Method for generating targeted medical-care scheme in combination with traditional Chinese medicine constitution of old people

Abstract

The invention relates to the technical field of medical care, in particular to a method for generating a targeted medical care scheme in combination with the traditional Chinese medicine constitution of old people. The method comprises the following steps: obtaining diagnosis results obtained by different diagnosis types of users with different physiques; according to the quantity distribution of different physiques with the same expression characteristics, determining the association degree of the diagnosis type and the physique; according to the type distribution of the medical care schemes corresponding to different samples, determining an adaptive proportion coefficient; determining the specific gravity evaluation of the medical care scheme by combining the characteristic extensibility and the adaptive specific gravity coefficient; determining the direct recommendation degree of each medical care scheme; determining a recommendation deviation in combination with the medical care schemes actually used by different samples and the recommended medical care schemes; and determining an inhibition coefficient according to the proportion evaluation and the recommendation deviation, and correcting the direct recommendation degree of the to-be-analyzed sample to the medical and nursing scheme according to the inhibition coefficient to obtain a corrected recommendation degree. According to the method, more accurate medical care scheme recommendation is realized, so that the obtained recommendation degree result is more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of health care, in particular to a medical care scheme generation method combined with old TCM constitution. BACKGROUND

[0002] Generally, with the increase of age, the human body will appear the phenomenon of weakened heart and lung function, decreased immunity, and declined digestive system, and many old people also suffer from chronic diseases and are in a sub-healthy state. It has a positive significance to help old people improve the biased constitution and maintain physical health by formulating a medical care scheme for health management for the old people.

[0003] In related technologies, medical care scheme recommendation analysis is directly based on a collaborative filtering algorithm. In this way, since many old people show similar performance characteristics (for example, both physical weakness and depression will show the characteristic of pale complexion) due to the decline of physical functions, many medical care schemes are for similar performance characteristics. The recommendation result for these similar performance characteristics will be high when the recommendation is made for the old individual, the recommendation degree of the medical care scheme is greatly affected, and the recommendation effect accuracy is low. SUMMARY

[0004] In order to solve the technical problems that the old individual's condition in related technologies makes the recommendation result for these similar performance characteristics high when the recommendation is made, the recommendation degree of the medical care scheme is greatly affected, and the recommendation effect accuracy is low, the present application provides a medical care scheme generation method combined with old TCM constitution, and the technical solution adopted is as follows: The present application provides a medical care scheme generation method combined with old TCM constitution, which comprises the following steps: Obtaining diagnosis results obtained by different constitution users for different diagnosis types, wherein the diagnosis results include different performance characteristics and medical care schemes; Determining the characteristic universality of each diagnosis type performance characteristic according to the quantity distribution of all different constitutions in the sample of each diagnosis type performance characteristic; and determining the correlation degree between the diagnosis type and the constitution by combining the characteristic universality of all performance characteristics in the same diagnosis type; Determining the adaptation proportionality coefficient of each medical care scheme for each diagnosis type performance characteristic according to the category distribution of the medical care scheme corresponding to different samples; and determining the proportionality evaluation of the medical care scheme by combining the characteristic universality of each diagnosis type performance characteristic and the adaptation proportionality coefficient of the medical care scheme corresponding to the diagnosis type performance characteristic; According to the correlation degree, the performance characteristic similarity between two samples, and the distribution of medical care programs adopted by each sample, the direct recommendation degree of each sample to each medical care program is determined; combined with the actual medical care program used by each sample and the direct recommendation degree value change of all other medical care programs, the recommendation deviation of the medical care program is determined; According to the proportion evaluation and the recommendation deviation, the inhibition coefficient of the medical care program is determined, and according to the inhibition coefficient, the direct recommendation degree of the sample to be analyzed to the medical care program is corrected to obtain a corrected recommendation degree.

[0005] Further, according to the number distribution of all different constitutions in the samples of each diagnostic type performance characteristic, the characteristic generality of each diagnostic type performance characteristic is determined, including: The information entropy of the number of all different constitutions in the samples of each diagnostic type performance characteristic is subjected to maximum-minimum value normalization processing as the characteristic generality.

[0006] Further, the correlation degree of the diagnostic type and the constitution is determined by combining the characteristic generality of all performance characteristics in the same diagnostic type, including: The sum value of the characteristic generality of all performance characteristics in the same diagnostic type is taken as the type analysis index of the corresponding diagnostic type; The type analysis index is subjected to negative correlation mapping and normalization processing as the correlation degree of the diagnostic type and the constitution.

[0007] Further, according to the category distribution of the medical care programs corresponding to different samples, the adaptation proportion coefficient of each medical care program to each diagnostic type performance characteristic is determined, and the corresponding calculation formula is: In the formula, represents the adaptation proportion coefficient of the a-th medical care program to the m-th performance characteristic of the k-th diagnostic type, represents the number of samples using the a-th medical care program in all samples of the m-th performance characteristic of the k-th diagnostic type, represents the total number of samples of the m-th performance characteristic of the k-th diagnostic type, represents the variance of the number of samples using each medical care program in all samples of the m-th performance characteristic of the k-th diagnostic type, represents the total number of medical care programs used in all samples showing the m-th symptom of the k-th aspect.

