabilitation plan.

[0049] Among them, the Word2Vec algorithm model is a technical means well known to those skilled in the art and will not be elaborated here.

[0050] Step S2: According to the difference in word vectors of phrases in the examination reports between the patient to be tested and each historical patient, the symptom similarity of each historical patient is obtained; based on the symptom similarity, similar patients of the patient to be tested are screened out from all historical patients; according to the difference in word vectors of phrases in the rehabilitation plans between similar patients, the plan difference index of the patient to be tested is obtained.

[0051] Since the degree of difference in the physical condition between the patient to be tested and different historical patients is different, in the TCM rehabilitation program recommendation system, the rehabilitation program of historical patients with similar physical condition to the patient to be tested should be recommended to the patient to be tested first. The examination report contains information about the physical condition of the patient to be tested or the historical patients. Therefore, the difference in the word vectors of the phrases in the examination report between the patient to be tested and each historical patient can be analyzed. The similarity of the physical condition between each historical patient and the patient to be tested can be reflected by the obtained symptom similarity, which is convenient for the subsequent accurate screening of historical patients with similar physical condition to the patient to be tested based on symptom similarity, so as to better target the actual condition of the patient to be tested and accurately recommend the corresponding TCM rehabilitation program to the patient to be tested.

[0052] Preferably, in one embodiment of the present invention, the method for obtaining the symptom similarity of each historical patient specifically includes:

[0053] First, any historical patient is taken as the target historical patient, and a phrase is randomly selected from the examination report of the patient to be tested and the examination report of the target historical patient respectively, and the two selected phrases are used as a first phrase combination. Then, for the two phrases in the first phrase combination, one belongs to the examination report of the patient to be tested, and the other belongs to the examination report of the target historical patient.

[0054] Then, the absolute value of the cosine similarity between the word vectors of the two phrases in each first phrase combination is used as the similarity of each first phrase combination. The greater the similarity, the more similar the two phrases in the first phrase combination are. Then, the average value of the similarity of all first phrase combinations can be used as the symptom similarity of the target historical patient. The greater the symptom similarity, the more similar the content information of the examination report of the target historical patient and the examination report of the patient to be tested is, and then the physical condition of the target historical patient and the patient to be tested is more similar.

[0055] As an example, in one embodiment of the present invention, the expression of the symptom similarity of the target historical patient may be specifically, for example, as follows:

[0056]

[0057] in, Indicates the symptom similarity of the target historical patient; Indicates The cosine similarity between the word vectors of the two words in the first word combination; Indicates The similarity of the first phrase combination; Indicates the number of first phrase combinations.

[0058] The symptom similarity of each historical patient can be obtained through the same method as mentioned above. The greater the symptom similarity, the more similar the physical condition of each historical patient is to the patient to be tested. Therefore, based on the symptom similarity, similar patients to the patient to be tested can be screened out from all historical patients. Subsequently, the best rehabilitation plan can be selected from the rehabilitation plans of similar patients and recommended to the patient to be tested, thereby improving the accuracy of the TCM rehabilitation plan recommendation system.

[0059] Preferably, in one embodiment of the present invention, the method for obtaining similar patients of the patient to be tested specifically includes:

[0060] According to the symptom similarity of each historical patient, all historical patients are clustered to obtain multiple clusters. The physical conditions of historical patients in the same cluster are relatively similar. The average value of the symptom similarity of all historical patients in each cluster is used as the cluster center value of each cluster. The larger the cluster center value, the more similar the physical condition of the historical patients in the cluster is to the patient to be tested. Therefore, the historical patients in the cluster corresponding to the maximum value of the cluster center value can be used as similar patients to the patient to be tested. In one embodiment of the present invention, the existing ISODATA clustering algorithm can be used to cluster historical patients. In other embodiments of the present invention, for example, the K-means clustering algorithm or the DBSCAN clustering algorithm can also be used, which is not limited here.

[0061] In other embodiments of the present invention, the symptom similarity of each historical patient can be compared with the threshold by setting a threshold, so as to screen out similar patients to the patient to be tested. For example, historical patients whose symptom similarity is greater than a preset similarity threshold can be used as similar patients to the patient to be tested, wherein the preset similarity threshold is set to 0.6, and the specific value of the preset similarity threshold can also be set by the implementer according to the specific implementation scenario, which is not limited here.

