CN118969320B

CN118969320B - A method and system for automatically optimizing parameters of mathematical model for individualized medication

Info

Publication number: CN118969320B
Application number: CN202411444204.4A
Authority: CN
Inventors: 董亚琳; 王陶陶; 陈思颖; 王楚慧; 孙金钥; 曾卫东; 王光启
Original assignee: First Affiliated Hospital of Xian Jiaotong University
Current assignee: First Affiliated Hospital of Xian Jiaotong University
Priority date: 2024-10-16
Filing date: 2024-10-16
Publication date: 2025-01-24
Anticipated expiration: 2044-10-16
Also published as: CN118969320A

Abstract

The application relates to an automatic optimization method and system for parameters of an individualized medication mathematical model. The method comprises the steps of obtaining physiological data, medication data, drug concentration monitoring data and inspection data of a trained patient object, processing by using a group medicine dynamic model to obtain a predicted value of the blood concentration, automatically optimizing model parameters of the group medicine dynamic model based on error information between the predicted value of the blood concentration and an actual measured blood concentration value to obtain a trained group medicine dynamic model, and processing by using the trained group medicine dynamic model to obtain a predicted result, wherein the predicted result is the predicted value of the blood concentration of the tested patient object. Therefore, the parameter of the group pharmacokinetic model can be adjusted by utilizing an error feedback mechanism in an automatic iteration mode, so that the predicted blood concentration result is as close to an actual measured value as possible, and the prediction accuracy of the group pharmacokinetic model is improved.

Description

Automatic optimization method and system for parameters of personalized medicine mathematical model

Technical Field

The application relates to the technical field of personalized medicine, in particular to an automatic optimization method and system for parameters of a personalized medicine mathematical model.

Background

With advances in medical technology and biometrics, personalized medicine is becoming a reality. The aim of accurate medication is to determine the most appropriate type of medicament and the dosage thereof according to the individual characteristics (such as genotype, physiological state and the like) of a patient so as to formulate the most appropriate treatment scheme to achieve the best treatment effect and reduce the occurrence of side effects. The effectiveness and safety of drug therapy is largely dependent on the concentration of the drug in the patient, which is affected by a variety of factors, including the physiology, pathological state, degree of drug metabolism, drug interactions, etc. of the patient. Therefore, accurate prediction of blood concentration is critical to achieving personalized medicine.

Bulk pharmacokinetic models are statistical tools used to describe the absorption, distribution, metabolism, and excretion (ADME) processes of drugs in a particular population. Such a model takes into account not only inter-individual differences, but also inter-individual variability. It can integrate data from multiple sources, such as clinical trials and laboratory test data, to more accurately predict pharmacokinetic behavior of drugs in individuals. However, in practical applications, population pharmacokinetic models are challenging, mainly because the core parameters of the model are of large variance, resulting in a dispersion of the predicted results, which suggests that there may be large errors for the actual patient. In addition, many of the model parameters currently in use are calculated based on actual medical cases in the past, and these data may be affected by factors such as differences in geographical distribution, differences in crowd characteristics, errors in measurement equipment, and insufficient sample size.

Therefore, it is desirable to develop an automatic optimization scheme for parameters of an individualized medication mathematical model to enhance the generalization ability of the model and improve the predictive performance of the model.

Disclosure of Invention

This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

In a first aspect, the present application provides a method for automatic optimization of parameters of an individualized medication mathematical model, the method comprising:

Acquiring physiological data, medication data, drug concentration monitoring data and test data of a training patient object;

Processing physiological data, medication data, drug concentration monitoring data and test data of the trained patient object by using a group pharmacokinetic model to obtain a predicted value of the blood drug concentration;

automatically optimizing model parameters of the group pharmacokinetic model based on error information between the predicted value of the blood concentration and the actual measured blood concentration value to obtain a trained group pharmacokinetic model;

And processing the physiological data, the drug concentration monitoring data and the test data of the tested patient object by using the trained group pharmacokinetic model to obtain a prediction result, wherein the prediction result is a predicted value of the blood concentration of the tested patient object.

The method comprises the steps of obtaining a low-dimensional embedded coding vector of training physiological data by using a physiological data embedded coding matrix, obtaining a low-dimensional embedded coding vector of training physiological data by using a drug data embedded coding matrix, obtaining a low-dimensional embedded coding vector of training drug data by using drug data of the training patient object, obtaining a low-dimensional embedded coding vector of training drug data by using a drug concentration monitoring data embedded coding matrix, obtaining a low-dimensional embedded coding vector of training drug concentration monitoring data by using a test data embedded coding matrix, obtaining a low-dimensional embedded coding vector of training test data by using a test data embedded coding matrix, obtaining a low-dimensional embedded coding vector of training physiological data, the low-dimensional embedded coding vector of training drug concentration monitoring data and the low-dimensional embedded coding vector of training test data by using the group drug dynamic model, and obtaining a dynamic coded vector by using a dynamic sequence of the group drug dynamic coding.

Optionally, the physiological data of the trained patient includes clinical test indexes such as basic disease, age, sex, weight, height, body surface area, combined medication, genotype, liver function index, kidney function index, blood convention and the like of the trained patient, the medication data of the trained patient includes administration time, administration dose, administration route, administration frequency, infusion time and the like, the drug concentration monitoring data of the trained patient includes monitoring time, blood concentration value, detection method and instrument type, and the test data includes liver and kidney function test data, blood convention test data and infection index test data.

Optionally, the group pharmacokinetic model is used for carrying out pharmacokinetic encoding on the training physiological data low-dimensional embedded coding vector, the training drug concentration monitoring data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector to obtain a pharmacokinetic timing coding vector, and the group pharmacokinetic model is used for carrying out the pharmacokinetic encoding on the training physiological data low-dimensional embedded coding vector, the training drug concentration monitoring data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector based on node potential energy time-space information transmission to obtain the pharmacokinetic timing coding vector.

Optionally, the group pharmacokinetics model is used for carrying out the pharmacokinetic encoding based on node potential energy space-time information transfer on the training physiological data low-dimensional embedded encoding vector, the training drug concentration monitoring data low-dimensional embedded encoding vector and the training test data low-dimensional embedded encoding vector to obtain the pharmacokinetic timing encoding vector, the group pharmacokinetics model comprises the steps of taking the training drug concentration monitoring data low-dimensional embedded encoding vector as a target data low-dimensional embedded encoding vector, taking the training physiological data low-dimensional embedded encoding vector, the training drug data low-dimensional embedded encoding vector and the training test data low-dimensional embedded encoding vector as a sequence of message transmission data low-dimensional embedded encoding vectors, calculating the information transmission dynamic energy space-time transfer weight of each message transmission data low-dimensional embedded encoding vector in the sequence of the message transmission data low-dimensional embedded encoding vector relative to the target data low-dimensional embedded encoding vector to obtain a sequence of information transmission dynamic energy space-time transfer weight, and carrying out fusion on the training physiological data low-dimensional embedded encoding vector and the training test data low-dimensional embedded encoding vector based on the sequence of the information transmission dynamic energy space-time information transfer weight to obtain the training vector fusion vector.

