Long-term data tracking and analysis method and platform for patients with hepatolenticular degeneration

Abstract

The present invention discloses a long-term data tracking and analysis method and platform for patients with Wilson's disease, relates to the technical field of data recommendation, and is used to solve the problem that current dietary recommendation plans lack personalization and refinement, and usually simply make recommendations based on a copper food exchange table and a copper food exchange map, and often fail to consider the dietary needs of different patients; a two-level matching strategy and a data-driven clustering analysis method are adopted to find a user group that is most similar to the initial state of a new user; a cluster analysis is performed on the dietary habits of people with a reasonable copper intake, different dietary pattern groups are divided, and the main characteristics of each group are identified, and the recommendation priority is determined in combination with the group proportion distribution; the technical means integrates data analysis methods such as similarity matching, cluster analysis and feature extraction, so that the recommendation system can finely process user data and realize personalized and scientific dietary plan recommendations.

Description

Technical Field

[0001] The present invention relates to the technical field of data recommendation, and more specifically, to a method and platform for long-term data tracking and analysis of patients with Wilson's disease. Background Art

[0002] Wilson's disease is a rare genetic disease characterized by disordered copper metabolism, which leads to the deposition of copper in organs such as the liver and brain. Dietary control is one of the key strategies for managing the disease, and reasonable control of dietary copper intake can help alleviate the patient's condition. However, due to individual differences, each patient's living habits, dietary needs, and economic conditions are different, so a single diet plan is difficult to meet the needs of all patients. To this end, a technical solution is needed that can combine the copper food exchange table and the copper food exchange map for personalized recommendations, thereby improving the effectiveness and compliance of dietary management.

[0003] The prior art has the following deficiencies:

[0004] However, the current dietary recommendation plans lack personalization and refinement. They usually simply make recommendations based on the copper food exchange table and copper food exchange map, and often fail to take into account the dietary needs of different patients.

[0005] In view of the above problems, the present invention proposes a solution. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a method and platform for long-term data tracking and analysis of patients with Wilson's disease, so as to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] The long-term data tracking and analysis method and platform for patients with hepatolenticular degeneration include the following steps:

[0009] Quantize and standardize the initial state features and eating habit features of all users;

[0010] Screen out the user groups with reasonable copper intake, and perform similarity matching with the people with reasonable copper intake based on the initial status characteristics of new users to determine similar people with reasonable copper intake;

[0011] Obtain the dietary habit characteristics of users with similar copper intake and determine dietary recommendations for new users based on them;

[0012] Among them, the initial state characteristics are the basic physiological health status of users when they first use the platform;

[0013] Eating habits are characterized by the types of food that users often eat and their frequency or intake.

[0014] In a preferred embodiment, the feature vectorization method is as follows:

[0015] Quantize the initial state and eating habit features of each user. Assume that the number of initial state features is m and the number of eating habit features is n. Then, for each user i, two vectors can be constructed:

[0016] Initial state vector: ai=(ai1,ai2,...,aim);

[0017] Eating habit vector: di=(di1,di2,...,din);

[0018] The matching method between new users and people with appropriate copper intake can be as follows:

[0019] Assume that the initial state vector of the new user is y=(y1,y2,...,ym);

[0020] Use cosine similarity to calculate the initial state similarity between new users and users with reasonable copper intake:

[0021] The initial state similarity is compared with the preset initial state similarity threshold, and the corresponding reasonable copper intake users whose similarity threshold is greater than the initial state similarity threshold are regarded as similar reasonable copper intake users of the new user. Based on the dietary habit characteristics of similar reasonable copper intake users, the food preferences of the new user in the copper food exchange table and copper food exchange map are determined to make subsequent recipe recommendations.

[0022] In a preferred embodiment, the dietary habit characteristics of users with reasonable copper intake on the platform are obtained, and groups are divided according to the similarity of the dietary habit characteristics of users with reasonable copper intake, and cluster groups with different dietary habit characteristics are identified, and the most representative main characteristics within the clustered groups are analyzed;

[0023] New users are matched with people with reasonable copper intake to determine similar users with reasonable copper intake. The spectrum recommendation strategy is then determined comprehensively based on the groups to which similar users with reasonable copper intake belong and the main characteristics of the groups.