[0062] Since TCM pays more attention to the differences in individual conditions of patients and the overall treatment of patients, there may be large differences in the rehabilitation plans used even between patients with similar conditions. The rehabilitation plan is a personal treatment plan provided by medical staff to patients based on the patient's examination report and combined with the patient's personal living habits, physical condition, medical history and other information. Therefore, the rehabilitation plan can reflect the individual characteristics of the patient. Therefore, it is necessary to obtain the plan difference index of the patient to be tested based on the difference in the word vectors of the phrases in the rehabilitation plans between similar patients. The plan difference index reflects the degree of difference between the rehabilitation plans used by similar patients of the patient to be tested. The smaller the plan difference index, the more similar the rehabilitation plans used by similar patients of the patient to be tested. Conversely, the greater the difference in the rehabilitation plans used by similar patients of the patient to be tested. Subsequently, different rehabilitation plans are recommended for the patient to be tested based on the plan difference index to improve the accuracy of the TCM rehabilitation plan recommendation system.

[0063] Preferably, in one embodiment of the present invention, the method for obtaining the regimen difference index of the patient to be tested specifically includes:

[0064] See also Figure 2 , which shows a flow chart of a method for obtaining a regimen difference index for a patient to be tested provided by an embodiment of the present invention.

[0065] Step S201: taking any two similar patients as a first patient group, selecting any one phrase from the rehabilitation programs of the two similar patients in each first patient group, and taking the selected two phrases as a second phrase combination of each first patient group.

[0066] The two phrases in the second phrase combination are respectively from the rehabilitation programs of two similar patients in the first patient group.

[0067] Step S202: negatively correlate the Euclidean distance between the word vectors of the two word groups in each second word group combination to obtain the word group similarity parameter of each second word group combination, and normalize the average value of the similarity parameters of all second word group combinations to obtain the scheme similarity of each first patient group.

[0068] The smaller the Euclidean distance between the word vectors of the two word groups in the second word group combination, the more similar the two word groups in the second word group combination are. Therefore, the Euclidean distance between the word vectors of the two word groups in each second word group combination can be negatively correlated to obtain the similarity parameters of each second word group combination, and then the average values ​​of the similarity parameters of all second word group combinations are normalized to obtain the scheme similarity of each first patient group. The greater the scheme similarity, the more similar the content information of the rehabilitation schemes of two similar patients in the first patient group is. Subsequently, based on the scheme similarities of each first patient group, the scheme difference indicators of the patients to be tested can be accurately calculated and analyzed.

[0069] In one embodiment of the present invention, the normalization processing can be specifically, for example, maximum and minimum value normalization processing, and the normalization in subsequent steps can all adopt maximum and minimum value normalization processing. In other embodiments of the present invention, other normalization methods can be selected according to the specific range of numerical values, which will not be described in detail.

[0070] As an example, in one embodiment of the present invention, the expression of the scheme similarity of each first patient group may be specifically, for example, as follows:

[0071]

[0072] in, Indicates The similarity of the regimens in the first patient group; Indicates The first patient group The Euclidean distance between the word vectors of the two word groups in the second word group combination; Indicates The first patient group Similar parameters for the second phrase combination; Indicates the number of second word group combinations in the first patient group; Represents a normalization function, which is used to limit the calculation results to scope; Indicates the preset first adjustment parameter, which is used to prevent the denominator from being 0. The value range is In one embodiment of the present invention, Set to 0.01, The specific value of can also be set by the implementer according to the specific implementation scenario and is not limited here.

[0073] It should be noted that in other embodiments of the present invention, negative correlation mapping may be achieved through other basic mathematical operations, which will not be elaborated herein.

[0074] Step S203: Obtaining a regimen difference index of the patient to be tested according to the distribution of regimen similarities of all first patient groups.

[0075] The smaller the overall level of scheme similarity of all first patient groups and the worse the consistency of scheme similarity of all first patient groups, the greater the difference in rehabilitation schemes used between similar patients to be tested. Therefore, the scheme difference index of the patient to be tested can be obtained based on the distribution of scheme similarity of all first patient groups.