[0008] Further, combined with the characteristic generality of each diagnostic type performance characteristic and the adaptation proportion coefficient of the medical care program and the diagnostic type performance characteristic, the proportion evaluation of the medical care program is determined, including: The product of the fit ratio coefficient between the medical and elderly care program and the performance characteristics of each diagnostic type and the breadth of the corresponding characteristics is calculated as the influence coefficient between the medical and elderly care program and the performance characteristics of the corresponding diagnostic type. The sum of the influence coefficients of the medical and elderly care plan with all diagnostic types and all manifestations is normalized and used as the weight evaluation of the medical and elderly care plan.

[0009] Furthermore, based on the degree of association, the similarity of performance characteristics between two samples, and the distribution of the medical and elderly care plans adopted by each sample, the degree of direct recommendation of each sample for each medical and elderly care plan is determined, including: The sample similarity between two samples is determined based on the similarity of their different manifestations under the same diagnostic type and the degree of correlation. Based on collaborative filtering, the degree of direct recommendation of each sample to each medical and elderly care plan is determined according to sample similarity and the adoption distribution of medical and elderly care plans.

[0010] Furthermore, based on the similarity and degree of association between any two samples with different manifestations under the same diagnostic type, the sample similarity between the two samples is determined, including: Determine the presence of each performance feature in two samples. If one sample does not have the performance feature and the other sample has the same performance feature, record the corresponding performance feature as 1; otherwise, record it as 0. Use the obtained record value as the performance analysis value. The sum of the performance analysis values ​​of all manifestation features under the same diagnostic type is used as the diagnostic analysis index for the diagnostic type. The product of the correlation degree and the diagnostic analysis index is calculated as the difference index between two samples in the same diagnostic type. The sum of the difference indicators of two samples across all diagnostic types is negatively correlated, normalized, and used as the sample similarity between the two samples.

[0011] Furthermore, by combining the changes in the direct recommendation degree between each sample's actual medical and elderly care plan and all other medical and elderly care plans, the recommendation bias of the medical and elderly care plan is determined, and the corresponding calculation formula is as follows: In the formula: This represents the total number of all samples. This represents the direct recommendation degree of the r-th medical and elderly care option for the i-th sample. This indicates the adoption status of the i-th sample with respect to the r-th medical and elderly care plan; a value of 1 indicates that the plan was used, and a value of 0 indicates that it was not used. This represents the direct recommendation degree of the medical and elderly care plan actually adopted by the i-th sample. Represents a linear normalization function. This represents the ReLU function.

[0012] Further, the inhibition coefficient of the medical and elderly care plan is determined based on the weighted evaluation and the recommendation deviation, including: The product of the weighted evaluation and the recommended deviation is calculated and normalized to serve as the inhibition coefficient of the medical and elderly care plan.

[0013] Furthermore, based on the inhibition coefficient, the direct recommendation level of the analyzed sample to the medical and elderly care plan is corrected to obtain the corrected recommendation level, including: Calculate the ratio of the direct recommendation degree to the inhibition coefficient for any medical and elderly care plan in the sample to be analyzed, and normalize the maximum and minimum values ​​to obtain the corrected recommendation degree.

[0014] The present invention has the following beneficial effects: This invention analyzes the diagnostic results of different users to recommend healthcare plans by combining three dimensions: diagnostic type, manifestation characteristics, and physical condition. Based on the distribution of diagnostic type manifestation characteristics across different physical conditions, the correlation between diagnostic type and physical condition is determined. Then, through correlation and sample similarity analysis, a direct recommendation degree analysis is performed. The recommendation bias of different healthcare plans is determined based on the numerical changes in the direct recommendation degree, accurately representing the bias effect of direct recommendations. Next, the weighting of healthcare plans is determined by the distribution of healthcare plans and the breadth of diagnostic type manifestation characteristics, representing the versatility of the plans. Finally, recommendation suppression is performed by combining the recommendation bias and weighting evaluation values ​​to correct the direct recommendation degree, resulting in a corrected recommendation degree. This invention considers the relationship between elderly physical condition, manifestation characteristics, and healthcare plans from multiple dimensions, achieving more accurate healthcare plan recommendations. It provides systematic assistance in analyzing elderly physical condition and developing elderly healthcare plans, making the obtained recommendation results more accurate and better suited to the current individual elderly person's situation. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating a method for generating a targeted medical and health care program for the elderly based on their traditional Chinese medicine constitution, provided in one embodiment of the present invention; Figure 2 This is a schematic diagram of the direct recommendation degree determination process provided in one embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method for generating a targeted medical and health care program based on the constitution of the elderly according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, 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 pertains.

[0019] It should be noted that, in order to ensure that the calculation results are meaningful, when performing fractional operations, if the denominator is 0, a parameter adjustment factor greater than 0 needs to be added to the denominator to prevent the denominator from being 0. The value of the parameter adjustment factor shall be set by the implementer according to the actual situation, and this application does not impose any special restrictions.