[0024] In a preferred embodiment, the groups are divided according to the similarity of the dietary habit characteristics of users with reasonable copper intake, and the cluster groups with different dietary habit characteristics are identified by the K-means clustering method; including:

[0025] Initialize the cluster center and randomly select K cluster centers. Each center ck is an n-dimensional vector:

[0026] ck=(ck1,ck2,...,ckn);where k represents the kth cluster center, k=1,2,...,K;

[0027] ckj represents the value of the k-th cluster center on the j-th feature, which is randomly selected initially;

[0028] Calculate the distance and assign samples. For each user di, calculate the Euclidean distance between it and all K cluster centers, and assign the user to the cluster with the nearest center. The formula for the Euclidean distance is: ; Where Dik represents the distance between the i-th user and the k-th cluster center;

[0029] Assign user di to the cluster Ck that minimizes the distance, and calculate the assignment as: ; In the formula is the cluster label of user di, that is, assigned to the nearest cluster center Ck, Indicates finding the cluster center k with the minimum distance;

[0030] Update the cluster center. When all users with reasonable copper intake are assigned to their respective cluster centers, calculate the mean vector of each cluster to update the cluster center. The calculation of the updated cluster center ck is as follows: ; |Ck| represents the number of users currently belonging to cluster Ck;

[0031] If the change in the updated cluster center is less than the set threshold ϵ compared to the previous center, the algorithm ends.

[0032] In a preferred embodiment, combining the group of users with similar copper intake and the main characteristics of the group, comprehensively determining the spectrum recommendation strategy mainly includes the following steps:

[0033] Calculate and obtain the group proportion distribution value pk of similar users with reasonable copper intake corresponding to the new user, obtain the second state vector ab of the new user, where the second state vector ab includes the main feature set corresponding to the group, and calculate the feature matching value Mk of the second state vector ab of the new user and the main feature set;

[0034] The group proportion distribution value and the feature matching value are weightedly calculated to determine the recommendation priority coefficient Sk for each group. The recommendation priority coefficients Sk are sorted from high to low, and the basic group for recommending recipes for the group with the highest value is selected. A recipe recommendation strategy is formulated based on the dietary habits of similar users with reasonable copper intake within the basic group.

[0035] In a preferred embodiment, the population proportion distribution value pk can be specifically calculated as follows:

[0036] For each reasonable copper intake user di in the set of similar reasonable copper intake users P corresponding to the new user, determine the group Ck to which it belongs, and count the distribution of these users in each group; that is:

[0037] Count the number nk of the kth group in the set P: ;

[0038] When user di belongs to group Ck , otherwise 0;

[0039] Calculate the proportion pk of each group in the set P, pk=nk / |P|; where |P| is the number of users with similar reasonable copper intake in the set P.

[0040] In a preferred embodiment, the steps for calculating the matching degree of the main feature set may be as follows:

[0041] For the main feature set tk={tk1,tk2,...,tkpk} of group Ck, the matching degree Mk between the new user and this set is calculated as follows: ; In the formula, when the new user has the same main feature value as the group on the jth categorical or continuous feature , otherwise 0.

[0042] The long-term data tracking and analysis platform for patients with hepatolenticular degeneration includes a data acquisition module, a feature matching module, a recipe recommendation module, and a data storage module, and the signal connections between the modules;

[0043] The data acquisition module is used to acquire the initial state characteristics, eating habit characteristics and second state characteristics of each user, and perform data preprocessing;

[0044] The feature matching module is used to perform similarity matching between the initial state of the new user and the states of these users, and find the user group that is most similar to the initial state of the new user;

[0045] The recipe recommendation module is used to perform cluster analysis on the dietary habits of people with a reasonable copper intake, divide them into different dietary pattern groups, identify the main characteristics of each group, and determine the recipe recommendation strategy based on the proportion distribution of the groups;

[0046] The data storage module is used to store all data during the platform processing process.