[0076] Preferably, in one embodiment of the present invention, the method for obtaining the regimen difference index of the patient to be tested further comprises:

[0077] The average values ​​of the scheme similarities of all the first patient groups are negatively correlated and mapped to obtain the first difference parameter of the patient to be tested. The larger the first difference parameter is, the smaller the overall level of scheme similarity of all the first patient groups is, which further indicates that the difference in rehabilitation schemes used between similar patients is greater. The standard deviation of the scheme similarities of all the first patient groups is used as the second difference parameter of the patient to be tested. The larger the second difference parameter is, the worse the consistency of the scheme similarities of all the first patient groups is, which further indicates that the difference in rehabilitation schemes used between similar patients to be tested is greater.

[0078] Then, the first difference parameter and the second difference parameter may be integrated and normalized to obtain a regimen difference index for the patient to be tested.

[0079] In the embodiment of the present invention, the integration of the first difference parameter and the second difference parameter can be achieved by calculating the sum or product of the two, which is not limited here.

[0080] As an example, in one embodiment of the present invention, the expression of the regimen difference index of the patient to be tested can be specifically, for example, as follows:

[0081]

[0082] in, represents the regimen difference index for the patients to be tested; represents the average of the regimen similarity of all first patient groups; represents a first difference parameter of the patient to be tested; represents the standard deviation of the similarity of the regimens of all first patient groups, i.e., the second difference parameter of the patients to be tested; Represents the hyperbolic tangent function, which is used for normalization and limits the calculation results to scope; Indicates the preset second adjustment parameter, which is used to prevent the denominator from being 0. The value range is In one embodiment of the present invention, Set to 0.01, The specific value of can also be set by the implementer according to the specific implementation scenario and is not limited here.

[0083] It should be noted that in other embodiments of the present invention, negative correlation mapping may be achieved through other basic mathematical operations, which will not be elaborated herein.

[0084] At this point, the regimen difference index of the patient to be tested has been obtained. Subsequently, based on the regimen difference index, different methods can be used to recommend rehabilitation regimens for the patient to be tested, thereby improving the accuracy of TCM rehabilitation regimen recommendations.

[0085] Step S3: If the plan difference index of the patient to be tested is less than the preset difference threshold, then according to the symptom similarity of each similar patient, a recommended rehabilitation plan for the patient to be tested is selected from the rehabilitation plans of all similar patients; if the plan difference index of the patient to be tested is not less than the preset difference threshold, then any similar patient is taken as the target similar patient, and according to the difference in word vectors of phrases in the rehabilitation plans between the target similar patient and other similar patients except the target similar patient, and the symptom similarity of the target similar patient, the priority recommendation coefficient of the target similar patient is obtained, and based on the priority recommendation coefficient of each similar patient, a recommended rehabilitation plan for the patient to be tested is selected from the rehabilitation plans of all similar patients.

[0086] The smaller the program difference index of the patient to be tested, the more similar the rehabilitation programs used by similar patients of the patient to be tested are, which means that the individual situation characteristics of the patient to be tested are relatively common, and further, it means that the rehabilitation programs of similar patients are also suitable for the patient to be tested. Conversely, the larger the program difference index of the patient to be tested, the more different the rehabilitation programs used by similar patients of the patient to be tested are, which means that the individual situation characteristics of the patient to be tested are relatively niche, and further, it means that not all rehabilitation programs for similar patients are suitable for the patient to be tested. Therefore, it is necessary to recommend rehabilitation programs for the patient to be tested based on the program difference index of the patient to be tested and using different methods.

[0087] The embodiment of the present invention uses a threshold setting method. If the plan difference index of the patient to be tested is less than a preset difference threshold, it means that the rehabilitation plans of similar patients are also suitable for the patient to be tested. In order to ensure that the recommended rehabilitation plan can better target the actual condition of the patient to be tested, the recommended rehabilitation plan for the patient to be tested can be selected from the rehabilitation plans of all similar patients according to the symptom similarity of each similar patient, thereby improving the accuracy of recommending a traditional Chinese medicine rehabilitation plan for the patient to be tested. The preset difference threshold is set to 0.5, and the specific value of the preset difference threshold can also be set by the implementer according to the specific implementation scenario, which is not limited here.