[0020] The following description, in conjunction with the accompanying drawings, details a specific method for generating targeted medical and health care programs for the elderly based on traditional Chinese medicine constitution, provided by this invention.

[0021] Please see Figure 1 The diagram illustrates a flowchart of a method for generating a targeted medical and health care program for the elderly based on their traditional Chinese medicine constitution, according to an embodiment of the present invention. The method includes: S101: Obtain diagnostic results for users with different physical conditions and different diagnostic types. The diagnostic results include different manifestations and medical and health care plans.

[0022] Among them, TCM constitution is an important component of the basic theory of TCM. Constitutions can be divided into nine types: balanced constitution, qi deficiency constitution, yang deficiency constitution, yin deficiency constitution, phlegm-dampness constitution, damp-heat constitution, blood stasis constitution, qi stagnation constitution, and special constitution. Of course, in other embodiments of this invention, different constitution classifications can also be implemented, and this is not a limitation.

[0023] Among them, diagnostic types mainly refer to diagnostic methods within the TCM system. Different diagnostic methods yield different manifestations. For example, diagnostic types can be specifically divided into four types: observation, auscultation and olfaction, inquiry, and palpation. Regarding pulse palpation, the manifestations of pulse palpation can specifically include various pulse characteristics such as floating pulse, deep pulse, slow pulse, and rapid pulse.

[0024] Among them, medical and health care programs usually include dietary conditioning, traditional Chinese medicine conditioning, exercise conditioning, emotional management, sleep management, environmental adjustment, etc. Taking the conditioning of Yin deficiency and Yang deficiency as examples, the Yin deficiency type usually uses medicines such as wolfberry and ophiopogon japonicus to nourish Yin, while the Yang deficiency type uses medicines such as cinnamon twig and dried ginger to warm Yang.

[0025] It should be noted that this plan aims to provide simple medical and elderly care plan recommendations based on diagnostic results, that is, mainly for the purpose of recommending the effectiveness of the plan, and is mainly used to assist experienced doctors in making plan recommendations, without involving the specific disease diagnosis process.

[0026] A person's constitution may become biased towards a particular constitution due to certain factors, manifesting as an abnormality or imbalance in a certain aspect; this is called a biased constitution. Typically, with age, people experience weakened cardiopulmonary function, decreased immunity, and digestive system decline. Many elderly people also suffer from chronic diseases and are in a sub-healthy state. Developing medical and elderly care plans and providing health management for the elderly is of positive significance in helping them improve their biased constitution and maintain their health.

[0027] Therefore, one implementation scenario of the present invention is to assist the elderly in the analysis of TCM constitution-based medical and health care programs. This is mainly achieved by analyzing the proportion of TCM health care methods for groups with the same symptoms and the general applicability of these methods. The analysis aims to determine whether all medical and health care programs are targeted at similar symptoms and to suppress recommendations for medical and health care programs that target similar symptoms.

[0028] In this embodiment of the invention, the diagnostic results can be specifically obtained by taking all the manifestation features as two-component inputs, taking 1 if there is a corresponding manifestation feature, and taking 0 if there is no corresponding manifestation feature.

[0029] It should be noted that the diagnosis result of each sample also corresponds to at least one medical and nursing care plan, so as to facilitate subsequent analysis based on all samples.

[0030] S102: Based on the distribution of the number of different constitutions in the samples of each diagnostic type's manifestation characteristics, determine the feature breadth of each diagnostic type's manifestation characteristics; combine the feature breadth of all manifestation characteristics in the same diagnostic type to determine the degree of correlation between the diagnostic type and the constitution.

[0031] It should be noted that there is a certain correlation between physical condition and diagnostic results. By analyzing this correlation, an effective auxiliary recommendation effect can be achieved. In this embodiment of the invention, the correlation analysis is mainly performed by analyzing the quantitative distribution of physical condition among samples exhibiting the same diagnostic type characteristics.

[0032] Furthermore, in some embodiments of the present invention, the feature breadth of each diagnostic type manifestation feature is determined based on the quantity distribution of all different constitutions in the samples of each diagnostic type manifestation feature, including: normalizing the information entropy of the quantity of all different constitutions in the samples of each diagnostic type manifestation feature to the maximum and minimum values, and using it as the feature breadth.

[0033] When samples exhibiting a certain trait all belong to the same constitution type, it indicates that the trait is distributed in extremes across different constitution types and has a strong correlation with constitution classification. Conversely, when samples exhibiting a certain trait are distributed across diverse constitution types, it indicates that the correlation between the trait and constitution classification is weak. Based on this logic, the degree of diversity in the distribution of each trait across different constitution types is denoted as the feature breadth.

[0034] Taking the m-th manifestation of the k-th diagnostic type as an example, we analyze the distribution of this manifestation in historical individual samples of different constitution types. The feature breadth of the m-th manifestation of the k-th diagnostic type is discussed. The calculation method is as follows: In the formula, This indicates the total number of body types. This represents the number of samples exhibiting the m-th phenotypic characteristic in the k-th aspect, representing the t-th physique. This represents the total number of samples exhibiting the m-th characteristic of the k-th aspect. Represents the logarithmic function with base 2. This represents the Sigmoid function, used here for value normalization.

