A method for generating doctor's order information based on patient identification

Abstract

The present invention relates to the technical field of order information generation based on patient identification, and discloses a method for generating order information based on patient identification. This method establishes a multi-dimensional uncertainty model of health data, integrates multi-dimensional health data and eliminates scale differences. First, physiological data and medical history data are collected from the health records authorized by the patient, and the maximum-minimum normalization is used to eliminate the dimensional differences of the data, so that all data are within the same range, avoiding the deviation caused by unit and range differences. Then, the fluctuation amplitude and correlation of the health data are calculated, the health indicators with large fluctuations are identified and their mutual influences are clarified, providing a basis for uncertainty propagation analysis. Finally, by assigning weights to each item of health data and introducing a regulation factor, the influence of the indicators is dynamically adjusted to adapt to the health conditions of different patients, thereby improving the personalization and accuracy of order optimization.

Description

Technical Field

[0001] The present invention relates to the technical field of order information generation based on patient identification, and specifically provides a method for generating order information based on patient identification. Background Art

[0002] With the continuous development of medical technology, patient health management has gradually developed towards personalization and precision, especially in the generation and adjustment of medical orders. Through in-depth analysis of patient health data, doctors can be assisted in providing more personalized treatment plans for patients. However, existing medical solutions generally face multiple problems, including incomplete data, excessive data uncertainty, inaccurate health risk prediction, etc., resulting in significant errors in the generation and adjustment of medical orders.

[0003] Most existing order generation technologies rely on traditional medical record data and simple medical rules, lacking comprehensive modeling and uncertainty analysis of multi-dimensional features of patient health data. Traditional methods usually rely on historical medical data and doctor experience to formulate medical orders, but this approach cannot adapt to the real-time changes and individual differences in the health status of patients, and cannot effectively handle the noise and uncertainty in health data. To improve the accuracy of medical orders, some existing technologies have introduced rule-based decision-making systems to process medical record data in combination with certain algorithms, but these methods still rely on fixed rules, lack flexibility, and cannot adaptively adjust medical orders. At the same time, existing technologies also appear to be relatively rough in modeling the multi-dimensional features of patient health data and cannot comprehensively capture the complexity and uncertainty of patient health changes.

[0004] Therefore, this case proposes a method for generating order information based on patient identification, which generates more personalized, accurate, and safe order plans through more precise health data analysis, uncertainty modeling, and dynamic optimization algorithms. Summary of the Invention

[0005] The present invention provides a method for generating order information based on patient identification, which helps to solve the problems mentioned in the above background art.

[0006] The present invention provides the following technical solution: A method for generating order information based on patient identification, comprising:

[0007] S1. Establish a multi-dimensional uncertainty model of health data:

[0008] Set a health index data set, denoted as X;

[0009] Collect multiple indicators from the health records authorized by the patient;

[0010] The multiple indicators include physiological data and medical history data. The collected data is added to the health indicator dataset to obtain X = [x1, x2,..., x n , where n is the total number of dimensions of the health data, and x i is the i-th health indicator data in the health indicator dataset;

[0011] Perform min-max normalization on each health indicator data:

[0012]

[0013] where min(x i ) and max(x i ) are the minimum and maximum values of the i-th health indicator data in the health indicator dataset respectively; x' i is the value after min-max normalization;

[0014] Update the values after min-max normalization of each health indicator data to the corresponding positions in the health indicator dataset;

[0015] Calculate the fluctuation range of each data in the health indicator dataset;

[0016] Calculate the correlation coefficient between the i-th and j-th data in the health indicator dataset;

[0017] Obtain the mean of the dataset X, denoted as μ;

[0018] Set a weight for each health data in the health indicator dataset. The calculation method is:

[0019]

[0020] where k is a regulation factor; ω i is the weighted coefficient of the i-th health indicator data;

[0021] S2. Uncertainty propagation analysis;

[0022] S3. Optimization generation of medical order uncertainty.

[0023] Optionally, the calculation of the fluctuation range of each data in the health indicator dataset specifically includes:

[0024] Obtain the variance of the i-th health indicator data in the health indicator dataset after min-max normalization, denoted as Var(x' i );

[0025] Obtain the mean of the i-th health indicator data in the health indicator dataset after min-max normalization, denoted as E(x' i );

[0026] Estimate the fluctuation range for each data in the health indicator dataset. The formula for calculating the fluctuation range is:

[0027] δ i = σ i ·x' i ;

[0028] Wherein, δ i is the fluctuation range of the i-th health indicator data;

[0029] Set the fluctuation range dataset, denoted as Δ;

[0030] Add all the calculated fluctuation ranges to the fluctuation range dataset in sequence.

[0031] Optionally, calculating the correlation coefficient between the i-th and j-th data in the health indicator dataset specifically includes:

[0032] Obtain the variance of the maximum-minimum normalization of the j-th health indicator data in the health indicator dataset, denoted as Var(x' j );

[0033] Calculate the correlation coefficient between the i-th and j-th data in the health indicator dataset, denoted as C ij , and the calculation formula is:

[0034]

[0035] Optionally, the uncertainty propagation analysis specifically includes:

[0036] Set the function of the overall health risk, denoted as R, specifically as:

[0037]

[0038] Wherein, α i is the risk weight of the i-th health indicator data; δ i is the fluctuation range of the i-th health indicator data; x' i is the value of the i-th health indicator data after maximum-minimum normalization;

[0039] Calculate the influence degree of each health indicator data item on the overall risk through partial derivatives. The sensitivity metric formula is:

[0040]

[0041] Set the sensitivity distribution set, denoted as S;

[0042] Add each calculated S iAdded to the set of sensitivity distributions to obtain S = [S1, S2,..., S n ;

[0043] Construct a risk propagation matrix, denoted as M, through correlation and sensitivity to describe the mutual influence between variables:

[0044] M[i, j] = S i ·C ij ;

[0045] where M[i, j] is the data in the i-th row and j-th column of the risk propagation matrix M; S i is the i-th data in the set of sensitivity distributions; C ij is the correlation coefficient between the i-th and j-th data in the health indicator dataset;

[0046] Calculate the total propagation effect of each variable in the health indicator dataset:

[0047]

[0048] Sort all T i and obtain the maximum propagation effect.