Optionally, calculating information transfer dynamic energy space-time transfer weights of each message-passing data low-dimensional embedded coding vector in the sequence of message-passing data low-dimensional embedded coding vectors relative to the target data low-dimensional embedded coding vector to obtain a sequence of information transfer dynamic energy space-time transfer weights, including calculating data low-dimensional embedded coding feature static energy values of each message-passing data low-dimensional embedded coding vector in the sequence of message-passing data low-dimensional embedded coding vectors; the method comprises the steps of determining a data low-dimensional embedded coding feature energy space attenuation factor based on a space span coefficient between each message transmission data low-dimensional embedded coding vector and the target data low-dimensional embedded coding vector, determining a data low-dimensional embedded coding feature energy time attenuation factor based on a time span coefficient between each message transmission data low-dimensional embedded coding vector and the target data low-dimensional embedded coding vector, performing space-time modulation on a data low-dimensional embedded coding feature static energy value of each message transmission data low-dimensional embedded coding vector based on the data low-dimensional embedded coding feature energy space attenuation factor and the data low-dimensional embedded coding feature energy time attenuation factor to obtain a data low-dimensional embedded coding feature dynamic energy space-time transfer value of each message transmission data low-dimensional embedded coding vector relative to the target data low-dimensional embedded coding vector, and inputting the data low-dimensional embedded coding feature dynamic energy transfer value of each message transmission data low-dimensional embedded coding vector into an energy space-time gating module based on a mask function to obtain a sequence of the information transfer dynamic energy transfer weight.

The method comprises the steps of calculating the mean value and standard deviation of each message transmission data low-dimensional embedded coding vector in a sequence of message transmission data low-dimensional embedded coding vectors to obtain the message transmission data low-dimensional embedded coding feature mean value and the message transmission data low-dimensional embedded coding standard deviation, calculating the fourth power of the position difference between the message transmission data low-dimensional embedded coding vector and the message transmission data low-dimensional embedded coding feature mean value, calculating the expected value of the obtained fourth power modulation message transmission data low-dimensional embedded coding offset vector to obtain the data low-dimensional embedded coding feature offset expected factor, calculating the division between the data low-dimensional embedded coding feature offset expected factor and the fourth power of the message transmission data low-dimensional embedded coding standard deviation, and adding the maximum feature value in the message transmission data low-dimensional embedded coding vector to obtain the data low-dimensional embedded coding feature static energy value.

Optionally, based on the sequence of the information transfer dynamic energy space-time transfer weights, fusing the training physiological data low-dimensional embedded coding vector, the training drug data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector, and then transferring the obtained node information fusion representation to the training drug concentration monitoring data low-dimensional embedded coding vector to obtain the pharmacokinetic time sequence coding vector, wherein the method comprises the steps of calculating the weighted sum of the training physiological data low-dimensional embedded coding vector, the training drug data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector based on the sequence of the information transfer dynamic energy space-time transfer weights to obtain a message transmission data low-dimensional embedded coding node characteristic transfer aggregate modulation vector as the node information fusion representation; and calculating the position-based summation of the characteristic transfer aggregation modulation vector of the low-dimensional embedded coding node of the message transmission data and the low-dimensional embedded coding vector of the training drug concentration monitoring data to obtain the pharmacokinetic timing coding vector.

In a second aspect, the present application provides an automated optimization system for personalized medicine mathematical model parameters, the system comprising:

the data acquisition module is used for acquiring physiological data, medication data, drug concentration monitoring data and inspection data of a training patient object;

the data processing module is used for processing the physiological data, the drug concentration monitoring data and the test data of the training patient object by using the group pharmacokinetics model so as to obtain a predicted value of the blood drug concentration;

the model parameter training module is used for automatically optimizing model parameters of the group pharmacokinetic model based on error information between the predicted value of the blood concentration and the actual measured blood concentration value to obtain a trained group pharmacokinetic model;

And the prediction result generation module is used for processing the physiological data, the drug concentration monitoring data and the test data of the tested patient object by using the trained group pharmacokinetic model to obtain a prediction result, wherein the prediction result is a prediction value of the blood concentration of the tested patient object.

The data processing module comprises a physiological data embedded coding unit, a medication concentration monitoring data embedded coding unit and a prediction order coding unit, wherein the physiological data embedded coding unit is used for performing embedded coding on physiological data of a training patient object by using a physiological data embedded coding matrix to obtain a training physiological data low-dimensional embedded coding vector, the medication data embedded coding unit is used for performing embedded coding on medication data of the training patient object by using a medication data embedded coding matrix to obtain a training medication concentration monitoring data low-dimensional embedded coding vector, the examination data embedded coding unit is used for performing embedded coding on examination data of the training patient object by using a group medication dynamic model to obtain a training examination data low-dimensional embedded coding vector, the training medication concentration monitoring data low-dimensional embedded coding vector and the training medication concentration monitoring data low-dimensional embedded coding vector are used for performing embedded coding on the training patient object by using a medication concentration monitoring data embedded coding matrix to obtain a training medication concentration monitoring data low-dimensional embedded coding vector, and the prediction order coding unit is used for performing dynamic decoding on the examination data of the training patient object to obtain a prediction order coding vector.

By adopting the technical scheme, physiological data, medication data, drug concentration monitoring data and inspection data of a trained patient object are acquired, a group medicine dynamic model is used for processing to obtain a predicted value of the blood concentration, model parameters of the group medicine dynamic model are automatically optimized to obtain a trained group medicine dynamic model based on error information between the predicted value of the blood concentration and an actual measured blood concentration value, and the trained group medicine dynamic model is used for processing to obtain a predicted result which is the predicted value of the blood concentration of the tested patient object. Therefore, the parameter of the group pharmacokinetic model can be adjusted by utilizing an error feedback mechanism in an automatic iteration mode, so that the predicted blood concentration result is as close to an actual measured value as possible, and the prediction accuracy of the group pharmacokinetic model is improved.

Additional features and advantages of the application will be set forth in the detailed description which follows.

Drawings

The above and other features, advantages and aspects of embodiments of the present application will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. The same or similar reference numbers will be used throughout the drawings to refer to the same or like elements. It should be understood that the figures are schematic and that elements and components are not necessarily drawn to scale. In the drawings:

FIG. 1 is a flow chart illustrating a method of automatic optimization of personalized medicine mathematical model parameters, according to an exemplary embodiment.

FIG. 2 is a flow chart showing the step S102 of an automatic optimization method for personalized medicine mathematical model parameters, according to the embodiment shown in FIG. 1.

FIG. 3 is a block diagram illustrating an automated optimization system for individualized medication mathematical model parameters, according to an exemplary embodiment.

Fig. 4 is a block diagram of an electronic device, according to an example embodiment.

FIG. 5 is an application scenario diagram illustrating an automatic optimization method for personalized medicine mathematical model parameters, according to an exemplary embodiment.

Detailed Description

Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. While the application is susceptible of embodiment in the drawings, it is to be understood that the application may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided to provide a more thorough and complete understanding of the application. It should be understood that the drawings and embodiments of the application are for illustration purposes only and are not intended to limit the scope of the present application.

It should be understood that the various steps recited in the method embodiments of the present application may be performed in a different order and/or performed in parallel. Furthermore, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the application is not limited in this respect.