[0035] in The information entropy represents the distribution of the m-th manifestation feature of the k-th diagnostic type. The calculation of information entropy is a well-known technique.

[0036] The larger the information entropy value, the more chaotic the distribution of the m-th manifestation of the k-th diagnostic type, that is, the greater its feature breadth. Conversely, the smaller the information entropy value, the more the m-th manifestation of the k-th diagnostic type is biased towards a certain constitution, and the smaller the corresponding feature breadth value.

[0037] Therefore, by combining the feature breadth of all manifestations in the same diagnostic type, the degree of correlation between the diagnostic type and the constitution is determined, including: using the sum of the feature breadth of all manifestations in the same diagnostic type as the type analysis index of the corresponding diagnostic type; performing negative correlation mapping and normalization on the type analysis index as the degree of correlation between the diagnostic type and the constitution.

[0038] It should be noted that the smaller the information entropy value, the more it is biased towards a certain constitution, that is, the higher the correlation with the constitution. Therefore, the type analysis index is directly calculated and negative correlation mapping is normalized to represent the degree of correlation between the diagnostic type and the constitution.

[0039] In this embodiment of the invention, a negative correlation indicates an inverse relationship between the independent and dependent variables, where a smaller independent variable corresponds to a larger dependent variable. The specific manifestation of the negative correlation is determined by the actual application, and this application does not impose any special limitations. That is, by performing negative correlation mapping and normalization on the type analysis indicators to obtain the degree of correlation, it is shown that the larger the value of the type analysis indicator, the smaller the value of the resulting degree of correlation. In this embodiment of the invention, the reciprocal of the type analysis indicator can be directly calculated and normalized to obtain the degree of correlation; alternatively, the negative value of the type analysis indicator can be calculated and normalized to obtain the degree of correlation. No limitation is imposed on this approach.

[0040] S103: Based on the distribution of different types of medical and elderly care solutions corresponding to different samples, determine the matching weight coefficient of each medical and elderly care solution for each diagnostic type's performance characteristics; combine the feature breadth of each diagnostic type's performance characteristics with the matching weight coefficient of the medical and elderly care solution to the diagnostic type's performance characteristics to determine the weight evaluation of the medical and elderly care solution.

[0041] Different medical and elderly care plans are not well-suited to different manifestations, and the suitability of medical and elderly care plans is also a relatively important item in the recommendation analysis. Therefore, it is necessary to conduct a specific analysis of the suitability.

[0042] Based on the distribution of different medical and elderly care solutions for different samples, the fit ratio coefficient of each medical and elderly care solution for each diagnostic type is determined. The corresponding calculation formula is as follows: In the formula, This represents the fitting ratio coefficient of the a-th type of medical and elderly care plan for the m-th manifestation of the k-th diagnostic type. This represents the number of samples using the a-th medical and elderly care plan among all samples exhibiting the m-th manifestation of the k-th diagnostic type. This represents the total number of samples for the m-th manifestation of the k-th diagnostic type. This represents the variance of the number of samples using each healthcare protocol across all samples for the m-th manifestation of the k-th diagnostic type. This represents the total number of healthcare programs used in all samples exhibiting the m-th symptom in the k-th aspect, and Sigmoid represents the Sigmoid function, used here for value normalization.

[0043] As can be seen from the above formula, the distribution of TCM health regimens among groups with similar phenotypic characteristics... And the proportion of medical and elderly care solutions among all medical and elderly care solutions, to obtain the matching ratio coefficient of medical and elderly care solutions.

[0044] By comparing all samples with the same performance characteristics, we can analyze the distribution ratio of medical and elderly care solutions corresponding to these samples with the same performance characteristics. The matching ratio coefficient of medical and elderly care solutions can reflect whether there is a situation where a certain cause leads to a large distribution of performance characteristics.

[0045] If, for a group exhibiting the same symptom, a certain medical and health care plan has a large proportion of participants, and the variety of such plans for that group is limited and unevenly distributed, then the recommendation level is more likely to be inflated due to the large proportion of such plans. For example, in a sample group where all participants have a pale complexion, blood and qi deficiency is the most common and prevalent condition. Therefore, medical and health care plans targeting blood and qi deficiency would have a higher recommendation level, while plans addressing other causes of paleness would have a relatively smaller proportion. A high value indicates that the corresponding recommendation may be too high, and its adaptation weight coefficient needs to be reduced.

[0046] Similarly, the variance of the sample size using each healthcare and elderly care plan A higher value indicates a more dispersed sample distribution and a lower suitability of the medical and elderly care plan, requiring a reduction in its suitability ratio coefficient. Therefore, the suitability ratio coefficient of the medical and elderly care plan is calculated using the above formula.