[0049] Optionally, the optimization generation of the doctor's order uncertainty specifically includes:

[0050] Obtain the target health value of each data in the health indicator dataset, denoted as is the target health value of the i-th data in the health indicator dataset;

[0051] Obtain the set of doctor's order parameters, denoted as u, and denote the j-th data in the set as u j ;

[0052] Obtain the target dose corresponding to each data in the set of doctor's order parameters, and denote the target dose of the j-th data in the set of doctor's order parameters as

[0053] Obtain the expected fluctuation range of the fluctuation range of each health indicator data in the health indicator dataset, and denote the obtained i-th expected fluctuation range as

[0054] Set the objective function for doctor's order optimization, denoted as

[0055]

[0056] where ψ i is the weight coefficient of the i-th data in the health indicator dataset in the objective function; γ jis the weight coefficient of the j-th data in the set of medical order parameters in the objective function; β i is the weight coefficient of the fluctuation range of the i-th data in the health index dataset in the objective function; m is the total number of data in the set of medical order parameters;

[0057] Update the medical order parameter u based on the gradient descent method j and the fluctuation range δ i ;

[0058] Set the iteration stop condition;

[0059] After each iteration, correct the fluctuation range δ i ;

[0060] Obtain the finally optimized j-th medical order parameter, denoted as

[0061] Obtain the fluctuation range of the i-th data in the finally optimized health index dataset, denoted as

[0062] Set the optimized set of medical orders, denoted as

[0063]

[0064] Optionally, the updating of the medical order parameter u j and the fluctuation range δ i , specifically includes:

[0065] Update the medical order parameter u j and the fluctuation range δ i :

[0066]

[0067] wherein, is the value of the j-th data in the set of medical order parameters at the t-th iteration; is the value of the fluctuation range of the i-th data in the health index dataset at the t-th iteration; η is the learning rate.

[0068] Optionally, the setting of the iteration stop condition specifically includes:

[0069] Set the tolerance error value threshold, denoted as ∈;

[0070] Set the maximum number of iterations, denoted as T max ;

[0071] Obtain the value of the objective function at the t-th iteration, denoted as

[0072] Obtain the value of the objective function at the (t + 1)-th iteration, denoted as

[0073] When the objective function has a change less than the tolerance error value threshold, stop the iteration:

[0074]

[0075] When the number of iterations is greater than the maximum number of iterations, force the iteration to stop:

[0076] t > T max .

[0077] Optionally, after each iteration, correct the fluctuation amplitude δ i , specifically including:

[0078]

[0079] where is the value of the fluctuation amplitude of the i-th item of data in the health index dataset at the t-th iteration; σ i is the correction coefficient of the i-th item of data in the health index dataset; M ij is the data in the i-th row and j-th column of the risk propagation matrix M; S j is the j-th item of data in the sensitivity distribution set.

[0080] The present invention has the following beneficial effects:

[0081] 1. By establishing a multi-dimensional uncertainty model for health data, this method solves the problem of how to integrate multi-dimensional health data and eliminate the scale differences between different data. Through max-min normalization processing, the dimension differences between different health index data are eliminated, enabling the data to be within the same standard range, which is convenient for subsequent analysis and comparison. This step avoids analysis biases that may be caused by different data units and ranges. By calculating the fluctuation amplitude and its correlation of health data, it is possible to identify which health indicators have large fluctuations and clarify the mutual influence between different health data. This provides a valuable reference basis for subsequent uncertainty propagation analysis. By assigning weights to each item of health data and introducing a regulation factor, this method can dynamically adjust the influence of health indicators according to the health conditions of different patients, making the data model more personalized and targeted in practical applications, thereby improving the accuracy and effectiveness of medical advice generation.

[0082] 2. By obtaining the variance and mean of each health indicator data after maximum-minimum normalization, the volatility of each data item can be accurately quantified. This step effectively solves the problem of how to convert volatility into a quantifiable value, thus enabling the volatility of health data to be clearly presented for further analysis. By calculating the fluctuation range, this method can identify and quantify the fluctuation range of each health data item, enabling accurate assessment of data uncertainty when faced with multi-dimensional and different-source health data. This not only enhances the understanding of the patient's health status but also provides a reliable basis for subsequent health risk assessment, avoiding misjudgments caused by ignoring data fluctuations. Setting up a dataset of fluctuation ranges and sequentially adding the fluctuation ranges of each health data item lays the foundation for subsequent uncertainty propagation analysis. This dataset provides a comprehensive reference framework of fluctuation ranges for the system, enabling comprehensive evaluation from multiple dimensions when analyzing the impact of health data on the patient's health, thus enhancing the accuracy and comprehensiveness of the analysis. By quantifying the fluctuation range of health data, the changing trend of the patient's health status can be predicted more precisely. The introduction of this volatility data makes the formulation of medical orders and treatment plans more in line with the actual situation of individual patients, reducing the possibility of misdiagnosis and over-treatment, thereby improving the treatment effect and patient satisfaction.

[0083] 3. By calculating the correlation coefficient, this method can effectively quantify the mutual relationship between each health data item, solving the problem of how to identify potential associations among multiple health indicators. Health data usually has multi-dimensional characteristics. Analyzing each indicator separately may not fully reflect the true health status. Through the calculation of the correlation coefficient, it can be revealed which health indicators have a strong relationship and which have a weak relationship, thus providing strong support for data analysis. By calculating the correlation among various health indicators, the dependence between different health parameters can be deeply analyzed, avoiding the problem of relying solely on individual data points for health assessment. This method enables multi-dimensional data to complement each other during the analysis process, thereby enhancing the comprehensiveness of the assessment. For example, an abnormal fluctuation in a certain health indicator may be caused by changes in other related indicators. Through correlation analysis, potential health problems can be detected early. Calculating the correlation coefficient between health indicators helps to more accurately predict the patient's health risks. During the patient health assessment process, correlation analysis can reveal the causal relationship and influence between different health data. Especially in complex health states, it can help doctors discover potential health problems and intervene in a timely manner, reducing the possibility of misdiagnosis and missed diagnosis. By calculating the correlation coefficient between health data, a more accurate basis for formulating medical orders can be provided. By identifying highly correlated health indicators, the formulation of medical orders can be optimized to ensure that the medical orders are more in line with the actual health status of the patient. This not only improves the effect of personalized treatment but also reduces unnecessary interference factors during the treatment process.