[0047] Technical effects and advantages of the long-term data tracking and analysis method and platform for hepatolenticular degeneration patients of the present invention:

[0048] The present invention adopts a two-level matching strategy and a data-driven cluster analysis method to achieve personalized dietary recommendations. First, by screening user data with reasonable copper intake, and similarity matching the initial state of new users with the state of these users, find the user group that is most similar to the initial state of the new user. Then, cluster analysis is performed on the dietary habits of people with reasonable copper intake, different dietary pattern groups are divided, and the main characteristics of each group are identified. After extracting the main features using statistical methods such as frequency analysis and variance analysis, the system calculates the comprehensive recommendation score of each group based on the matching degree between the new user and the main characteristics of each group, combined with the proportion distribution of the group, so as to determine the recommendation priority. This technical means combines data analysis methods such as similarity matching, cluster analysis and feature extraction, so that the recommendation system can process user data in a refined manner and achieve personalized and scientific dietary plan recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a flow chart of the first method for long-term data tracking and analysis of patients with Wilson's disease in the present invention;

[0050] Figure 2 This is a flow chart of the second method for long-term data tracking and analysis of patients with Wilson's disease according to the present invention. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] The present invention adopts a two-level matching strategy and a data-driven cluster analysis method to achieve personalized dietary recommendations. First, by screening user data with reasonable copper intake, and similarity matching the initial state of new users with the states of these users, find the user group that is most similar to the initial state of new users. Then, cluster analysis is performed on the dietary habits of people with reasonable copper intake, different dietary pattern groups are divided, and the main characteristics of each group are identified. After extracting the main features using statistical methods such as frequency analysis and variance analysis, the system calculates the comprehensive recommendation score of each group based on the matching degree between the new user and the main characteristics of each group, combined with the proportion distribution of the group, to determine the recommendation priority. This technical means combines data analysis methods such as similarity matching, cluster analysis and feature extraction, so that the recommendation system can finely process user data and achieve personalized and scientific dietary plan recommendations.

[0053] Embodiment 1, as Figure 1As shown, the long-term data tracking and analysis method for Wilson's disease patients of the present invention comprises the following steps:

[0054] Quantize the initial status features and eating habits features of all users to ensure the uniformity of data format and perform standardized processing;

[0055] Screen out the user groups with reasonable copper intake, and perform similarity matching with the people with reasonable copper intake based on the initial status characteristics of new users to determine similar people with reasonable copper intake;

[0056] Obtain the dietary habit characteristics of users with similar copper intake and use them to determine dietary recommendations for new users.

[0057] The initial status characteristics are the basic physiological health status of the user when using the platform for the first time; for example:

[0058] Age: Users of different ages may have different tolerance and ability to metabolize copper;

[0059] Gender: Gender may affect copper metabolism due to physiological differences;

[0060] Physical condition: such as weight, BMI index, basal metabolic rate, etc., which can reflect the user's nutritional needs and metabolic characteristics;

[0061] Health status: including liver function indicators (such as ALT and AST levels) and copper metabolism-related indicators (such as serum copper, urine copper, etc.).

[0062] The dietary habit characteristics are the types of food that the user often eats and their frequency or intake. It should be noted that, in this embodiment, the foods in the copper food exchange table and the copper food exchange map are used as feature items, that is, the user's food preferences in the copper food exchange table and the copper food exchange map are recorded.

[0063] By analyzing these eating habits characteristics of different users, they can be assigned to groups with similar eating preferences so that healthy eating plans that suit their eating habits can be more accurately matched in subsequent recommendations.

[0064] Specific:

[0065] The feature vectorization method is as follows:

[0066] Quantize the initial state and eating habit features of each user. Assuming the number of initial state features is m and the number of eating habit features is n, for each user i, two vectors can be constructed:

[0067] Initial state vector: ai=(ai1,ai2,...,aim);

[0068] Eating habit vector: di=(di1,di2,...,din).

[0069] It should be noted that aij represents the value of the i-th user on the j-th feature. The subsequent expressions are similar and will not be repeated here.

[0070] The eigenvalues ​​are normalized or standardized to reduce the impact of scale differences. This process is a conventional method in the prior art and will not be described in detail here.

[0071] The method of screening the population with reasonable copper intake can be as follows:

[0072] Set a reasonable copper intake range [L,U]. Filter out the user group S that meets the following conditions:

[0073] S={i|L≤Ci≤U}; where Ci is the average copper intake of user i.

[0074] The matching method between new users and people with appropriate copper intake can be as follows:

[0075] Assume that the initial state vector of the new user is y=(y1,y2,...,ym);

[0076] Use cosine similarity to calculate the initial state similarity between new users and users with reasonable copper intake:

[0077] ;

[0078] Where j is the feature index, which ranges from 1 to m and traverses all features, so that the cosine similarity formula can calculate the similarity between multi-dimensional feature vectors.