[0047] In other embodiments of the present invention, for each symptom group of each diagnostic type, only the fit ratio coefficient of the most frequently used medical and elderly care solution can be used. For medical and elderly care solutions that are not used most frequently, their fit ratio coefficient in the symptom group can be recorded as 0. Similarly, all fit ratio coefficients are obtained.

[0048] After determining the fit ratio coefficient, an evaluation analysis is required. Based on the fit ratio coefficient of each medical and elderly care plan for all groups with the same symptom characteristics, and combined with the feature breadth of each symptom, the weight evaluation of each medical and elderly care plan is obtained.

[0049] Furthermore, in some embodiments of the present invention, the weight evaluation of the medical and elderly care plan is determined by combining the feature breadth of each diagnostic type's performance characteristics and the matching weight coefficient between the medical and elderly care plan and the performance characteristics of the diagnostic type. This includes: calculating the product of the matching weight coefficient between the medical and elderly care plan and the corresponding feature breadth as the influence coefficient between the medical and elderly care plan and the corresponding diagnostic type's performance characteristics; and normalizing the sum of the influence coefficients between the medical and elderly care plan and all performance characteristics of all diagnostic types as the weight evaluation of the medical and elderly care plan.

[0050] Therefore, the calculation formula for weighted evaluation can specifically include: In the formula, This indicates the weighting of the r-th type of medical and elderly care option. This represents the total number of all diagnostic types included. This represents the total number of manifestation features included in the k-th diagnostic type. This represents the proportion coefficient of the r-th medical and elderly care option in the same symptom group for the m-th manifestation of the k-th diagnostic type. This represents the breadth of the m-th manifestation of the k-th diagnostic type. This represents the linear normalization function.

[0051] Since a higher feature breadth indicates a wider prevalence of a trait among elderly individuals with different physical conditions, it is more likely that this trait is a similar symptom resulting from age-related decline in bodily functions. Consequently, the most widely applied medical and elderly care solutions are more likely to target this similar trait. Therefore, feature breadth is used as a weight to increase the weight of the proportion coefficient of the corresponding group with the same trait. The product of the proportion coefficient and the corresponding feature breadth is calculated as the influence coefficient. Furthermore, the sum of the influence coefficients of the medical and elderly care plan with all diagnostic types and all manifestations was normalized and used as the weighting evaluation of the medical and elderly care plan.

[0052] S104: Based on the degree of association, the similarity of performance characteristics between two samples, and the distribution of the medical and elderly care plans adopted by each sample, determine the degree of direct recommendation of each sample to each medical and elderly care plan; combine the changes in the direct recommendation values ​​of each sample's actual medical and elderly care plan with all other medical and elderly care plans to determine the recommendation bias of the medical and elderly care plan.

[0053] It's important to note that a high proportion of medical and elderly care solutions may not only target similar symptoms in the elderly, but may also be widely used methods. For example, elderly people often have poor bodily regulation and experience symptoms like qi deficiency and fatigue, requiring adjustments to sleep and rest. Among groups exhibiting similar symptoms, medical and elderly care solutions focusing on sleep adjustment tend to have a high proportion, thus indicating a high recommendation rate for such solutions. Therefore, the inhibitory effect on medical and elderly care solutions cannot be determined solely by proportion evaluation. It's also necessary to analyze whether highly recommended solutions are actually being used. If a high recommendation rate is coupled with a significant number of non-adoptions, it suggests that the solution is not highly applicable but rather targets similar symptoms, requiring suppression of its recommendation rate. The specific inhibitory effect needs to be determined through analysis of recommendation bias.

[0054] Furthermore, in some embodiments of the present invention, see [link to relevant documentation]. Figure 2 ,Figure 2 This is a schematic diagram of a direct recommendation degree determination process provided in one embodiment of the present invention; based on the degree of association, the similarity of performance characteristics between two samples, and the distribution of medical and elderly care plans adopted by each sample, the direct recommendation degree of each sample for each medical and elderly care plan is determined, including: S201: Determine the sample similarity between two samples based on the similarity of their different manifestations under the same diagnostic type and the degree of correlation.

[0055] Among them, sample similarity can specifically represent the similarity between two samples. This similarity includes the similarity of performance features. This application adds a correlation degree parameter to adjust the sample similarity and avoid the influence of generalized performance features.

[0056] Based on the similarity and correlation of different manifestations of any two samples under the same diagnostic type, the sample similarity between the two samples is determined, including: determining the presence of each manifestation feature in the two samples; if one sample does not have a manifestation feature and the other sample has the same manifestation feature, the corresponding manifestation feature is recorded as 1, otherwise it is recorded as 0, and the obtained recorded value is used as the manifestation analysis value; the sum of the manifestation analysis values ​​of all manifestation features under the same diagnostic type is used as the diagnostic analysis index of the diagnostic type; the product of the correlation degree and the diagnostic analysis index is calculated as the difference index between the two samples in the same diagnostic type; the sum of the difference indices of the two samples across all diagnostic types is negatively correlated and normalized to be used as the sample similarity between the two samples.