[0084] 4. By constructing a risk propagation matrix, this step can effectively quantify the risk coupling relationship between different health indicators and calculate the impact degree of each health indicator on the overall health risk through sensitivity analysis. Traditional health risk analysis often only focuses on a single indicator and ignores the interaction between various indicators. The implementation of this method can effectively solve this problem and make the health risk assessment more comprehensive and accurate. By calculating the sensitivity of health indicators, it is possible to identify which health indicators contribute more to the overall health risk, helping doctors and medical staff to focus on those key health factors and thus formulate more personalized and precise treatment plans. For example, some health indicators contribute more to risk propagation and may be the key to potential health problems. Early intervention in these key indicators can significantly reduce the health risk of patients. By quantifying the risk propagation effect between different health indicators, the risk propagation matrix provides a comprehensive and dynamic health risk map, enabling decision-makers to evaluate the impact of changes in each health indicator on the overall health status from a global perspective. This process can effectively guide medical decisions, such as adjusting treatment strategies and formulating health intervention measures, to improve the accuracy and effectiveness of health management. The risk propagation matrix can provide a clear propagation path for the uncertainty in health data, ensuring that the health warning system can identify the risk source in advance. The maximum propagation effect obtained through ranking can provide the priority for real-time monitoring of the intelligent health monitoring system, giving priority to those health indicators with a large propagation effect, thereby improving the response speed of disease prevention and risk management.

[0085] 5. By obtaining the target health values of each health indicator data and optimizing the medical order parameters, this step can achieve the optimization of the medical order plan according to the specific conditions of the patient. Traditional medical order generation methods usually rely on experience and general rules and cannot accurately optimize for the health status of each patient. Through this step, however, the medical order can be quantitatively analyzed based on factors such as the patient's individual health indicator data and fluctuation range, thus improving the personalization level and treatment effect of the medical plan. By adding the weight coefficient of the fluctuation range to the objective function and optimizing it, the adaptability between the medical order parameters and the fluctuation range of health data is ensured, thereby effectively reducing health risks. In practical applications, the patient's health condition may fluctuate. By correcting the fluctuation range, it is ensured that the result after each iteration is more matched with the patient's current health condition, avoiding medical decision-making mistakes caused by not adapting to risk changes. By using the gradient descent method for parameter update and setting the iteration stop condition, the optimized medical order can be continuously adjusted until an optimal solution is reached. The gradient descent method can gradually optimize the parameters and fluctuation range of the medical order, precisely adjusting the dosage of each medical order to match the patient's health condition. In this way, not only can the treatment effect of the patient be effectively improved, but also the treatment plan can be dynamically adjusted according to the changes in health data to ensure the continuous effectiveness of the treatment. By establishing an optimization objective function and correcting the fluctuation range after each iteration, it is ensured that the final optimization result adapts to the changes and sensitivity of the patient's health data, making medical decisions more accurate. Traditional medical decisions often rely on doctors' experience and conventional treatment plans, while this method realizes the accurate assessment of the patient's health risks through a data-driven approach, thus being able to provide a more scientific decision-making basis. By continuously iterating and optimizing and correcting the fluctuation range, the medical order can adapt to the risk distribution and health status changes of different patients. With the dynamic changes in health data, this method can adjust the medical order in real time, ensuring the flexibility and adaptability of the treatment plan. For example, the health data of some patients may fluctuate during the treatment process. Through the optimization step, the medical order can be adjusted in a timely manner, thus improving the treatment effect. By correcting and optimizing the fluctuation range, this method effectively reduces the risks brought by the uncertainty of health data and improves the stability of the medical order. Especially when facing complex patient health conditions, this uncertainty optimization generation strategy can minimize the deviation of the medical plan to the greatest extent and improve the stability and reliability of the treatment plan.

[0086] 6. The doctor's order parameters and fluctuation amplitude are gradually updated by the gradient descent method, and the current value is adjusted according to the partial derivative of the objective function in each iteration. In this way, the fluctuation amplitude of the doctor's order parameters and health data can be continuously optimized and gradually tend to the optimal state. Traditional treatment plans are usually determined once and lack dynamic adjustment for changes in patient health. The gradient descent method allows the doctor's order to be adjusted in real time according to the patient's health data, ensuring that the treatment plan after each iteration can more accurately respond to the fluctuation of the patient's health. The gradient descent method calculates the gradient of the objective function to ensure that each iteration is optimized in the direction of the target minimum value, reducing the possibility of falling into the local optimal solution. For complex medical decision-making problems, local optimality may lead to unsatisfactory treatment effects. The application of the gradient descent method can minimize the objective function as a whole, so that the doctor's order parameters and fluctuation amplitude can find the global optimal solution, thereby improving the accuracy of the treatment effect. By setting the learning rate, the size of each iteration step can be controlled to prevent the update process from being too drastic or slow. An appropriate learning rate helps to speed up the optimization process and avoid unsatisfactory convergence speed or oscillation problems caused by too large or too small update step sizes. In actual medical applications, the adjustment of the learning rate can be flexibly controlled according to the specific situation of the patient, thereby improving the efficiency and stability of the optimization of medical orders. Through the iterative optimization of the objective function, the gradient descent method ensures that the parameters and fluctuations of the medical orders are constantly adjusted to accurately reflect the changes in the patient's health needs. For patients with multiple diseases or large fluctuations in health status, this optimization method can better adapt to the changes in individual health status, provide personalized treatment plans, and significantly improve the treatment effect. The application of the gradient descent method ensures that the fluctuation range of health data can be corrected after each iteration, so that the medical order can adapt to the patient's health changes in real time. For example, some patients may experience large health fluctuations. The gradient descent method ensures that the medical order can always be adjusted according to the patient's actual health status at different stages by updating the fluctuation range. This adaptability greatly improves the flexibility of the treatment process and avoids the problem of mismatch between the medical order and the patient's health fluctuations. Through the optimization process of the gradient descent method, the medical order can be personalized according to each patient's unique health data, fluctuation range, and optimization goals. This not only avoids a one-size-fits-all treatment plan, but also ensures that the medical order plan is highly consistent with the patient's health, improving the treatment effect and patient compliance.