[0079] The initial state similarity is compared with the preset initial state similarity threshold, and the corresponding copper intake reasonable users whose copper intake is greater than the initial state similarity threshold are regarded as similar copper intake reasonable users of the new user. Thus, according to the dietary habit characteristics of similar copper intake reasonable users, the food preferences of the new user in the copper food exchange table and copper food exchange map are determined, and subsequent recipe recommendations are made.

[0080] Example 2, Example 1 of the present invention introduces that the platform screens user data with reasonable copper intake, and matches the initial state of new users with the states of these users for similarity, finds the user group with the most similar initial state of the new user, and determines the recipe recommendation strategy for the new user based on the most similar user group. However, sometimes there may be multiple similar user groups, and at this time there may be multiple different user dietary habits and preferences, resulting in certain deficiencies in the subsequent recipe recommendation strategy. Based on this problem, Figure 2 As shown, this embodiment 2 introduces a recipe recommendation improvement method, which specifically includes the following steps:

[0081] Obtain the dietary habit characteristics of users with reasonable copper intake on the platform, divide the groups according to the similarity of the dietary habit characteristics of users with reasonable copper intake, identify cluster groups with different dietary habit characteristics, and analyze the most representative main characteristics within the clustered groups;

[0082] New users are matched with people with reasonable copper intake to determine similar users with reasonable copper intake. The spectrum recommendation strategy is then determined comprehensively based on the groups to which similar users with reasonable copper intake belong and the main characteristics of the groups.

[0083] Specifically, the groups are divided according to the similarity of the dietary habits of users with reasonable copper intake. The clustering groups with different dietary habits can be determined by the K-means clustering method. K-means is an unsupervised learning algorithm used to divide data into K clusters. The goal of this algorithm is to maximize the similarity of samples within each cluster and minimize the similarity of samples between different clusters. The specific steps include initial center point selection, sample allocation, and center update until the convergence condition is reached. That is:

[0084] Get the user's eating habit vector: di=(di1,di2,...,din);

[0085] Initialize the cluster center and randomly select K cluster centers. Each center ck is an n-dimensional vector:

[0086] ck=(ck1,ck2,…,ckn);where k represents the kth cluster center, k=1,2,…,K.

[0087] ckj represents the value of the k-th cluster center on the j-th feature, which is initially randomly selected.

[0088] Calculate the distance and assign samples. For each user di, calculate the Euclidean distance between it and all K cluster centers, and assign the user to the cluster with the nearest center. The formula for Euclidean distance is: ; Where Dik represents the distance between the i-th user and the k-th cluster center;

[0089] Assign user di to the cluster Ck that minimizes the distance, and calculate the assignment as: ; In the formula is the cluster label of user di, that is, assigned to the nearest cluster center Ck, Indicates finding the cluster center k with the minimum distance.

[0090] Update the cluster center. When all users with reasonable copper intake are assigned to their respective cluster centers, calculate the mean vector of each cluster to update the cluster center. The calculation of the updated cluster center ck is as follows: ; |Ck| represents the number of users currently belonging to cluster Ck.

[0091] If the change in the updated cluster center is less than a certain threshold ϵ (i.e., the convergence condition) compared with the previous center, the algorithm ends.

[0092] Furthermore, for the clustered groups, analyzing the most representative main features within them includes the following steps:

[0093] For each group Ck, the categorical characteristics and continuous characteristics are distinguished, and the main characteristics are analyzed separately, including:

[0094] For categorical features, calculate the frequency of each category and select the category with the highest frequency as the main categorical feature;

[0095] For continuous features, calculate the feature variance. If the variance is less than the preset threshold, the feature is recorded as a continuous main feature.

[0096] The categorical main features and the continuous main features are combined to obtain the main feature set tk of each group.

[0097] It should be noted that categorical features include the user's region, cultural background, etc. Different regions affect the user's taste, thereby affecting the recipe recommendation. The cultural background can be specifically Han, Tibetan, etc. Different cultural backgrounds also affect the user's choice of food, thereby affecting the recipe recommendation.

[0098] Continuous features include user economic conditions, etc., which can be expressed as a specific income amount or annual family income and other numerical values, which are input by the user. Different economic conditions lead to different food choices by users, thus affecting recipe recommendations.

[0099] Meanwhile, it should be noted that calculating the feature variance of continuous features is an existing technology and will not be described in detail here.