[0088] Preferably, in one embodiment of the present invention, the method for obtaining the recommended rehabilitation plan for the patient to be tested specifically includes:

[0089] From the above analysis, it can be seen that in this case, the rehabilitation plans of each similar patient are also applicable to the patient to be tested. At this time, the rehabilitation plans of similar patients with higher similarity to the physical condition of the patient to be tested need to be recommended to the patient to be tested first, so as to ensure that the recommended rehabilitation plan can better target the actual condition of the patient to be tested. Therefore, the rehabilitation plans of similar patients with the greatest symptom similarity corresponding to a preset number are selected as the recommended rehabilitation plans for the patient to be tested. In one embodiment of the present invention, the preset number is set to 3, and the specific value of the preset number can also be set by the implementer according to the specific implementation scenario, which is not limited here.

[0090] If the program difference index of the patient to be tested is not less than the preset difference threshold, it means that the individual situation characteristics of the patient to be tested are relatively niche, which means that not all rehabilitation programs for similar patients are suitable for the patient to be tested. At this time, when recommending a traditional Chinese medicine rehabilitation program for the patient to be tested, it is necessary to consider not only the similarity of the disease conditions between the patient to be tested and similar patients, but also the independence of the individual characteristics of the patient to be tested. That is to say, in this case, it is necessary to give priority to recommending the rehabilitation programs of similar patients with similar disease conditions to the patient to be tested and with equally niche individual characteristics to the patient to be tested, so as to ensure the accuracy of the recommendation of traditional Chinese medicine rehabilitation programs for the patient to be tested. Therefore, any similar patient can be analyzed first, and any similar patient can be used as the target similar patient. The difference in the word vectors of the phrases in the rehabilitation program between the target similar patient and other similar patients except the target similar patient, as well as the symptom similarity of the target similar patient, are analyzed. The priority recommendation coefficient obtained reflects the degree to which the rehabilitation program of the target similar patient is given priority to the patient to be tested. Subsequently, based on the priority recommendation coefficient, the rehabilitation program can be accurately recommended to the patient to be tested.

[0091] Preferably, in one embodiment of the present invention, the method for obtaining the priority recommendation coefficient of the target similar patient specifically includes:

[0092] First, the target similar patient and any other similar patient except the target similar patient are regarded as a second patient group. For each second patient group, it contains the target similar patient and one other similar patient, and the number of second patient groups is equal to the number of other similar patients. Then, the calculation method of the scheme similarity of the first patient group is used to obtain the scheme similarity coefficient of each second patient group. The specific process is as follows: randomly select a phrase from the rehabilitation schemes of two similar patients in each second patient group, and use the selected two phrases as a third phrase combination for each second patient group; negatively correlate the Euclidean distance between the word vectors of the two groups in each third phrase combination to obtain the similarity factor of each third phrase combination, and normalize the average value of the similarity factors of all third phrase combinations to obtain the scheme similarity coefficient of each second patient group.

[0093] As an example, in one embodiment of the present invention, the expression of the regimen similarity coefficient of each second patient group may be specifically, for example, as follows:

[0094]

[0095] in, Indicates The regimen similarity coefficient for the second patient group; Indicates The second patient group The Euclidean distance between the word vectors of the two words in the third word combination; Indicates The second patient group Similarity factor of the third phrase combination; Indicates the number of third word combinations in the second patient group; Represents a normalization function, which is used to limit the calculation results to scope; Indicates the preset first adjustment parameter, which is used to prevent the denominator from being 0.

[0096] The smaller the scheme similarity coefficient of the second patient group is, the greater the difference in rehabilitation schemes between the target similar patients and other similar patients is. Therefore, the average values ​​of the scheme similarity coefficients of all the second patient groups can be negatively correlated to obtain the scheme independence of the target similar patients. The greater the scheme independence, the more independent the individual characteristics of the target similar patients are. This means that the rehabilitation schemes of the target similar patients are more suitable for the patients to be tested and need to be recommended to the patients to be tested first. At the same time, in order to ensure that the recommended rehabilitation schemes can better target the actual conditions of the patients to be tested, the rehabilitation schemes of similar patients with greater symptom similarity need to be recommended to the patients to be tested first. Therefore, the scheme independence and symptom similarity of the target similar patients can be combined and normalized to obtain the priority recommendation coefficient of the target similar patients.

[0097] In the embodiment of the present invention, the combination of the scheme independence degree and the symptom similarity can be achieved by calculating the sum or product of the two, which is not limited here.