[0087] 7. By setting a tolerance error value threshold, when the change of the objective function after each iteration is less than the predetermined tolerance value, the iteration can be stopped. This effectively avoids unnecessary calculations caused by over-iteration and reduces the waste of computing resources. When the change of the objective function tends to be stable, the optimization process terminates, ensuring the efficiency and economy of the optimization. By setting a maximum number of iterations, it prevents unlimited iteration caused by the extremely slow change or ineffective convergence of the objective function. In practice, the gradient descent method may fail to converge the objective function in a short time due to improper initial value selection or learning rate setting. The limit of the maximum number of iterations ensures that even if the convergence of the objective function is slow, the iteration can be forced to stop within a reasonable number of times, avoiding the situation where the algorithm cannot terminate. The iteration stop condition can avoid continuing the optimization when the objective function has not changed significantly, thus avoiding the instability or overfitting of the model caused by excessive iterations. Through reasonable stop criteria, the optimization process can be maintained within a stable range, ensuring that the obtained medical order plan is both accurate and general, and avoiding the occurrence of overfitting problems. In the medical system, an overly long optimization process may delay the implementation time of the treatment plan and affect the patient's treatment experience. By setting reasonable stop conditions, the medical order optimization can be completed in a short time, and the treatment plan can be given as soon as possible, thus improving the patient's treatment experience and satisfaction.

[0088] 8. By modifying the fluctuation range according to the risk propagation matrix and sensitivity distribution, the fluctuation range can be dynamically adjusted to match the current risk distribution and sensitivity state. The fluctuation range of health indicator data usually varies with the change of the patient's health condition. A simple fixed fluctuation range may lead to the optimization scheme not conforming to the actual situation. By introducing a correction coefficient, this method can reflect the actual situation of the fluctuation of health indicators in real time, making the optimization process more adaptable and targeted. Modifying the fluctuation range can ensure that after each iteration, the optimization model can better adapt to the mutual influence between different health data. Especially when dealing with the risks of mutual coupling between multiple health indicators, the correction of the fluctuation range helps to more accurately measure the contribution of each health indicator to the overall risk, thus improving the accuracy of medical orders. This is of great significance for the generation of personalized treatment plans for patients, ensuring that the treatment plan is more scientific and refined. By modifying the fluctuation range after each iteration, the complex risk coupling relationship between health indicators can be effectively addressed. When the risk association between health indicators changes, the corrected fluctuation range can better capture these changes, thus optimizing the health risk prediction and avoiding wrong judgments and medical order mistakes caused by inaccurate risk coupling relationships. Dynamically adjusting the fluctuation range ensures the flexibility and stability of the optimization process when dealing with health data. Especially when the health data fluctuates greatly or there is a certain degree of uncertainty, modifying the fluctuation range can balance the influence of each health indicator, avoiding the deviation of the entire optimization process caused by too large or too small fluctuations of a certain health indicator, thus enhancing the stability of the treatment effect. During the optimization process, the modification of the fluctuation range helps to guide the gradient descent method to quickly find the optimal solution. By timely adjusting the fluctuation range, the convergence of the optimization goal can be accelerated, avoiding the problems of slow convergence or mis-convergence caused by improper setting of the fluctuation range. Therefore, this method can improve the optimization efficiency, reduce the number of iterations, and make the generation of treatment plans more efficient. The health data and fluctuation range of each patient may vary significantly. By modifying the fluctuation range, the weight of each data item in the optimization process can be precisely adjusted, making the medical order more personalized. For different patients, the dynamic adjustment of the fluctuation range ensures the individualized characteristics of the medical order, can better adapt to the changes in the health conditions of different patients, and improve the treatment effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0090] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0091] Example, refer to Figure 1 , A method for generating medical order information based on patient identification, including:

[0092] S1. Establish a multi-dimensional uncertainty model of health data:

[0093] Set a health index data set, denoted as X;

[0094] Collect multiple indicators from the health records authorized by the patient;

[0095] The multiple indicators include physiological data and medical history data, and the collected data is added to the health index data set to obtain X = [x1, x2,..., x n , where n is the total number of dimensions of health data, and x i is the i-th health index data in the health index data set;

[0096] Perform max-min normalization on each health index data:

[0097]

[0098] Among them, min(x i ) and max(x i ) are the minimum and maximum values of the i-th health index data in the health index data set respectively; x' i is the value after max-min normalization;

[0099] Update the value after max-min normalization of each health index data to the data at the corresponding position in the health index data set;

[0100] Calculate the fluctuation range of each data in the health index data set;

[0101] Calculate the correlation coefficient between the i-th and j-th data in the health index data set;

[0102] Obtain the mean value of the data set X, denoted as μ;

[0103] Set a weight for each health data in the health index data set, and the calculation method is:

[0104]

[0105] Among them, k is a regulation factor used to amplify the weight difference; ω i is the weighting coefficient of the i-th health index data;

[0106] S2. Uncertainty propagation analysis;

[0107] S3. Optimization generation of medical advice uncertainty.