[0100] Furthermore, in this embodiment, the profile recommendation strategy is comprehensively determined based on the group of users with similar reasonable copper intake and the main characteristics of the group, and mainly includes the following steps:

[0101] Calculate and obtain the group proportion distribution value pk of similar copper intake reasonable users corresponding to the new user, and obtain the second state vector ab of the new user, where the second state vector ab includes the main feature set (category features and continuous features) corresponding to the group, that is, including regional features, cultural features, economic conditions, etc. And calculate the feature matching value Mk of the second state vector ab of the new user and the main feature set;

[0102] The group proportion distribution value and the feature matching value are weightedly calculated to determine the recommendation priority coefficient Sk for each group, Sk=αpk+βMk. The recommendation priority coefficients Sk are sorted from high to low, and the basic group for recommending recipes for the group with the highest value is selected. A recipe recommendation strategy is formulated based on the dietary habits of users with similar copper intake in the basic group.

[0103] It should be noted that the population proportion distribution value pk can be calculated as follows:

[0104] For each reasonable copper intake user di in the set of similar reasonable copper intake users P corresponding to the new user, determine the group Ck to which it belongs, and count the distribution of these users in each group. That is:

[0105] Count the number nk of the kth group in the set P: ;

[0106] When user di belongs to group Ck , otherwise 0.

[0107] Calculate the proportion pk of each group in the set P, pk=nk / |P|. Where |P| is the number of users with similar copper intake in the set P.

[0108] Furthermore, the steps for calculating the matching degree of the main feature set can be as follows:

[0109] For the main feature set tk={tk1,tk2,...,tkpk} of group Ck, the matching degree Mk between the new user and this set is calculated as follows: ; In the formula, when the new user has the same main feature value as the group on the jth categorical or continuous feature , otherwise 0.

[0110] The present invention also discloses a long-term data tracking and analysis platform for patients with Wilson's disease to implement the method described in the above embodiment, including a data acquisition module, a feature matching module, a recipe recommendation module and a data storage module, and signal connections between the modules.

[0111] The data acquisition module is used to acquire the initial state characteristics, eating habit characteristics and second state characteristics of each user, and perform data preprocessing;

[0112] The feature matching module is used to perform similarity matching between the initial state of the new user and the states of these users, and find the user group that is most similar to the initial state of the new user;

[0113] The recipe recommendation module is used to perform cluster analysis on the dietary habits of people with a reasonable copper intake, divide them into different dietary pattern groups, identify the main characteristics of each group, and determine the recipe recommendation strategy based on the proportion distribution of the groups;

[0114] The data storage module is used to store all data during the platform processing process.

[0115] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0116] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0117] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0118] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0119] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0120] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. Long-term data tracking and analysis method for patients with Wilson's disease, It is characterized in that The steps include: Quantize and standardize the initial state features and eating habit features of all users; Screen out the user groups with reasonable copper intake, and perform similarity matching with the people with reasonable copper intake based on the initial status characteristics of new users to determine similar people with reasonable copper intake; Obtain the dietary habit characteristics of users with similar copper intake and determine dietary recommendations for new users based on them; Among them, the initial state characteristics are the basic physiological health status of users when they first use the platform; Eating habits are characterized by the types of food that users often eat and their frequency or intake; The feature vectorization method is as follows: Quantize the initial state and eating habit features of each user. Assume that the number of initial state features is m and the number of eating habit features is n. Then, for each user i, two vectors can be constructed: Initial state vector: ai=(ai1,ai2,...,aim); Eating habit vector: di=(di1,di2,...,din); The matching method between new users and people with appropriate copper intake can be as follows: Assume that the initial state vector of the new user is y=(y1,y2,...,ym); Use cosine similarity to calculate the initial state similarity between new users and users with reasonable copper intake: The initial state similarity is compared with a preset initial state similarity threshold, and the corresponding copper intake reasonable users whose copper intake is greater than the initial state similarity threshold are regarded as similar copper intake reasonable users of the new user, and based on the dietary habit characteristics of similar copper intake reasonable users, the food preference of the new user in the copper food exchange table and the copper food exchange map is determined to make subsequent recipe recommendations; A recipe recommendation improvement method is also provided, which specifically comprises the following steps: Obtain the dietary habit characteristics of users with reasonable copper intake on the platform, divide the groups according to the similarity of the dietary habit characteristics of users with reasonable copper intake, identify cluster groups with different dietary habit characteristics, and analyze the most representative main characteristics within the clustered groups; New users are matched with people with reasonable copper intake to determine similar users with reasonable copper intake. Then, the spectrum recommendation strategy is determined comprehensively based on the groups and main characteristics of similar users with reasonable copper intake. Combining the groups of users with similar copper intake and their main characteristics, the comprehensive determination of the spectrum recommendation strategy mainly includes the following steps: Calculate and obtain the group proportion distribution value pk of similar users with reasonable copper intake corresponding to the new user, obtain the second state vector ab of the new user, where the second state vector ab includes the main feature set corresponding to the group, and calculate the feature matching value Mk of the second state vector ab of the new user and the main feature set; The group proportion distribution value and the feature matching value are weightedly calculated to determine the recommendation priority coefficient Sk for each group. The recommendation priority coefficients Sk are sorted from high to low, and the basic group for recommending recipes for the group with the highest value is selected. A recipe recommendation strategy is formulated based on the dietary habits of similar users with reasonable copper intake within the basic group.