[0057] In this case, when one sample does not have any performance characteristics and another sample has the same performance characteristics, the corresponding performance characteristics are recorded as 1. That is, the performance analysis value obtained represents the difference effect of performance characteristics between the two samples. The product of the correlation degree and the diagnostic analysis index is calculated as the difference index between the two samples in the same diagnostic type. That is, the larger the value of the difference index, the fewer the same performance characteristics are between the two samples, and the lower the correlation between the corresponding performance characteristics and the constitution, that is, the lower the sample similarity.

[0058] S202: Based on collaborative filtering, determine the degree of direct recommendation of each sample to each medical and elderly care plan according to sample similarity and the adoption distribution of medical and elderly care plans.

[0059] Then, the direct recommendation level can be calculated directly based on the collaborative filtering algorithm. The steps to generate medical and elderly care plans using the collaborative filtering algorithm are as follows: using the similarity between all sample individuals and the current individual as weights, calculate the weighted average of all sample individuals for each medical and elderly care plan, thereby obtaining the direct recommendation level of each medical and elderly care plan.

[0060] As people age, various bodily functions decline, which manifests in outward signs. For example, kidney failure often causes fatigue, gastrointestinal decline leads to bad breath, and cardiopulmonary decline results in shortness of breath. With age, these bodily functions gradually decline, leading many elderly individuals to exhibit similar symptoms. This results in many samples displaying similar symptoms across multiple diagnostic categories, causing a higher proportion of healthcare plans addressing these similar symptoms. Consequently, recommendations for these more prevalent plans may be overstated, reducing the specificity and accuracy of the overall healthcare recommendations.

[0061] For example, most elderly people suffer from qi and blood deficiency. In traditional Chinese medicine, this deficiency is often characterized by a pale complexion. As people age, the physiological functions of the spleen and kidneys gradually decline, making qi and blood deficiency very common among the elderly. Besides qi and blood deficiency, other causes of a pale complexion include heart failure, excessive fatigue, and side effects of certain medications. However, these other causes are much less common than qi and blood deficiency. Different causes of pale complexion require different treatment and care plans. Therefore, when recommending a treatment plan for an individual, even if the individual's pale complexion is not due to qi and blood deficiency, the recommended treatment plan for qi and blood deficiency may still be more strongly recommended due to the large number of cases with this condition.

[0062] The direct recommendation score of a healthcare plan is calculated directly using the collaborative filtering algorithm based on the sample similarity between different samples and the adoption distribution of healthcare plans. If the obtained recommendation score is high but the individual in that sample has not adopted that healthcare plan, it is considered that a recommendation bias has occurred. First, the direct recommendation score is obtained using the collaborative filtering algorithm based on the similarity between samples. Taking the r-th healthcare plan as an example, the adoption status of the r-th healthcare plan for each individual sample is considered unknown. The recommendation degree of the r-th healthcare plan for each individual sample is calculated using the collaborative filtering algorithm, denoted as the direct recommendation score. The direct recommendation score of the r-th healthcare plan for the i-th sample is... The calculation method is as follows: In the formula: This represents the total number of all samples. This represents the number of samples excluding the i-th sample itself. This represents the similarity between the i-th sample and the j-th sample. This indicates the adoption status of the j-th sample with respect to the r-th medical and elderly care plan. If the plan is used, the value is 1; otherwise, it is 0.

[0063] In this embodiment of the invention, the specific calculation of the direct recommendation degree is the calculation process of the collaborative filtering algorithm. The collaborative filtering algorithm is an algorithm well known to those skilled in the art, and will not be further limited or described.

[0064] The higher the direct recommendation rate, but the less likely the healthcare plan is to be adopted, the more likely it is to be a healthcare plan targeting a particular performance characteristic that accounts for a large proportion. Therefore, a recommendation bias has occurred, and the recommendation results of the healthcare plan need to be suppressed.

[0065] Recommendation bias reflects situations where the recommended outcomes for each healthcare and elderly care plan are too high, but it is not actually adopted. The greater the recommendation bias, the more likely the healthcare and elderly care plan is to be a method for similar performance characteristics rather than a general approach. The more cases of over-recommendation there are, the more necessary it is to suppress the recommendation results of that method.

[0066] By combining the changes in the direct recommendation strength of each sample's actual medical and elderly care plan with all other medical and elderly care plans, the recommendation bias of the medical and elderly care plan is determined. The corresponding calculation formula is as follows: In the formula: This represents the total number of all samples. This represents the direct recommendation degree of the r-th medical and elderly care option for the i-th sample. This indicates the adoption status of the i-th sample with respect to the r-th medical and elderly care plan; a value of 1 indicates that the plan was used, and a value of 0 indicates that it was not used. This represents the direct recommendation degree of the medical and elderly care plan actually adopted by the i-th sample. Represents a linear normalization function. This represents the ReLU function.