[0108] By establishing a multi-dimensional uncertainty model of health data, this method solves the problem of how to integrate multi-dimensional health data and eliminate the scale differences between different data. Through maximum-minimum normalization processing, the dimensional differences between different health index data are eliminated, enabling the data to be within the same standard range, which is convenient for subsequent analysis and comparison. This step avoids analysis biases that may be caused by different data units and ranges. By calculating the fluctuation amplitude and its correlation of health data, it is possible to identify which health indicators have large fluctuations and clarify the mutual influence between different health data. This provides a valuable reference basis for subsequent uncertainty propagation analysis. By assigning weights to each piece of health data and introducing a regulation factor, this method can dynamically adjust the influence of health indicators according to the health conditions of different patients, making the data model more personalized and targeted in practical applications, thereby improving the accuracy and effectiveness of medical advice generation.

[0109] The specific steps for calculating the fluctuation amplitude of each data in the health index dataset include:

[0110] Obtain the variance of the maximum-minimum normalized i-th health index data in the health index dataset, denoted as Var(x' i );

[0111] Obtain the mean of the maximum-minimum normalized i-th health index data in the health index dataset, denoted as E(x' i );

[0112] Estimate the fluctuation amplitude of each data in the health index dataset. The formula for the fluctuation amplitude is:

[0113] δ i = σ i ·x' i ;

[0114] Among them, δ i is the fluctuation amplitude of the i-th health index data;

[0115] Set the fluctuation amplitude dataset, denoted as Δ;

[0116] Add all the calculated fluctuation amplitudes to the fluctuation amplitude dataset in sequence.

[0117] By obtaining the variance and mean after maximum-minimum normalization of each health indicator data, the volatility of each data can be accurately quantified. This step effectively solves the problem of how to convert volatility into quantifiable numerical values, thus enabling the volatility of health data to be clearly presented for further analysis. By calculating the fluctuation range, this method can identify and quantify the fluctuation range of each health data, enabling accurate assessment of the uncertainty of data in the face of multi-dimensional and different-source health data. This not only enhances the understanding of the patient's health status but also provides a reliable basis for subsequent health risk assessment, avoiding misjudgments caused by ignoring data fluctuations. Setting up a dataset of fluctuation ranges and sequentially adding the fluctuation ranges of each health data lays the foundation for subsequent uncertainty propagation analysis. This dataset provides a comprehensive reference framework of fluctuation ranges for the system, enabling comprehensive evaluation from multiple dimensions when analyzing the impact of health data on the patient's health and improving the accuracy and comprehensiveness of the analysis. By quantifying the fluctuation range of health data, the change trend of the patient's health status can be predicted more accurately. The introduction of this volatility data makes the formulation of medical orders and treatment plans more in line with the actual situation of the patient, reducing the possibility of misdiagnosis and over-treatment, thereby improving the treatment effect and patient satisfaction.

[0118] Calculating the correlation coefficient between the i-th and j-th data in the health indicator dataset specifically includes:

[0119] Obtaining the variance of the j-th health indicator data in the health indicator dataset after maximum-minimum normalization, denoted as Var(x' j );

[0120] Calculating the correlation coefficient between the i-th and j-th data in the health indicator dataset, denoted as C ij , and the calculation formula is:

[0121]

[0122] By calculating the correlation coefficient, this method can effectively quantify the mutual relationship between various health data, solving the problem of how to identify potential associations among multiple health indicators. Health data usually has multi-dimensional characteristics. Analyzing each indicator separately may not fully reflect the true health status. Through the calculation of the correlation coefficient, it can be revealed which health indicators have strong relationships and which have weak relationships, thus providing strong support for data analysis. By calculating the correlation between various health indicators, the dependence between different health parameters can be deeply analyzed, avoiding the problem of relying solely on individual data points for health assessment. This method enables multi-dimensional data to complement each other during the analysis process, thereby enhancing the comprehensiveness of the assessment. For example, an abnormal fluctuation in a certain health indicator may be caused by changes in other related indicators. Through correlation analysis, potential health problems can be detected early. Calculating the correlation coefficient between health indicators helps to more accurately predict the health risks of patients. During the process of patient health assessment, correlation analysis can reveal the causal relationships and influences between different health data. Especially in complex health states, it can help doctors discover potential health problems and intervene in a timely manner, reducing the possibility of misdiagnosis and missed diagnosis. By calculating the correlation coefficient between health data, a more accurate basis can be provided for formulating medical orders. By identifying highly correlated health indicators, the formulation of medical orders can be optimized to ensure that the medical orders are more in line with the actual health conditions of patients. This not only improves the effect of personalized treatment but also reduces unnecessary interference factors during the treatment process.

[0123] The uncertainty propagation analysis specifically includes:

[0124] Set the function of the overall health risk, denoted as R, specifically as:

[0125]

[0126] where α i is the risk weight of the i-th health indicator data; δ i is the fluctuation range of the i-th health indicator data; x' i is the value of the i-th health indicator data after maximum-minimum normalization;

[0127] Calculate the influence degree of each health indicator data item on the overall risk through partial derivatives. The sensitivity metric formula is:

[0128]

[0129] Set the sensitivity distribution set, denoted as S;

[0130] Add each calculated S i to the sensitivity distribution set to obtain S = [S1, S2,..., Sn ;

[0131] Construct a risk propagation matrix, denoted as M, through correlation and sensitivity to describe the mutual influence between variables:

[0132] M[i,j] = S i ·C ij ;

[0133] where M[i,j] is the data in the i-th row and j-th column of the risk propagation matrix M; S i is the i-th data in the sensitivity distribution set; C ij is the correlation coefficient between the i-th and j-th data in the health index dataset;

[0134] The risk propagation matrix M quantifies the risk coupling relationship between different health indicators and calculates the total propagation effect of each variable in the health index dataset:

[0135]

[0136] Sort all T i and obtain the maximum propagation effect.