The term "including" and variations thereof as used herein are intended to be open-ended, i.e., including, but not limited to. The term "based on" is based at least in part on. The term "one embodiment" means "at least one embodiment," another embodiment "means" at least one additional embodiment, "and" some embodiments "means" at least some embodiments. Related definitions of other terms will be given in the description below.

It should be noted that the terms "first," "second," and the like herein are merely used for distinguishing between different devices, modules, or units and not for limiting the order or interdependence of the functions performed by such devices, modules, or units.

It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those skilled in the art will appreciate that "one or more" is intended to be construed as "one or more" unless the context clearly indicates otherwise.

The names of messages or information interacted between the devices in the embodiments of the present application are for illustrative purposes only and are not intended to limit the scope of such messages or information.

The following describes specific embodiments of the present application in detail with reference to the drawings.

FIG. 1 is a flow chart illustrating a method of automatic optimization of personalized medicine mathematical model parameters, as shown in FIG. 1, according to an exemplary embodiment, the method comprising:

step S101, physiological data, medication data, drug concentration monitoring data and inspection data of a training patient object are obtained;

Step S102, using a group pharmacokinetics model to process physiological data, medication data, drug concentration monitoring data and test data of the trained patient object so as to obtain a predicted value of the blood drug concentration;

step S103, automatically optimizing model parameters of the group pharmacokinetic model based on error information between the predicted value of the blood concentration and the actual measured blood concentration value to obtain a trained group pharmacokinetic model;

And step S104, using the trained group pharmacokinetic model to process the physiological data, the drug concentration monitoring data and the test data of the tested patient object to obtain a prediction result, wherein the prediction result is a predicted value of the blood concentration of the tested patient object.

It will be appreciated that in clinical practice, the pharmacokinetic profile will also be different compared to healthy volunteers due to variations in the physiological and pathological characteristics of the patient. This may result in poor efficacy in some patients due to too low a level of drug in the blood or in adverse effects due to too high a level of drug in the blood. Thus, the standard dosing regimen recommended by the instructions or guidelines for administration cannot be adapted to all patients. Physicians need to formulate personalized dosing regimens with the aid of blood concentration monitoring to ensure the effectiveness and safety of drug therapy.

The group pharmacokinetics allows the physiological differences among the subjects of the patients to be considered when the medicine is taken, so that the absorption, distribution, metabolism and excretion processes of the medicine in the body are clarified, the personalized prediction of the blood medicine concentration is realized, and a scientific basis is provided for the personalized medicine taking. However, the existing group pharmacokinetic model has limitations, and needs to be further improved to improve the accuracy and reliability in practical application. In the background of big data age, more and more data such as physiological parameters, human body information, medication information, inspection data and the like become available, and the big data provide possibility for constructing a more accurate personalized medication model. The machine learning algorithm and the deep learning framework can be utilized to effectively extract useful information from a complex data set and be used for optimizing parameters of a group pharmacokinetic model and learning error information sources.

FIG. 2 is a flow chart showing the step S102 of an automatic optimization method for personalized medicine mathematical model parameters, according to the embodiment shown in FIG. 1. As shown in FIG. 2, step S102 of processing the physiological data, the drug concentration monitoring data and the test data of the trained patient object by using the group pharmacokinetic model to obtain a predicted value of the blood drug concentration comprises the steps of S1021 of performing embedded encoding on the physiological data of the trained patient object by using a physiological data embedded encoding matrix to obtain a training physiological data low-dimensional embedded encoding vector;

Step S1022, using the medication data embedded coding matrix to perform embedded coding on the medication data of the training patient object so as to obtain a training medication data low-dimensional embedded coding vector;

Step S1023, performing embedded coding on the drug concentration monitoring data of the training patient object by using a drug concentration monitoring data embedded coding matrix to obtain a low-dimensional embedded coding vector of the training drug concentration monitoring data;

Step S1024, embedding and encoding the test data of the training patient object by using a test data embedding and encoding matrix to obtain a training test data low-dimensional embedded and encoded vector;

Step S1025, performing pharmacokinetic coding on the training physiological data low-dimensional embedded coding vector, the training medication concentration monitoring data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector by using the group pharmacokinetic model to obtain a pharmacokinetic time sequence coding vector;

and step S1026, decoding the pharmacokinetic time sequence coding vector to obtain the predicted value of the blood concentration.

Based on the above, in the technical scheme of the application, an automatic optimization method of individual medication mathematical model parameters is provided, which comprises the steps of obtaining physiological data, medication concentration monitoring data and inspection data of a training patient object, processing the physiological data, the medication concentration monitoring data and the inspection data of the training patient object by using a group medication mathematical model to obtain a predicted value of the blood concentration, automatically optimizing model parameters of the group medication mathematical model based on error information between the predicted value of the blood concentration and an actual measured blood concentration value to obtain a trained group medication mathematical model, obtaining the physiological data, the medication concentration monitoring data and the inspection data of a testing patient object, and processing the physiological data, the medication concentration monitoring data and the inspection data of the testing patient object by using the trained group medication mathematical model to obtain a predicted result which is the predicted value of the blood concentration of the testing patient object.

In one embodiment of the application, the physiological data reflects the patient's personal health and physical condition, generally including, but not limited to, basal disease, age, sex, weight, height, body surface area, co-medication, genotype, liver function index (e.g., ALT, AST, etc.), kidney function index (e.g., creatinine clearance), blood routine, and other factors that may affect drug metabolism. Medication data refers to information about the medication administered to a patient, including the time point of administration, the dosage administered, the route of administration (oral, intravenous, etc.), the frequency of administration, the time of infusion, other medications administered simultaneously, and their interactions. The drug concentration monitoring data mainly relates to monitoring of drug concentration in a patient, and specifically comprises actually measured monitoring time, a blood drug concentration value, a measured time point, a detection method and an instrument type. Test data include, in addition to the above drug therapy specific data, more extensive clinical test data such as liver and kidney function test data, blood routine test data, infection index test data, and other disease specific marker tests.

In the parameter optimization scheme of the personalized medication mathematical model, in the process of generating the predicted value of the blood concentration by using the group pharmacokinetic model, physiological data, medication concentration monitoring data and inspection data of a trained patient object are input into a data processing and analyzing algorithm based on artificial intelligence and deep learning to be processed, so that physiological data low-dimensional embedded coding features, medication concentration monitoring data low-dimensional embedded coding features and inspection data low-dimensional embedded coding features are captured, and information transfer semantics between different parameter data features relative to the medication detection embedded coding features based on node potential energy time and space are extracted from the features by the group pharmacokinetic model to reflect time sequence coding information of the pharmacokinetics, so that the prediction of the blood concentration and the parameter optimization of the group pharmacokinetic model are performed. Therefore, the parameters of the group pharmacokinetic model can be adjusted by utilizing an error feedback mechanism in an automatic iteration mode, and the parameters in the group pharmacokinetic model are continuously adjusted to better fit the actually observed drug concentration monitoring data, so that the predicted blood drug concentration result is as close to an actual measured value as possible, and the prediction accuracy of the group pharmacokinetic model is improved. The method not only improves the accuracy of the blood concentration prediction of the model, but also reduces the requirement of manual intervention, so that the establishment and updating of the group pharmacokinetic model are more efficient, and the method is favorable for supporting doctors to make more scientific and reasonable treatment decisions.