2. The long-term data tracking and analysis method for patients with hepatolenticular degeneration according to claim 1, Features ； The groups were divided according to the similarity of dietary habit characteristics of users with reasonable copper intake, and the cluster groups with different dietary habit characteristics were identified by K-means clustering method; include: Initialize the cluster center and randomly select K cluster centers. Each center ck is an n-dimensional vector: ck=(ck1,ck2,...,ckn);where k represents the kth cluster center, k=1,2,...,K; ckj represents the value of the k-th cluster center on the j-th feature, which is randomly selected initially; Calculate the distance and assign samples. For each user di, calculate the Euclidean distance between it and all K cluster centers, and assign the user to the cluster with the nearest center. The formula for the Euclidean distance is: ; Where Dik represents the distance between the i-th user and the k-th cluster center; Assign user di to the cluster Ck that minimizes the distance, and calculate the assignment as: ; In the formula is the cluster label of user di, that is, assigned to the nearest cluster center Ck, Indicates finding the cluster center k with the minimum distance; Update the cluster center. When all users with reasonable copper intake are assigned to their respective cluster centers, calculate the mean vector of each cluster to update the cluster center. The calculation of the updated cluster center ck is as follows: ; |Ck| represents the number of users currently belonging to cluster Ck; If the change in the updated cluster center is less than the set threshold ϵ compared to the previous center, the algorithm ends.

3. The long-term data tracking and analysis method for patients with Wilson's disease according to claim 1, Features: The population proportion distribution value pk can be calculated as follows: For each reasonable copper intake user di in the set of similar reasonable copper intake users P corresponding to the new user, determine the group Ck to which it belongs, and count the distribution of these users in each group; that is: Count the number nk of the kth group in the set P: ; When user di belongs to group Ck , otherwise 0; Calculate the proportion pk of each group in the set P, pk=nk / |P|; where |P| is the number of users with similar reasonable copper intake in the set P.

4. The long-term data tracking and analysis method for patients with Wilson's disease according to claim 1, Features: The steps for calculating the matching degree of the main feature set can be as follows: For the main feature set tk={tk1,tk2,...,tkpk} of group Ck, the matching degree Mk between the new user and this set is calculated as follows: ; In the formula, when the new user has the same main feature value as the group on the jth categorical or continuous feature , otherwise 0.

5. A long-term data tracking and analysis platform for patients with hepatolenticular degeneration, used to implement the long-term data tracking and analysis method for patients with hepatolenticular degeneration as described in any one of claims 1 to 4 above, Features: It includes a data acquisition module, a feature matching module, a recipe recommendation module and a data storage module, and the signal connections between the modules; The data acquisition module is used to acquire the initial state characteristics, eating habit characteristics and second state characteristics of each user, and perform data preprocessing; The feature matching module is used to perform similarity matching between the initial state of the new user and the states of these users, and find the user group that is most similar to the initial state of the new user; The recipe recommendation module is used to perform cluster analysis on the dietary habits of people with a reasonable copper intake, divide them into different dietary pattern groups, identify the main characteristics of each group, and determine the recipe recommendation strategy based on the proportion distribution of the groups; The data storage module is used to store all data during the platform processing process.

Images