[0137] By constructing the risk propagation matrix, this step can effectively quantify the risk coupling relationship between different health indicators and calculate the influence degree of each health indicator on the overall health risk through sensitivity analysis. Traditional health risk analysis often only focuses on a single indicator and ignores the interaction between various indicators. The implementation of this method can effectively solve this problem and make the health risk assessment more comprehensive and accurate. By calculating the sensitivity of health indicators, it is possible to identify which health indicators contribute more to the overall health risk, helping doctors and medical staff to focus on those key health factors, and thus formulating more personalized and precise treatment plans. For example, some health indicators contribute more to risk propagation and may be the key to potential health problems. Early intervention on these key indicators can significantly reduce the health risk of patients. By quantifying the risk propagation effect between different health indicators, the risk propagation matrix provides a comprehensive and dynamic health risk map, enabling decision-makers to evaluate the impact of changes in each health indicator on the overall health status from a global perspective. This process can effectively guide medical decisions, such as adjusting treatment strategies and formulating health intervention measures, to improve the accuracy and effectiveness of health management. The risk propagation matrix can provide a clear propagation path for the uncertainty in health data, ensuring that the health warning system can identify the risk source in advance. The maximum propagation effect obtained through sorting can provide the priority of real-time monitoring for the intelligent health monitoring system, giving priority to those health indicators with larger propagation effects, thereby improving the response speed of disease prevention and risk management.

[0138] The optimization generation of the medical advice uncertainty specifically includes:

[0139] Obtain the target health value of each data in the health index dataset, denoted as The target health value of the i-th data in the health index dataset;

[0140] Obtain the medical advice parameter set, denoted as u, and denote the j-th data in the set as u j ;

[0141] Obtain the target dose corresponding to each data in the medical advice parameter set, and denote the target dose of the j-th data in the medical advice parameter set as

[0142] Obtain the expected fluctuation range of the fluctuation range of each health index data in the health index dataset, and denote the obtained i-th expected fluctuation range as

[0143] Set the objective function for optimizing the medical advice, denoted as With the main goal of minimizing and optimizing the patient's health:

[0144]

[0145] Among them, ψ i Is the weight coefficient of the i-th data in the health index dataset in the objective function; γ j Is the weight coefficient of the j-th data in the medical advice parameter set in the objective function; β i Is the weight coefficient of the fluctuation range of the i-th data in the health index dataset in the objective function; m is the total number of data in the medical advice parameter set;

[0146] Update the medical advice parameter u j And the fluctuation range δ i ;

[0147] Set the iteration stop condition;

[0148] After each iteration, correct the fluctuation range δ i To ensure that it adapts to the current risk distribution and sensitivity;

[0149] Obtain the finally optimized j-th medical advice parameter, denoted as

[0150] Obtain the fluctuation range of the i-th data in the finally optimized health index dataset, denoted as

[0151] Set the optimized medical advice set, denoted as

[0152]

[0153] By obtaining the target health values of each health index data and optimizing the medical order parameters, this step can achieve the optimization of the medical order plan according to the specific conditions of the patient. Traditional medical order generation methods usually rely on experience and general rules and cannot accurately optimize for the health status of each patient. Through this step, the medical order can be quantitatively analyzed based on factors such as the patient's individual health index data and fluctuation range, thereby improving the personalization degree and treatment effect of the medical plan. By adding a weight coefficient of the fluctuation range to the objective function and optimizing it, the adaptability between the medical order parameters and the fluctuation range of the health data is ensured, thus effectively reducing health risks. In practical applications, the health status of the patient may fluctuate. By correcting the fluctuation range, it is ensured that the result after each iteration is more matched with the current health status of the patient, avoiding medical decision-making mistakes caused by non-adaptation to risk changes. By using the gradient descent method to update the parameters and setting the iteration stop condition, the optimized medical order can be continuously adjusted until an optimal solution is reached. The gradient descent method can gradually optimize the parameters and fluctuation range of the medical order, accurately adjusting the dosage of each medical order to match the health status of the patient. In this way, not only can the treatment effect of the patient be effectively improved, but also the treatment plan can be dynamically adjusted according to the changes in health data to ensure the continuous effectiveness of the treatment. By establishing an optimization objective function and correcting the fluctuation range after each iteration, it is ensured that the final optimization result adapts to the changes and sensitivities of the patient's health data, making medical decisions more accurate. Traditional medical decisions often rely on the experience of doctors and conventional treatment plans, while this method realizes the accurate assessment of the patient's health risks through a data-driven approach, thereby being able to provide a more scientific decision-making basis. By continuously iterating and optimizing and correcting the fluctuation range, the medical order can adapt to the risk distribution and health status changes of different patients. With the dynamic changes in health data, this method can adjust the medical order in real time, ensuring the flexibility and adaptability of the treatment plan. For example, the health data of some patients may fluctuate during the treatment process. Through the optimization step, the medical order can be adjusted in a timely manner, thereby improving the treatment effect. Through the correction and optimization of the fluctuation range, this method effectively reduces the risks brought by the uncertainty of health data and improves the stability of the medical order. Especially in the face of complex patient health conditions, this strategy generated by uncertainty optimization can minimize the deviation of the medical plan to the greatest extent and improve the stability and reliability of the treatment plan.

[0154] Updating the medical order parameter u j and the fluctuation range δ i , specifically including:

[0155] Updating the medical order parameter u j and the fluctuation range δ i, the goal is to minimize the optimization objective function

[0156]

[0157] in, is the value of the j-th item of data in the medical order parameter set at the t-th iteration; is the fluctuation amplitude of the i-th data in the health indicator dataset at the t-th iteration; η is the learning rate, which controls the step size of each iteration.