[0067] When the ReLU function is used to numerically adjust the data, a large difference in the direct recommendation value between the r-th healthcare plan and the actually adopted plan indicates that the r-th healthcare plan has an overly high recommendation value, resulting in a significant recommendation bias. The value of is between [0,1], through Calculation, in Setting it to 1 indicates that the r-th healthcare plan is used, which corresponds to a smaller recommendation bias. It must be less than 0, then It is 0, while A value of 0 indicates that the r-th healthcare plan was not used, which corresponds to a significant recommendation bias. A larger value corresponds to a greater recommendation bias. Therefore, this application cleverly uses two ReLU functions to accurately calculate the recommendation bias.

[0068] S105: Determine the inhibition coefficient of the medical and elderly care plan based on the weighted evaluation and recommendation bias. Use the inhibition coefficient to correct the direct recommendation degree of the medical and elderly care plan for the sample to be analyzed, and obtain the corrected recommendation degree.

[0069] After determining the recommendation bias, the inhibition coefficient can be analyzed based on the weight evaluation and the recommendation bias. In this embodiment of the invention, the inhibition coefficient of the medical and elderly care plan is determined according to the weight evaluation and the recommendation bias, including: calculating the product of the weight evaluation and the recommendation bias, and normalizing it as the inhibition coefficient of the medical and elderly care plan.

[0070] Since a larger weighting evaluation value indicates that the healthcare plan is more likely to address multiple manifestations and has high versatility, it can lead to distorted recommendations, necessitating suppression. Conversely, a larger recommendation deviation value indicates a greater bias in the recommended healthcare plan compared to actual conditions, also requiring suppression. Therefore, this application calculates the product of the weighting evaluation and the recommendation deviation, and normalizes it to serve as the suppression coefficient for the healthcare plan.

[0071] Furthermore, in some embodiments of the present invention, the ratio of the direct recommendation degree to the inhibition coefficient of any medical and elderly care plan in the sample to be analyzed is calculated, and the maximum and minimum values ​​are normalized to obtain the corrected recommendation degree.

[0072] In other words, by using the inhibition coefficient to suppress the direct recommendation degree, a modified recommendation degree is obtained, which can more reliably characterize the recommendation effect of different medical and elderly care programs.

[0073] This invention analyzes the diagnostic results of different users to recommend healthcare plans by combining three dimensions: diagnostic type, manifestation characteristics, and physical condition. Based on the distribution of diagnostic type manifestation characteristics across different physical conditions, the correlation between diagnostic type and physical condition is determined. Then, through correlation and sample similarity analysis, a direct recommendation degree analysis is performed. The recommendation bias of different healthcare plans is determined based on the numerical changes in the direct recommendation degree, accurately representing the bias effect of direct recommendations. Next, the weighting of healthcare plans is determined by the distribution of healthcare plans and the breadth of diagnostic type manifestation characteristics, representing the versatility of the plans. Finally, recommendation suppression is performed by combining the recommendation bias and weighting evaluation values ​​to correct the direct recommendation degree, resulting in a corrected recommendation degree. This invention considers the relationship between elderly physical condition, manifestation characteristics, and healthcare plans from multiple dimensions, achieving more accurate healthcare plan recommendations. It provides systematic assistance in analyzing elderly physical condition and developing elderly healthcare plans, making the obtained recommendation results more accurate and better suited to the current individual elderly person's situation.

[0074] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0075] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for generating targeted medical and health care programs that combine traditional Chinese medicine (TCM) constitutions of the elderly, characterized in that, The method includes: Obtain diagnostic results for users with different physical conditions and different diagnostic types, including different manifestations and medical and health care plans; Based on the distribution of the number of different constitutions in the samples of each diagnostic type's manifestation characteristics, the feature breadth of each diagnostic type's manifestation characteristics is determined; combining the feature breadth of all manifestation characteristics in the same diagnostic type, the degree of correlation between the diagnostic type and the constitution is determined. Based on the distribution of different types of medical and elderly care solutions corresponding to different samples, determine the matching weight coefficient of each medical and elderly care solution for each diagnostic type's performance characteristics; combine the feature breadth of each diagnostic type's performance characteristics with the matching weight coefficient of the medical and elderly care solution to the diagnostic type's performance characteristics to determine the weight evaluation of the medical and elderly care solution. Based on the degree of correlation, the similarity of performance characteristics between two samples, and the distribution of the medical and elderly care plans adopted by each sample, the direct recommendation degree of each sample to each medical and elderly care plan is determined; combined with the changes in the direct recommendation degree values ​​of each sample's actual medical and elderly care plan to all other medical and elderly care plans, the recommendation bias of the medical and elderly care plan is determined. The inhibition coefficient of the medical and elderly care plan is determined based on the weighted evaluation and the recommendation deviation. The degree of direct recommendation of the medical and elderly care plan to the sample to be analyzed is corrected based on the inhibition coefficient, and the corrected recommendation degree is obtained.

2. The method for generating a targeted medical and health care program based on the constitution of elderly people according to claim 1, characterized in that, The determination of the feature breadth of each diagnostic type's manifestation characteristics based on the quantitative distribution of all different constitutions in the samples of each diagnostic type's manifestation characteristics includes: The information entropy of the number of different constitutions in the samples that demonstrate the characteristics of each diagnostic type is normalized by the maximum and minimum values, and this is used as the feature breadth.