In the technical scheme of the application, firstly, physiological data of a training patient object is embedded and encoded by using a physiological data embedded encoding matrix to obtain a training physiological data low-dimensional embedded encoding vector, the medication data of the training patient object is embedded and encoded by using a medication data embedded encoding matrix to obtain a training medication data low-dimensional embedded encoding vector, the medication concentration monitoring data of the training patient object is embedded and encoded by using a medication concentration monitoring data embedded encoding matrix to obtain a training medication concentration monitoring data low-dimensional embedded encoding vector, and the examination data of the training patient object is embedded and encoded by using the examination data embedded encoding matrix to obtain the training medication data low-dimensional embedded encoding vector.

It should be appreciated that physiological data, medication data, drug concentration monitoring data, and test data may be highly dimensional and contain a significant amount of redundant information due to the original training patient subjects. By embedding the codes, high-dimensional data can be mapped into a lower-dimensional space while preserving important characteristic information of the data. Therefore, the calculation cost can be reduced, the problem of dimension disasters can be avoided, and the subsequent model training is more efficient. Furthermore, since different data types may have different units and magnitudes, direct use of raw data may result in a learning process for certain feature-dominated models. By transcoding, different types of input data can be normalized, while the process of embedding the code is effectively a process of feature extraction. It can help extract the most relevant features in the data, which is critical to improving the predictive power of the model. In particular, for unstructured data, embedded codes can convert them into numerical form, facilitating machine learning model processing. Specifically, in the technical scheme of the application, the embedded coding mode can respectively capture the physiological data low-dimensional embedded coding feature, the medication concentration monitoring data low-dimensional embedded coding feature and the test data low-dimensional embedded coding feature of the training patient object.

Further, because the training physiological data low-dimensional embedded coding vector, the training medication concentration monitoring data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector respectively contain low-dimensional embedded coding characteristic representations related to physiological data, medication concentration monitoring data and test data of a training patient object, different data embedded representations have different importance on a prediction task of blood medication concentration, and the importance is reflected in a space-time information transmission function. That is, since the physiological differences among the individuals of the patient subjects may cause the absorption, distribution, metabolism and excretion capacities of the drug in the body to be different, the blood concentration monitoring value and the time sequence information expressed in the low-dimensional embedded encoding vector of the training drug concentration monitoring data may be affected by the physiological index, the administration parameter and the test data parameter of the training patient subjects, and the semantic information carried by these different parameters (including the physiological index, the administration parameter and the test data parameter) during the embedding process may have different degrees of influence on them, so that the following task of predicting the blood concentration needs to be performed in consideration of such time-space information transmissibility during the information aggregation process. Based on the above, in the technical scheme of the application, the group pharmacokinetic model is used for carrying out pharmacokinetic coding based on node potential energy space-time information transfer on the training physiological data low-dimensional embedded coding vector, the training medication concentration monitoring data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector so as to obtain a pharmacokinetic timing coding vector. The pharmacokinetic coding process based on node potential energy space-time information transmission can calculate node information fusion expression characteristics among different data sources through a dynamic energy space-time transmission weight sequence, and transmits the node information fusion expression characteristics to the training drug concentration monitoring data to embed coding vectors in a low-dimensional mode, so that information of various data sources is integrated, and prediction accuracy and robustness of a model are improved.

In one embodiment of the application, the group pharmacokinetic model is used for carrying out pharmacokinetic encoding on the training physiological data low-dimensional embedded coding vector, the training drug concentration monitoring data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector to obtain a pharmacokinetic timing coding vector, and the group pharmacokinetic model is used for carrying out the pharmacokinetic encoding on the training physiological data low-dimensional embedded coding vector, the training drug concentration monitoring data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector based on node potential energy time-space information transmission to obtain the pharmacokinetic timing coding vector.

The process of the pharmacokinetic coding based on node potential energy space-time information transmission comprises the steps of firstly taking the training medicine concentration monitoring data low-dimensional embedded coding vector as a target data low-dimensional embedded coding vector, taking the training physiological data low-dimensional embedded coding vector, the training medicine data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector as a sequence of message transmission data low-dimensional embedded coding vectors, calculating a data low-dimensional embedded coding characteristic static energy value of each message transmission data low-dimensional embedded coding vector in the sequence of message transmission data low-dimensional embedded coding vectors, wherein the static energy value is equal to the maximum characteristic value of each message transmission data low-dimensional embedded coding vector, the maximum characteristic value of each message transmission data low-dimensional embedded coding vector, the mean and variance are related, and the static energy value provides an inherent attribute measure of the quantized eigenvector in space, provides a benchmark for the space-time attenuation of potential energy, and helps the model understand the static importance of the data embedded representation represented by each node in the space dimension and the time dimension. next, a data low-dimensional embedded encoding feature energy spatial attenuation factor is determined based on the spatial span coefficient between the respective message-propagating data low-dimensional embedded encoding vector and the target data low-dimensional embedded encoding vector, and a data low-dimensional embedded encoding feature energy temporal attenuation factor is determined based on the temporal span coefficient between the respective message-propagating data low-dimensional embedded encoding vector and the target data low-dimensional embedded encoding vector. The spatial span coefficient and the temporal span coefficient reflect the spatial distance between the nodes and the temporal distance between the nodes and the inspection data, respectively, and the attenuation factor adjusts the weight of the information transfer between the nodes based on the coefficients. It should be appreciated that the energy temporal attenuation factor and the energy spatial attenuation factor enable the model to adjust the intensity of information transfer according to the distance between nodes, enhancing the sensitivity of the model to locality and timeliness, and enabling the data information transfer and aggregation processes to be more consistent with actual processes. And further, based on the data low-dimensional embedded coding characteristic energy space attenuation factor and the data low-dimensional embedded coding characteristic energy time attenuation factor, performing space-time modulation on the data low-dimensional embedded coding characteristic static energy value of each message transmission data low-dimensional embedded coding vector so as to combine the space attenuation factor and the time attenuation factor through space-time modulation to adjust the static energy value, thereby reflecting the data parameter embedded semantics of the node and the dynamic change of the characteristic in space-time. Further, the data low-dimensional embedded coding feature dynamic energy space-time transfer values of the information transmission data low-dimensional embedded coding vectors are input into an energy transfer gating module based on a mask function, wherein the mask function is used for screening and enhancing important space-time transfer parameter embedded information, and a gating mechanism determines which information is important and which information can be ignored through learning. Therefore, the accuracy and the effectiveness of information transmission are improved through energy transmission gating, the information flow is adaptively adjusted, so that the model can pay more attention to the characteristics contributing to the task, and only the information meaningful to the subsequent blood concentration prediction task is ensured to be transmitted, and the accuracy and the effectiveness of information transmission can be improved. Then, based on the sequence of information transfer dynamic energy space-time transfer weights, a weighted sum of the training physiological data low-dimensional embedded code vector, the training medication data low-dimensional embedded code vector and the training test data low-dimensional embedded code vector is calculated. That is, the information from the nodes where the different data parameters are embedded semantics are synthesized by weighted summing the dynamic energy spatiotemporal delivery values of the message-propagated data low-dimensional embedded encoding vectors. The obtained low-dimensional embedded coding node characteristic transfer aggregation modulation vector of the message propagation data is used as node information fusion representation, so that the data embedded semantic information from different nodes is synthesized, and the richness and the accuracy of characteristic representation are improved. Finally, the information transmission data low-dimensional embedded coding node feature transmission aggregation modulation vector and the training test data low-dimensional embedded coding vector are subjected to position summation to generate a pharmacokinetic timing coding vector which synthesizes time-space information transmission information among physiological data, drug data and test data of a training patient object relative to drug concentration monitoring data embedded coding representation, so that abundant context information is provided for a follow-up blood concentration prediction task, and the processing capacity of a model on complex tasks is enhanced.