[0158] The doctor's order parameters and fluctuation amplitude are gradually updated by the gradient descent method, and the current value is adjusted according to the partial derivative of the objective function in each iteration. In this way, the fluctuation amplitude of the doctor's order parameters and health data can be continuously optimized and gradually tend to the optimal state. Traditional treatment plans are usually determined once and lack dynamic adjustment for changes in patient health. The gradient descent method allows the doctor's order to be adjusted in real time according to the patient's health data, ensuring that the treatment plan after each iteration can more accurately respond to the fluctuation of the patient's health. The gradient descent method calculates the gradient of the objective function to ensure that each iteration is optimized in the direction of the target minimum value, reducing the possibility of falling into the local optimal solution. For complex medical decision-making problems, local optimality may lead to unsatisfactory treatment effects. The application of the gradient descent method can minimize the objective function as a whole, so that the doctor's order parameters and fluctuation amplitude can find the global optimal solution, thereby improving the accuracy of the treatment effect. By setting the learning rate, the size of each iteration step can be controlled to prevent the update process from being too drastic or slow. An appropriate learning rate helps to speed up the optimization process and avoid unsatisfactory convergence speed or oscillation problems caused by too large or too small update step sizes. In actual medical applications, the adjustment of the learning rate can be flexibly controlled according to the specific situation of the patient, thereby improving the efficiency and stability of the optimization of medical orders. Through the iterative optimization of the objective function, the gradient descent method ensures that the parameters and fluctuations of the medical orders are constantly adjusted to accurately reflect the changes in the patient's health needs. For patients with multiple diseases or large fluctuations in health status, this optimization method can better adapt to the changes in individual health status, provide personalized treatment plans, and significantly improve the treatment effect. The application of the gradient descent method ensures that the fluctuation range of health data can be corrected after each iteration, so that the medical order can adapt to the patient's health changes in real time. For example, some patients may experience large health fluctuations. The gradient descent method ensures that the medical order can always be adjusted according to the patient's actual health status at different stages by updating the fluctuation range. This adaptability greatly improves the flexibility of the treatment process and avoids the problem of mismatch between the medical order and the patient's health fluctuations. Through the optimization process of the gradient descent method, the medical order can be personalized according to each patient's unique health data, fluctuation range, and optimization goals. This not only avoids a one-size-fits-all treatment plan, but also ensures that the medical order plan is highly consistent with the patient's health, improving the treatment effect and patient compliance.

[0159] The setting of the iteration stop condition specifically includes:

[0160] Set a tolerance error value threshold, denoted as ∈;

[0161] Set the maximum number of iterations, denoted as T max ;

[0162] Obtain the value of the objective function at the t-th iteration, denoted as

[0163] Obtain the value of the objective function at the (t + 1)-th iteration, denoted as

[0164] When the objective function changes less than the tolerance error value threshold, stop the iteration:

[0165]

[0166] When the number of iterations is greater than the maximum number of iterations, force the iteration to stop:

[0167] t > T max .

[0168] By setting the tolerance error value threshold, when it is ensured that the change of the objective function after each iteration is less than the predetermined tolerance value, the iteration can be stopped. This effectively avoids unnecessary calculations caused by excessive iteration and reduces the waste of computing resources. When the change of the objective function tends to be stable, the optimization process terminates, ensuring the efficiency and economy of the optimization. By setting the maximum number of iterations, it prevents unlimited iteration caused by extremely slow change or ineffective convergence of the objective function. In practice, the gradient descent method may fail to converge the objective function in a short time due to improper initial value selection or learning rate setting. The limitation of the maximum number of iterations ensures that even if the convergence of the objective function is slow, it can be forced to stop iterating within a reasonable number of times, avoiding the situation where the algorithm cannot terminate. The iteration stop condition can prevent continuous optimization when the objective function has not changed significantly, thus avoiding instability or overfitting of the model caused by excessive iteration. Through reasonable stop criteria, the optimization process can be maintained within a stable range, ensuring that the obtained medical advice plan is both accurate and general, and avoiding the occurrence of overfitting problems. In the medical system, an overly long optimization process may delay the implementation time of the treatment plan and affect the patient's treatment experience. By setting reasonable stop conditions, the medical advice optimization can be completed in a short time, and the treatment plan can be given as soon as possible, thereby improving the patient's treatment experience and satisfaction.

[0169] After each iteration, the fluctuation amplitude δ i is corrected, specifically including:

[0170]

[0171] Among them, is the value of the fluctuation range of the i-th data in the health index dataset at the t-th iteration; σ i is the correction coefficient of the i-th data in the health index dataset; M ij is the data in the i-th row and j-th column of the risk propagation matrix M; S j is the j-th data in the sensitivity distribution set.

[0172] By modifying the fluctuation range according to the risk propagation matrix and the sensitivity distribution, the fluctuation range can be dynamically adjusted to match the current risk distribution and sensitivity state. The fluctuation range of health index data usually varies with the patient's health condition. A simple fixed fluctuation range may lead to an optimization plan that does not conform to the actual situation. By introducing a correction coefficient, this method can reflect the actual situation of health index fluctuations in real time, making the optimization process more adaptable and targeted. Modifying the fluctuation range can ensure that after each iteration, the optimization model can better adapt to the mutual influence between different health data. Especially when dealing with the risks of mutual coupling between multiple health indicators, the correction of the fluctuation range helps to more accurately measure the contribution of each health indicator to the overall risk, thereby improving the accuracy of medical advice. This is of great significance for the generation of personalized treatment plans for patients, ensuring that the treatment plan is more scientific and refined. By modifying the fluctuation range after each iteration, the complex risk coupling relationship between health indicators can be effectively addressed. When the risk association between health indicators changes, the corrected fluctuation range can better capture these changes, thereby optimizing health risk prediction and avoiding misjudgments and medical advice errors caused by inaccurate risk coupling relationships. Dynamically adjusting the fluctuation range ensures the flexibility and stability of the optimization process when dealing with health data. Especially when the health data fluctuates greatly or there is a certain degree of uncertainty, modifying the fluctuation range can balance the influence of each health indicator, avoiding deviations in the entire optimization process due to excessive or too small fluctuations in a certain health indicator, thereby enhancing the stability of the treatment effect. During the optimization process, the correction of the fluctuation range helps to guide the gradient descent method to quickly find the optimal solution. By adjusting the fluctuation range in a timely manner, the convergence of the optimization goal can be accelerated, avoiding slow convergence or mis-convergence problems caused by improper setting of the fluctuation range. Therefore, this method can improve the optimization efficiency, reduce the number of iterations, and make the generation of treatment plans more efficient. The health data and fluctuation range of each patient may vary significantly. By modifying the fluctuation range, the weight of each data in the optimization process can be precisely adjusted, making the medical advice more personalized. For different patients, the dynamic adjustment of the fluctuation range ensures the individualized characteristics of the medical advice, can better adapt to the changes in the health conditions of different patients, and improve the treatment effect.