[0098] As an example, in one embodiment of the present invention, the expression of the priority recommendation coefficient of the target similar patient can be specifically, for example, as follows:

[0099]

[0100] in, Indicates the priority recommendation coefficient for patients with similar targets; represents the mean of the regimen similarity coefficients for all second patient groups; Indicates the degree of protocol independence for patients with similar goals; Indicates the symptom similarity of target-similar patients; Represents a normalization function, which is used to limit the calculation results to scope; Indicates the preset third adjustment parameter, which is used to prevent the denominator from being 0. The value range is In one embodiment of the present invention, Set to 0.01, The specific value of can also be set by the implementer according to the specific implementation scenario and is not limited here.

[0101] The same method as above can be used to obtain the priority recommendation coefficient of each similar patient to the patient to be tested. The larger the priority recommendation coefficient, the more suitable the rehabilitation plan of the similar patient is for the patient to be tested, and the more it should be recommended to the patient to be tested. Therefore, based on the priority recommendation coefficient of each similar patient, the recommended rehabilitation plan for the patient to be tested can be selected from the rehabilitation plans of all similar patients, thereby improving the accuracy of recommending traditional Chinese medicine rehabilitation plans for the patient to be tested.

[0102] Preferably, in one embodiment of the present invention, the method for obtaining the recommended rehabilitation plan for the patient to be tested specifically includes:

[0103] The priority recommendation coefficient of each similar patient is used as a parameter used by the ant colony algorithm, and the rehabilitation plans of all similar patients are input into the ant colony algorithm, which outputs the recommended rehabilitation plan for the patient to be tested. The ant colony algorithm is a technical means well known to those skilled in the art and will not be described in detail here.

[0104] In other embodiments of the present invention, rehabilitation plans of similar patients corresponding to a preset number of largest priority recommendation coefficients may also be selected as the recommended rehabilitation plan for the patient to be tested, which is not limited here.

[0105] An embodiment of the present invention provides a TCM rehabilitation program recommendation system, which includes a memory, a processor, and a computer program, wherein the memory is used to store the corresponding computer program, and the processor is used to run the corresponding computer program. When the computer program runs in the processor, it can implement the method described in steps S1 to S3.

[0106] In summary, the embodiment of the present invention first obtains the examination reports of the patient to be tested and each historical patient, and obtains the rehabilitation plan of each historical patient, performs natural language processing on each examination report and rehabilitation plan, and obtains the word vector of the phrases in each examination report and rehabilitation plan; obtains the symptom similarity of each historical patient based on the difference in the word vector of the phrases in the examination report between the patient to be tested and each historical patient; based on the symptom similarity, screens out similar patients of the patient to be tested from all historical patients; obtains the plan difference index of the patient to be tested based on the difference in the word vector of the phrases in the rehabilitation plan between each similar patient; if the plan difference index of the patient to be tested is less than a preset difference threshold, then selects the recommended rehabilitation plan for the patient to be tested from the rehabilitation plans of all similar patients based on the symptom similarity of each similar patient, and if the plan difference index of the patient to be tested is not less than the preset difference threshold, then any similar patient is taken as the target similar patient, and the priority recommendation coefficient of the target similar patient is obtained based on the difference in the word vector of the phrases in the rehabilitation plan between the target similar patient and other similar patients except the target similar patient, and the symptom similarity of the target similar patient, and based on the priority recommendation coefficient of each similar patient, selects the recommended rehabilitation plan for the patient to be tested from the rehabilitation plans of all similar patients.