In one embodiment of the application, the group pharmacokinetic model is used for carrying out pharmacokinetic encoding based on node potential energy space-time information transfer on the training physiological data low-dimensional embedded encoding vector, the training drug concentration monitoring data low-dimensional embedded encoding vector and the training test data low-dimensional embedded encoding vector to obtain the pharmacokinetic time sequence encoding vector, the group pharmacokinetic model comprises the steps of taking the training drug concentration monitoring data low-dimensional embedded encoding vector as a target data low-dimensional embedded encoding vector, taking the training physiological data low-dimensional embedded encoding vector, the training drug concentration monitoring data low-dimensional embedded encoding vector and the training test data low-dimensional embedded encoding vector as a sequence of message transmission data low-dimensional embedded encoding vectors, calculating information transmission dynamic energy transfer weights of all message transmission data low-dimensional embedded encoding vectors in the sequence of the message transmission data low-dimensional embedded encoding vector relative to the target data low-dimensional embedded encoding vector to obtain an information transmission dynamic energy space-time weight sequence, and carrying out fusion on the training physiological data low-dimensional embedded encoding vector based on the information transmission dynamic energy transfer weight sequence to obtain the training drug concentration monitoring data low-dimensional embedded encoding vector.

Further, in one embodiment of the application, calculating information transfer dynamic energy space-time transfer weights of each message-passing data low-dimensional embedded code vector in the sequence of message-passing data low-dimensional embedded code vectors relative to the target data low-dimensional embedded code vector to obtain a sequence of information transfer dynamic energy space-time transfer weights includes calculating data low-dimensional embedded code feature static energy values of each message-passing data low-dimensional embedded code vector in the sequence of message-passing data low-dimensional embedded code vectors, determining data low-dimensional embedded code feature energy space-attenuation factors based on space-span coefficients between each message-passing data low-dimensional embedded code vector and the target data low-dimensional embedded code vector, determining data low-dimensional embedded code feature energy time-attenuation factors based on time-span coefficients between each message-passing data low-dimensional embedded code vector and the target data low-dimensional embedded code vector, performing space-time modulation on the data low-dimensional embedded coding feature static energy value of each message transmission data low-dimensional embedded coding vector to obtain a data low-dimensional embedded coding feature dynamic energy space-time transfer value of each message transmission data low-dimensional embedded coding vector relative to the target data low-dimensional embedded coding vector; and inputting the data low-dimensional embedded coding characteristic dynamic energy space-time transfer values of the information transmission data low-dimensional embedded coding vectors into an energy transfer gating module based on a mask function to obtain a sequence of the information transfer dynamic energy space-time transfer weights.

Further, in one embodiment of the application, calculating the data low-dimensional embedded coding feature static energy value of each message transmission data low-dimensional embedded coding vector in the sequence of message transmission data low-dimensional embedded coding vectors comprises the steps of calculating the mean value and standard deviation of the message transmission data low-dimensional embedded coding vectors to obtain the message transmission data low-dimensional embedded coding feature mean value and the message transmission data low-dimensional embedded coding standard deviation respectively, calculating the fourth power of the position difference between the message transmission data low-dimensional embedded coding vector and the message transmission data low-dimensional embedded coding feature mean value, calculating the expected value of the fourth power modulation message transmission data low-dimensional embedded coding offset vector to obtain the data low-dimensional embedded coding feature offset expected factor, calculating the division between the data low-dimensional embedded coding feature offset expected factor and the fourth power of the message transmission data low-dimensional embedded coding standard deviation, and adding the maximum feature value in the message transmission data low-dimensional embedded coding vector to obtain the data low-dimensional embedded coding feature static energy value.

In one embodiment of the application, based on the sequence of information transmission dynamic energy space-time transmission weights, the method comprises the steps of fusing the training physiological data low-dimensional embedded coding vector, the training drug data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector, transmitting the obtained node information fusion representation to the training drug concentration monitoring data low-dimensional embedded coding vector to obtain the pharmacokinetic timing coding vector, and calculating the weighted sum of the training physiological data low-dimensional embedded coding vector, the training drug data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector based on the sequence of information transmission dynamic energy space-time transmission weights to obtain a message transmission data low-dimensional embedded coding node characteristic transmission aggregate modulation vector as the node information fusion representation, and calculating the position summation of the message transmission data low-dimensional embedded coding node characteristic transmission aggregate modulation vector and the training drug concentration monitoring data low-dimensional embedded coding vector to obtain the pharmacokinetic timing coding vector.

Specifically, using the group pharmacokinetic model to perform pharmacokinetic coding based on node potential energy space-time information transfer on the training physiological data low-dimensional embedded coding vector, the training medication concentration monitoring data low-dimensional embedded coding vector and the training test data low-dimensional embedded coding vector according to the following characteristic message propagation formula so as to obtain the pharmacokinetic timing coding vector;

Wherein, the characteristic message propagation formula is: ;; ; ; ;;

wherein, A set of data low-dimensional embedded coding vectors consisting of the training physiological data low-dimensional embedded coding vectors, the training medication concentration monitoring data low-dimensional embedded coding vectors and the training test data low-dimensional embedded coding vectors,The training physiological data low-dimensional embedded coding vector, the training medication data low-dimensional embedded coding vector, the training test data low-dimensional embedded coding vector and the training medication concentration monitoring data low-dimensional embedded coding vector are respectively,The number of vectors of the set of encoded vectors is embedded for the data low dimension,Embedding the training physiological data into the coding vector in a low dimension, embedding the training medication data into the coding vector in a low dimension and embedding the training test data into the coding vector in a low dimensionThe individual data is embedded in the encoded vector in a low dimension,Embedding the training physiological data into the coding vector in a low dimension, embedding the training medication data into the coding vector in a low dimension and embedding the training test data into the coding vector in a low dimensionLow-dimensional embedded coding vector of dataThe characteristic value of the individual position is used,AndRespectively the firstThe characteristic mean and standard deviation of the individual data low-dimensional embedded encoding vectors,It is the calculation of the desired value that,The maximum value in the vector is represented,The characteristic static energy values are encoded for low-dimensional embedding of data,Representation ofAndThe coefficient of spatial span between them,Is thatAndAn energy spatial decay factor in between,AndRepresentation ofAndIs used for the time stamp of (a),A down-rounding operation is indicated,To control the inverse scale parameters of the time decay period,Is the function value of the natural index,Is thatAndAn energy time decay factor in between,AndIn order for the super-parameters to be trainable,Is thatRelative toIs a dynamic energy space-time transfer value of (1),In order for the masking operation to be performed,As a result of the threshold value being exceeded,Is thatThe function of the function is that,Vectors are encoded for the pharmacokinetic sequences.