3. The method for generating a targeted medical and health care program based on the constitution of elderly people according to claim 1, characterized in that, By combining the breadth of features across all manifestations within the same diagnostic type, the degree of association between the diagnostic type and the constitution is determined, including: The sum of the feature breadth of all manifestation features in the same diagnostic type is used as the type analysis index for the corresponding diagnostic type. The type analysis indicators were negatively correlated and normalized to represent the degree of correlation between diagnostic type and constitution.

4. The method for generating a targeted medical and health care program based on the constitution of elderly people according to claim 1, characterized in that, Based on the distribution of different medical and elderly care solutions for different samples, the fit ratio coefficient of each medical and elderly care solution for each diagnostic type is determined. The corresponding calculation formula is as follows: In the formula, This represents the fitting ratio coefficient of the a-th type of medical and elderly care plan for the m-th manifestation of the k-th diagnostic type. This represents the number of samples using the a-th medical and elderly care plan among all samples exhibiting the m-th manifestation of the k-th diagnostic type. This represents the total number of samples for the m-th manifestation of the k-th diagnostic type. This represents the variance of the number of samples using each healthcare protocol across all samples for the m-th manifestation of the k-th diagnostic type. This represents the total number of medical and nursing care programs used in all samples exhibiting the m-th symptom of the k-th aspect.

5. The method for generating a targeted medical and health care program based on the constitution of elderly people according to claim 1, characterized in that, By combining the breadth of features of each diagnostic type and the matching weight coefficient between the medical and elderly care plan and the corresponding diagnostic type features, the weight evaluation of the medical and elderly care plan is determined, including: The product of the fit ratio coefficient between the medical and elderly care program and the performance characteristics of each diagnostic type and the breadth of the corresponding characteristics is calculated as the influence coefficient between the medical and elderly care program and the performance characteristics of the corresponding diagnostic type. The sum of the influence coefficients of the medical and elderly care plan with all diagnostic types and all manifestations is normalized and used as the weight evaluation of the medical and elderly care plan.

6. The method for generating a targeted medical and health care program based on the constitution of elderly people according to claim 1, characterized in that, Based on the degree of correlation, the similarity of performance characteristics between two samples, and the distribution of the medical and elderly care plans adopted by each sample, the degree of direct recommendation of each sample for each medical and elderly care plan is determined, including: The sample similarity between two samples is determined based on the similarity of their different manifestations under the same diagnostic type and the degree of correlation. Based on collaborative filtering, the degree of direct recommendation of each sample to each medical and elderly care plan is determined according to sample similarity and the adoption distribution of medical and elderly care plans.

7. The method for generating a targeted medical and health care program based on the constitution of elderly people according to claim 6, characterized in that, Based on the similarity and degree of association between any two samples with different manifestations under the same diagnostic type, the sample similarity between the two samples is determined, including: Determine the presence of each performance feature in two samples. If one sample does not have the performance feature and the other sample has the same performance feature, record the corresponding performance feature as 1; otherwise, record it as 0. Use the obtained record value as the performance analysis value. The sum of the performance analysis values ​​of all manifestation features under the same diagnostic type is used as the diagnostic analysis index for the diagnostic type. The product of the correlation degree and the diagnostic analysis index is calculated as the difference index between two samples in the same diagnostic type. The sum of the difference indicators of two samples across all diagnostic types is negatively correlated, normalized, and used as the sample similarity between the two samples.

8. The method for generating a targeted medical and health care program based on the constitution of elderly people according to claim 1, characterized in that, By combining the changes in the direct recommendation strength of each sample's actual medical and elderly care plan with all other medical and elderly care plans, the recommendation bias of the medical and elderly care plan is determined. The corresponding calculation formula is as follows: In the formula: This represents the total number of all samples. This represents the direct recommendation degree of the r-th medical and elderly care option for the i-th sample. This indicates the adoption status of the i-th sample with respect to the r-th medical and elderly care plan; a value of 1 indicates that the plan was used, and a value of 0 indicates that it was not used. This represents the direct recommendation degree of the medical and elderly care plan actually adopted by the i-th sample. Represents a linear normalization function. This represents the ReLU function.

9. The method for generating a targeted medical and health care program based on the constitution of elderly people according to claim 1, characterized in that, The inhibition coefficient of the medical and elderly care plan is determined based on the weighted evaluation and the recommended deviation, including: The product of the weighted evaluation and the recommended deviation is calculated and normalized to serve as the inhibition coefficient of the medical and elderly care plan.

10. The method for generating a targeted medical and health care program based on the constitution of elderly people according to claim 1, characterized in that, Based on the inhibition coefficient, the direct recommendation level of the medical and elderly care plan for the sample to be analyzed is corrected to obtain the corrected recommendation level, including: Calculate the ratio of the direct recommendation degree to the inhibition coefficient for any medical and elderly care plan in the sample to be analyzed, and normalize the maximum and minimum values ​​to obtain the corrected recommendation degree.

Images