[0107] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0108] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for recommending a traditional Chinese medicine rehabilitation program, characterized in that: The method comprises: Obtain the examination reports of the patient to be tested and each historical patient, and obtain the rehabilitation plan of each historical patient, perform natural language processing on each examination report and rehabilitation plan, and obtain the word vector of each phrase in the examination report and rehabilitation plan; According to the difference in the word vectors of the phrases in the examination reports between the patient to be tested and each historical patient, the symptom similarity of each historical patient is obtained; based on the symptom similarity, similar patients of the patient to be tested are screened out from all historical patients; according to the difference in the word vectors of the phrases in the rehabilitation plans between the similar patients, the plan difference index of the patient to be tested is obtained; If the scheme difference index of the patient to be tested is less than the preset difference threshold, then according to the symptom similarity of each similar patient, a recommended rehabilitation scheme for the patient to be tested is selected from the rehabilitation schemes of all similar patients; if the scheme difference index of the patient to be tested is not less than the preset difference threshold, then any similar patient is taken as the target similar patient, and according to the difference in word vectors of phrases in the rehabilitation schemes between the target similar patient and other similar patients except the target similar patient, and the symptom similarity of the target similar patient, a priority recommendation coefficient of the target similar patient is obtained, and based on the priority recommendation coefficient of each similar patient, a recommended rehabilitation scheme for the patient to be tested is selected from the rehabilitation schemes of all similar patients; The scheme difference index of the patient to be tested is obtained including: Taking any two similar patients as a first patient group, selecting any one phrase from the rehabilitation programs of the two similar patients in each first patient group, and taking the selected two phrases as a second phrase combination of each first patient group; Performing negative correlation mapping on the Euclidean distance between the word vectors of two word groups in each second word group combination to obtain a similarity parameter of each second word group combination, and normalizing the average value of the similarity parameters of all second word group combinations to obtain the scheme similarity of each first patient group; Obtaining a regimen difference index for the patient to be tested according to the distribution of the regimen similarities of all the first patient groups; The obtaining of the regimen difference index of the patient to be tested according to the distribution of the regimen similarities of all the first patient groups comprises: Performing negative correlation mapping on the average values ​​of the scheme similarities of all first patient groups to obtain a first difference parameter of the patient to be tested; Using the standard deviation of the similarity of the regimens of all the first patient groups as the second difference parameter of the patients to be tested; The first difference parameter and the second difference parameter are integrated and normalized to obtain a regimen difference index for the patient to be tested.

2. A method for recommending a TCM rehabilitation program according to claim 1, characterized in that: The obtaining of the symptom similarity of each historical patient comprises: Taking any historical patient as the target historical patient, selecting any phrase from the examination report of the patient to be tested and the examination report of the target historical patient, respectively, and combining the two selected phrases as a first phrase; The absolute value of the cosine similarity between the word vectors of the two word groups in each first word group combination is used as the similarity degree of each first word group combination; The average value of the similarity levels of all first word group combinations is taken as the symptom similarity of the target historical patient.

3. A method for recommending a TCM rehabilitation program according to claim 1, characterized in that: Based on the symptom similarity, similar patients to the patient to be tested are screened from all historical patients, including: According to the similarity of the symptoms of each historical patient, all historical patients are clustered to obtain multiple clusters; The average of the symptom similarities of all historical patients in each cluster is taken as the cluster center value of each cluster; The historical patients in the cluster corresponding to the maximum value of the cluster center value are taken as similar patients to the patient to be tested.

4. A method for recommending a TCM rehabilitation program according to claim 1, characterized in that: The selecting of a recommended rehabilitation program for the patient to be tested from the rehabilitation programs of all similar patients according to the symptom similarity of each similar patient includes: The rehabilitation plans of similar patients corresponding to a preset number of the greatest symptom similarities are selected as the recommended rehabilitation plans for the patient to be tested.

5. A method for recommending a TCM rehabilitation program according to claim 1, characterized in that: The priority recommendation coefficients for obtaining target similar patients include: The target similar patient and any other similar patient except the target similar patient are regarded as a second patient group, and the scheme similarity coefficient of each second patient group is obtained by using the calculation method of the scheme similarity of the first patient group; Performing negative correlation mapping on the average values ​​of the scheme similarity coefficients of all second patient groups to obtain the scheme independence degree of target similar patients; The regimen independence degree and symptom similarity of target similar patients are integrated and normalized to obtain the priority recommendation coefficient of target similar patients.

6. A method for recommending a TCM rehabilitation program according to claim 1, characterized in that: The step of selecting a recommended rehabilitation plan for the patient to be tested from the rehabilitation plans of all similar patients based on the priority recommendation coefficient of each similar patient includes: The priority recommendation coefficient of each similar patient is used as a parameter used by the ant colony algorithm, and the rehabilitation plans of all similar patients are input into the ant colony algorithm, and the ant colony algorithm outputs a recommended rehabilitation plan for the patient to be tested.

7. A method for recommending a TCM rehabilitation program according to claim 1, characterized in that: The step of obtaining the word vector of each word group in the examination report and the rehabilitation plan includes: The jieba word segmentation tool is used to segment each examination report and rehabilitation plan to obtain the phrases in each examination report and rehabilitation plan. The Word2Vec algorithm is used to process the phrases in the examination report and rehabilitation plan to obtain the word vectors of the phrases in each examination report and rehabilitation plan.

8. A TCM rehabilitation program recommendation system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

Images