The method comprises the steps of calculating the space distance span between the message transmission data low-dimensional embedded coding vector and the target data low-dimensional embedded coding vector to obtain the message transmission data low-dimensional embedded coding space span coefficient, and calculating the division between the data low-dimensional embedded coding feature static energy value corresponding to the message transmission data low-dimensional embedded coding vector and the message transmission data low-dimensional embedded coding space span coefficient to obtain the data low-dimensional embedded coding feature energy space attenuation factor.

The method comprises the steps of calculating a time stamp of the target data low-dimensional embedded coding vector and a time stamp of the message transmission data low-dimensional embedded coding vector, then rounding down an obtained time stamp value to obtain a data low-dimensional embedded coding characteristic time span coefficient, dividing the data low-dimensional embedded coding characteristic time span coefficient by an inverse scale parameter of a control time attenuation period, taking the obtained time span attenuation representation coefficient as an exponential power, calculating a natural exponential function value based on a natural constant e to obtain a time span attenuation class support representation coefficient, and calculating a division between a static energy value of the data low-dimensional embedded coding characteristic and the time span attenuation class support representation coefficient to obtain the data low-dimensional embedded coding characteristic energy time attenuation factor.

Wherein inputting the data low-dimensional embedded code feature dynamic energy space-time transfer values of the respective message-propagating data low-dimensional embedded code vectors into a mask function-based energy transfer gating module to obtain a sequence of the information transfer dynamic energy space-time transfer weights comprises, in the mask function-based energy transfer gating module, passing, in response to the data low-dimensional embedded code feature dynamic energy space-time transfer values being greater than a threshold superparameter, byAnd carrying out normalization processing on the data low-dimensional embedded coding characteristic dynamic energy space-time transfer values larger than the threshold super-parameters by a function to obtain information transfer dynamic energy space-time transfer weights.

Further, the pharmacokinetic timing code vector is decoded to obtain a predicted value of the blood concentration. In this way, the time sequence coding information of the pharmacokinetics can be utilized to predict the blood concentration and optimize the parameters of the group pharmacokinetic model, in this way, the parameters of the group pharmacokinetic model can be adjusted by utilizing an error feedback mechanism in a subsequent automatic iteration way, and the parameters in the group pharmacokinetic model are continuously adjusted to better fit the actually observed drug concentration monitoring data, so that the predicted blood concentration result is as close to an actual measurement value as possible, and the prediction accuracy of the group pharmacokinetic model is improved. The method not only improves the accuracy of the blood concentration prediction of the model, but also reduces the requirement of manual intervention, so that the establishment and updating of the group pharmacokinetic model are more efficient, and the method is favorable for supporting doctors to make more scientific and reasonable treatment decisions.

Preferably, inputting the pharmacokinetic timing encoding vector into a decoder-based blood concentration prediction module to obtain a predicted value of blood concentration includes:

Calculating the sum of absolute values of all characteristic values of the pharmacokinetic timing sequence coding vector to obtain a first pharmacokinetic timing sequence coding and modulation value, and calculating the square root of the square sum of all characteristic values of the pharmacokinetic timing sequence coding vector to obtain a second pharmacokinetic timing sequence coding and modulation value;

After the point subtraction is carried out on the pharmacokinetic time sequence coding vector and the second pharmacokinetic time sequence coding and modulating value, the point multiplication is respectively carried out on the pharmacokinetic time sequence coding vector with the number of the characteristic values and the reciprocal of the first pharmacokinetic time sequence coding and modulating value, and the reciprocal of each characteristic value is taken to obtain a first pharmacokinetic time sequence coding phase conversion vector;

The method comprises the steps of carrying out point subtraction on the first and second pharmacokinetic time sequence coding vectors and the first and second pharmacokinetic time sequence coding and modulation values, carrying out point multiplication on the first and second pharmacokinetic time sequence coding vectors and the inverse of the characteristic value number of the first and second pharmacokinetic time sequence coding and modulation values respectively, taking the inverse of each characteristic value to obtain a second pharmacokinetic time sequence coding phase conversion vector, carrying out point subtraction on the first and second pharmacokinetic time sequence coding phase conversion vectors and the point multiplication vector of a weighting super parameter to obtain an optimized pharmacokinetic time sequence coding vector, and inputting the optimized pharmacokinetic time sequence coding vector into a blood concentration prediction module based on a decoder to obtain a predicted value of blood concentration.

Here, the pharmacokinetic timing encoding vector is denoted asThe optimization of (c) is expressed as:;;;;;

wherein, Is each characteristic value of the pharmacokinetic timing coding vector,Is the first pharmacokinetic timing code and modulation value,Is the second pharmacokinetic timing code and modulation value,Is the coding vector of the pharmacokinetic time sequence,Represents a set of real numbers,Is the number of eigenvalues of the pharmacokinetic timing encoding vector,Is the inverse of the first pharmacokinetic timing code and modulation value,Is the inverse of each of the eigenvalues,Is the first pharmacokinetic timing code phase conversion vector,Is the inverse of the second pharmacokinetic timing code and modulation value,Is a second pharmacokinetic timing code phase conversion vector,Is a weighted super-parameter that is used to determine the weight of the object,Is an optimized pharmacokinetic timing sequence coding vector,Is subtracted according to the position point,Is multiplied by a position point.

More specifically, the decoder-based blood concentration prediction module is used to perform decoding regression on the optimized pharmacokinetic time series coding vector to obtain a predicted value of the blood concentration according to a decoding formulaWherein, the method comprises the steps of, wherein,Is the optimized pharmacokinetic timing coding vector,Is a predicted value of the blood concentration,Is a matrix of weights for the decoder,Representing a matrix multiplication.

That is, considering that the training physiological data low-dimensional embedded code vector, the training medication concentration monitoring data low-dimensional embedded code vector and the training inspection data low-dimensional embedded code vector respectively represent low-dimensional semantic embedded code features of physiological data, medication concentration monitoring data and inspection data, when the low-dimensional embedded code features are input into a characteristic message propagation network based on node potential energy space-time transfer characteristics, the pharmacokinetic timing code vector also has global propagation type aggregation difference based on node potential energy space-time transfer differences of each data, so that it is expected to improve the detailed semantic aggregation expression effect of the pharmacokinetic timing code vector based on the local feature propagation type aggregation difference of each sample under the global sample domain.

Based on the method, the difference of the characteristic value of the pharmacokinetic time sequence coding vector relative to the difference of the characteristic set and the modulation representation of the vector whole of the pharmacokinetic time sequence coding vector is used as semantic change intensity information, the phase-like conversion corresponding to the position-based intensity modulation is carried out through the difference and the modulation representation form, so that the aggregation enhancement of semantic change phase perception can promote the axial aggregation receptive field along the characteristic aggregation direction through carrying out the spatial translation operation based on alternate stacking under the vector set scale balance of the pharmacokinetic time sequence coding vector, the perception effect of the aggregation semantics of the pharmacokinetic time sequence coding vector on detail semantic change is promoted, the expression effect of the pharmacokinetic time sequence coding vector is promoted, and the accuracy of the blood concentration predicted value input by a blood concentration predicting module based on a decoder is promoted. In this way, the parameter of the group pharmacokinetic model is adjusted by utilizing an error feedback mechanism in an automatic iteration mode, and the parameter in the group pharmacokinetic model is continuously adjusted to better fit the actually observed drug concentration monitoring data, so that the predicted blood concentration result is as close to an actual measurement value as possible, and the prediction accuracy of the group pharmacokinetic model is improved.