[0173] It should be noted that, in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0174] The above are only the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A method for generating medical order information based on patient identification, characterized in that, Including: S1. Establish a multi-dimensional uncertainty model for health data: Set up a health index data set, denoted as X; Collect multiple indicators from the health records authorized by the patient; The multiple indicators include physiological data and medical history data. The collected data is added to the health indicator dataset to obtain X = [x1, x2,..., x n , where n is the total number of dimensions of the health data, and x i is the i-th health indicator data in the health indicator dataset; Perform max-min normalization on each piece of health index data: where min(x i ) and max(x i ) are the minimum and maximum values of the i-th health indicator data in the health indicator dataset respectively; x' i is the value after max-min normalization; Update the values after max-min normalization of each piece of health index data to the data at the corresponding position in the health index data set; Calculate the fluctuation amplitude of each data in the health index data set; The calculation of the fluctuation amplitude of each data in the health index data set specifically includes: Obtain the variance after maximum-minimum normalization of the i-th health indicator data in the health indicator dataset, denoted as Var(x' i ); Obtain the mean value of the maximum-minimum normalization of the i-th health indicator data in the health indicator dataset, denoted as E(x' i ); Estimate the fluctuation amplitude for each data in the health index data set, and the fluctuation amplitude calculation formula is: δ i = σ i · x' i ; Among them, δ i is the fluctuation range of the i-th health index data; Set up a fluctuation amplitude data set, denoted as Δ; Add all the calculated fluctuation amplitudes to the fluctuation amplitude data set in sequence; Calculate the correlation coefficient between the i-th and j-th data in the health index data set; The calculation of the correlation coefficient between the i-th and j-th data in the health index data set specifically includes: Obtain the variance after maximum-minimum normalization of the j-th health indicator data in the health indicator dataset, denoted as Var(x' j ); The correlation coefficient between the $i$-th and $j$-th data in the health indicator dataset is denoted as $C$ ij , and the calculation formula is as follows: Obtain the mean value of the data set X, denoted as μ; Set a weight for each health data in the health index data set, and the calculation method is: where k is a regulation factor; ω i is the weighted coefficient of the i-th health index data; S2. Uncertainty propagation analysis; The uncertainty propagation analysis specifically includes: Set up a function for the overall health risk, denoted as R, specifically: where α i is the risk weight of the i-th health index data; δ i is the fluctuation range of the i-th health index data; x' i is the value of the i-th health index data after maximum-minimum normalization; Calculate the influence degree of each health index data item on the overall risk through partial derivatives, and the sensitivity metric formula is: Set up a sensitivity distribution set, denoted as S; Add each calculated S i to the sensitivity distribution set to obtain S = [S1, S2,..., S n ; Construct a risk propagation matrix, denoted as M, through correlation and sensitivity to describe the mutual influence between variables: M[i,j] = S i ·C ij ; Among them, M[i,j] is the data in the i-th row and j-th column of the risk propagation matrix M; S i is the i-th data in the sensitivity distribution set; C ij is the correlation coefficient between the i-th and j-th data in the health index dataset; Calculate the total propagation effect of each variable in the health index data set; For all Ts i sort them and obtain the maximum propagation effect; S3. Optimized generation of medical order uncertainty; The optimized generation of medical order uncertainty specifically includes: Obtain the target health value for each piece of data in the health indicator dataset, denoted as is the target health value of the i-th piece of data in the health indicator dataset; Obtain the set of medical order parameters, denoted as u, and denote the j-th item of data in the set as u j ; Obtain the target dose corresponding to each data in the doctor's order parameter set, and denote the target dose of the j-th data in the doctor's order parameter set as Obtain the expected fluctuation amplitude of the fluctuation amplitude of each health indicator data in the health indicator dataset, and denote the obtained i-th expected fluctuation amplitude as Set the objective function for optimizing medical orders, denoted as where, ψ i is the weight coefficient of the i-th data in the health index dataset in the objective function; γ j is the weight coefficient of the j-th data in the doctor's order parameter set in the objective function; β i is the weight coefficient of the fluctuation range of the i-th data in the health index dataset in the objective function; m is the total number of data in the doctor's order parameter set; Update the medical order parameter u based on the gradient descent method j and the fluctuation amplitude δ i ; Updating the medical advice parameter u based on the gradient descent method j and the fluctuation amplitude δ i , specifically including: Update the medical advice parameter u based on the gradient descent method j and the fluctuation amplitude δ i : where, is the value of the j-th item of data in the doctor's order parameter set at the t-th iteration; is the value of the fluctuation range of the i-th item of data in the health index dataset at the t-th iteration; η is the learning rate Set up an iteration stop condition; The setting of the iteration stop condition specifically includes: Set a tolerance error value threshold, denoted as ∈; Set the maximum number of iterations, denoted as T max ; Obtain the value of the objective function at the \(t\) -th iteration, denoted as Obtain the value of the objective function at the (t + 1)-th iteration, denoted as When the objective function changes by less than the tolerance error value threshold, stop the iteration: When the number of iterations is greater than the maximum number of iterations, force the iteration to stop: t>T max ; After each iteration, correct the fluctuation amplitude δ i ; After each iteration, the fluctuation amplitude δ i is corrected, specifically including: wherein, is the value of the fluctuation range of the i-th data in the health index dataset at the t-th iteration; σ i is the correction coefficient of the i-th data in the health index dataset; M ij is the data in the i-th row and j-th column of the risk propagation matrix M; S j the j-th data in the sensitivity distribution set Obtain the j-th medical order parameter after final optimization, denoted as Obtain the fluctuation amplitude of the i-th data in the finally optimized health index dataset, denoted as Set the optimized set of medical orders, denoted as

Images