In summary, by adopting the above scheme, in the process of generating the predicted value of the blood concentration by using the group pharmacokinetic model, physiological data, medication data, drug concentration monitoring data and inspection data of a training patient are input into a data processing and analyzing algorithm based on artificial intelligence and deep learning to be processed, so that physiological data low-dimensional embedded coding features, medication data low-dimensional embedded coding features, drug concentration monitoring data low-dimensional embedded coding features and inspection data low-dimensional embedded coding features are captured, and information transfer semantics between different parameter data features relative to the drug detection embedded coding features based on node potential energy space-time are extracted from the features by using the group pharmacokinetic model to reflect time sequence coding information of the pharmacokinetics, so that the prediction of the blood concentration and the parameter optimization of the group pharmacokinetic model are performed. Therefore, the parameters of the group pharmacokinetic model can be adjusted by utilizing an error feedback mechanism in an automatic iteration mode, and the parameters in the group pharmacokinetic model are continuously adjusted to better fit the actually observed drug concentration monitoring data, so that the predicted blood drug concentration result is as close to an actual measured value as possible, and the prediction accuracy of the group pharmacokinetic model is improved.

FIG. 3 is a block diagram illustrating an automated optimization system for individualized medication mathematical model parameters, according to an exemplary embodiment. As shown in fig. 3, the system 200 includes:

a data acquisition module 201 for acquiring physiological data, medication concentration monitoring data, and test data of a training patient subject;

A data processing module 202, configured to process the physiological data, the medication concentration monitoring data, and the test data of the trained patient object using a group pharmacokinetic model to obtain a predicted value of the blood concentration;

The model parameter training module 203 is configured to automatically optimize model parameters of the group pharmacokinetic model based on error information between the predicted blood concentration value and the actual measured blood concentration value to obtain a trained group pharmacokinetic model;

And the prediction result generating module 204 is configured to process the physiological data, the medication data, the drug concentration monitoring data and the test data of the test patient object by using the trained group pharmacokinetic model to obtain a prediction result, where the prediction result is a prediction value of the blood concentration of the test patient object.

In one embodiment of the application, the data processing module comprises a physiological data embedding encoding unit, a medication concentration monitoring data embedding encoding unit and a prediction and dynamic encoding unit, wherein the physiological data embedding encoding unit is used for embedding and encoding physiological data of a training patient object by using a physiological data embedding encoding matrix to obtain a training physiological data low-dimensional embedding encoding vector, the medication data embedding encoding unit is used for embedding and encoding medication data of the training patient object by using a medication data embedding encoding matrix to obtain a training medication concentration monitoring data low-dimensional embedding encoding vector, the examination data embedding encoding unit is used for embedding and encoding examination data of the training patient object by using a examination data embedding encoding matrix to obtain a training examination data low-dimensional embedding encoding vector, and the medication and dynamic encoding unit is used for decoding the training physiological data low-dimensional embedding encoding vector, the training medication concentration monitoring vector and the training medication concentration monitoring data low-dimensional embedding encoding vector and the training medication vector by using the group medication and dynamic model to obtain a prediction and dynamic encoding vector when the prediction and dynamic encoding unit decodes the training medication vector.

In another embodiment of the present application, there is provided a computer software system for automatic optimization of personalized medicine mathematical model parameters, the system comprising:

(1) The acquisition unit is used for collecting physiological data, medication data, drug concentration monitoring data and checking data of the patient;

(2) The prediction unit predicts the blood concentration by utilizing the data acquired by the acquisition unit based on the existing group pharmacokinetic model;

(3) The learning unit is used for determining a predicted error source by learning difference information between the predicted blood concentration and the actually measured blood concentration based on a machine learning algorithm, or automatically optimizing group pharmacokinetic model parameters and correcting model prediction accuracy;

(4) The diagnosis unit automatically gives a first-time medication scheme, an adjustment medication scheme, a medical record-like medication scheme and the like according to the patient data obtained by the acquisition unit and the prediction result of the prediction unit;

(5) And the database unit is used for maintaining the related data and providing related data inquiry.

The group pharmacokinetics model is a one-chamber model, a two-chamber model or a combination of the one-chamber model and the two-chamber model. The machine learning algorithm used is a Bayes maximum a posteriori estimation algorithm (MAP), or a reinforcement learning algorithm, or a fusion algorithm of maximum a posteriori estimation and reinforcement learning is used.

The acquiring unit can be a medical record input interface in software, and also has a data access interface of a hospital HIS system and an LIS system, and is used for automatically acquiring treatment information of the HIS system and the LIS system.

The model used in the prediction unit uses different model parameters for different patient groups, provides multiple sets of model parameters for the same patient group, and can automatically judge the effect of the multiple sets of model parameters.

Referring now to fig. 4, a schematic diagram of an electronic device 600 suitable for use in implementing embodiments of the present application is shown. The terminal device in the embodiment of the present application may include, but is not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), car terminals (e.g., car navigation terminals), and the like, and stationary terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 4 is only an example and should not be construed as limiting the functionality and scope of use of the embodiments of the application.

As shown in fig. 4, the electronic device 600 may include a processing means (e.g., a central processor, a graphic processor, etc.) 601, which may perform various appropriate actions and processes according to a program stored in a ROM (read only memory) 602 or a program loaded from a storage means 608 into a RAM (random access memory) 603. In the RAM603, various programs and data required for the operation of the electronic apparatus 600 are also stored. The processing device 601, the ROM602, and the RAM603 are connected to each other through a bus 604. An I/O (input/output) interface 605 is also connected to the bus 604.

In general, devices may be connected to I/O interface 605 including input devices 606, including for example, touch screens, touch pads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc., output devices 607, including for example, liquid Crystal Displays (LCDs), speakers, vibrators, etc., storage devices 608, including for example, magnetic tape, hard disk, etc., and communication devices 609. The communication means 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. While fig. 4 shows an electronic device 600 having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead.

In particular, according to embodiments of the present application, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer program embodied on a non-transitory computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via communication means 609, or from storage means 608, or from ROM 602. The above-described functions defined in the method of the embodiment of the present application are performed when the computer program is executed by the processing means 601.

The computer readable medium of the present application may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of a computer-readable storage medium may include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present application, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to electrical wiring, fiber optic cable, RF (radio frequency), and the like, or any suitable combination of the foregoing.

In some embodiments, the clients, servers may communicate using any currently known or future developed network protocol, such as HTTP (HyperText TransferProtocol ), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the internet (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed networks.

The computer readable medium may be included in the electronic device or may exist alone without being incorporated into the electronic device.

Computer program code for carrying out operations of the present application may be written in one or more programming languages, including, but not limited to, an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The modules involved in the embodiments of the present application may be implemented in software or in hardware. The name of the module is not limited to the module itself in some cases, and for example, the test parameter obtaining module may also be described as "a module for obtaining the device test parameter corresponding to the target device".

The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic that may be used include Field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems-on-a-chip (SOCs), complex Programmable Logic Devices (CPLDs), and the like.

In the context of the present application, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

FIG. 5 is an application scenario diagram illustrating an automatic optimization method for personalized medicine mathematical model parameters, according to an exemplary embodiment. As shown in fig. 5, in this application scenario, first, physiological data (e.g., C1 as illustrated in fig. 5), medication data (e.g., C2 as illustrated in fig. 5), medication concentration monitoring data (e.g., C3 as illustrated in fig. 5), and test data (e.g., C4 as illustrated in fig. 5) of a trained patient subject are acquired, and then, the acquired physiological data, medication concentration monitoring data, and test data are input to a server (e.g., S as illustrated in fig. 5) of an automatic optimization algorithm deployed with personalized medication mathematical model parameters, wherein the server is capable of processing the physiological data, medication concentration monitoring data, and test data based on the automatic optimization algorithm of personalized medication mathematical model parameters to obtain a prediction result, which is a prediction value of the blood concentration of the tested patient subject.

The above description is only illustrative of the preferred embodiments of the present application and of the principles of the technology employed. It will be appreciated by persons skilled in the art that the scope of the disclosure referred to in the present application is not limited to the specific combinations of technical features described above, but also covers other technical features formed by any combination of the technical features described above or their equivalents without departing from the spirit of the disclosure. Such as the above-mentioned features and the technical features disclosed in the present application (but not limited to) having similar functions are replaced with each other.

Moreover, although operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order. In certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the application. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.

Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are exemplary forms of implementing the claims. The specific manner in which the various modules perform the operations in the apparatus of the above embodiments have been described in detail in connection with the embodiments of the method, and will not be described in detail herein.

Claims

1. A method for automatically optimizing parameters of a mathematical model for personalized medication, comprising:

Obtaining physiological data, medication data, drug concentration monitoring data, and test data of training patient subjects;

Using a population pharmacokinetic model to process the physiological data, medication data, drug concentration monitoring data and test data of the training patient subject to obtain a predicted value of blood drug concentration;

Based on the error information between the predicted value of the blood drug concentration and the actual measured blood drug concentration value, automatically optimizing the model parameters of the population pharmacokinetic model to obtain a trained population pharmacokinetic model;

Using the trained population pharmacokinetic model to process the physiological data, medication data, drug concentration monitoring data and test data of the test patient subject to obtain a prediction result, wherein the prediction result is a predicted value of the blood drug concentration of the test patient subject;

The method of using a population pharmacokinetic model to process the physiological data, medication data, drug concentration monitoring data and test data of the training patient subjects to obtain a predicted value of blood drug concentration includes:

Embedding the physiological data of the training patient object using a physiological data embedding coding matrix to obtain a low-dimensional embedding coding vector of the training physiological data;

Using a medication data embedding coding matrix to embed the medication data of the training patient object to obtain a low-dimensional embedding coding vector for the training medication data;

Embedding the drug concentration monitoring data of the training patient object using a drug concentration monitoring data embedding coding matrix to obtain a low-dimensional embedding coding vector of the training drug concentration monitoring data;

Embedding the test data of the training patient object using a test data embedding coding matrix to obtain a low-dimensional embedding coding vector for the training test data;

Using the population pharmacokinetic model, the training physiological data low-dimensional embedded coding vector, the training medication data low-dimensional embedded coding vector, the training drug concentration monitoring data low-dimensional embedded coding vector, and the training test data low-dimensional embedded coding vector are pharmacokinetic encoded to obtain a pharmacokinetic time series coding vector;

Decoding the pharmacokinetic time series coding vector to obtain a predicted value of the blood drug concentration;

The step of inputting the pharmacokinetic time sequence coding vector into a decoder-based blood drug concentration prediction module to obtain a predicted value of blood drug concentration includes:

Calculating the sum of the absolute values of each eigenvalue of the pharmacokinetic timing coding vector to obtain a first pharmacokinetic timing coding and modulation value, and calculating the square root of the sum of the squares of each eigenvalue of the pharmacokinetic timing coding vector to obtain a second pharmacokinetic timing coding and modulation value;

After performing dot-wise subtraction on the pharmacokinetic timing coding vector and the second pharmacokinetic timing coding and modulation value, respectively perform dot-wise multiplication on the number of eigenvalues of the pharmacokinetic timing coding vector and the reciprocal of the first pharmacokinetic timing coding and modulation value, and take the reciprocal of each eigenvalue to obtain a first pharmacokinetic timing coding phase conversion vector;

After point-wise subtraction of the pharmacokinetic timing coding vector and the first pharmacokinetic timing coding and modulation value, point-wise multiplication is performed with the square root of the number of eigenvalues of the pharmacokinetic timing coding vector and the inverse of the second pharmacokinetic timing coding and modulation value, and the inverse of each eigenvalue is taken to obtain the second pharmacokinetic timing coding phase transition vector; point-wise subtraction of the first pharmacokinetic timing coding phase transition vector by the point product vector of the second pharmacokinetic timing coding phase transition vector and the weighted hyperparameter is performed to obtain an optimized pharmacokinetic timing coding vector; the optimized pharmacokinetic timing coding vector is input into a decoder-based blood drug concentration prediction module to obtain a predicted value of blood drug concentration.

2. The method for automatically optimizing the parameters of the mathematical model of personalized medication according to claim 1, characterized in that the physiological data of the training patient subjects include the underlying diseases, age, gender, weight, height, body surface area, concomitant medications, genotype, liver function index, kidney function index and blood routine of the training patient subjects, the medication data of the training patient subjects include administration time, dosage, administration route, administration frequency and infusion time, the drug concentration monitoring data of the training patient subjects include monitoring time, blood drug concentration value, detection method and instrument type, and the test data include liver and kidney function test data, blood routine test data, and infection index test data.

3. The method for automatically optimizing the parameters of the mathematical model of personalized medication according to claim 2 is characterized in that the low-dimensional embedded coding vector of the training physiological data, the low-dimensional embedded coding vector of the training medication data, the low-dimensional embedded coding vector of the training drug concentration monitoring data and the low-dimensional embedded coding vector of the training test data are pharmacokinetic encoded using the population pharmacokinetic model to obtain a pharmacokinetic time series coding vector, including: using the population pharmacokinetic model to perform pharmacokinetic encoding based on node potential spatiotemporal information transfer on the low-dimensional embedded coding vector of the training physiological data, the low-dimensional embedded coding vector of the training medication data, the low-dimensional embedded coding vector of the training drug concentration monitoring data and the low-dimensional embedded coding vector of the training test data to obtain the pharmacokinetic time series coding vector.

4. The method for automatically optimizing parameters of a mathematical model for personalized medication according to claim 3, characterized in that the population pharmacokinetic model is used to perform pharmacokinetic coding based on node potential spatiotemporal information transfer on the training physiological data low-dimensional embedded coding vector, the training medication data low-dimensional embedded coding vector, the training drug concentration monitoring data low-dimensional embedded coding vector, and the training test data low-dimensional embedded coding vector to obtain the pharmacokinetic time series coding vector, comprising:

The training drug concentration monitoring data low-dimensional embedding coding vector is used as the target data low-dimensional embedding coding vector, and the training physiological data low-dimensional embedding coding vector, the training medication data low-dimensional embedding coding vector and the training inspection data low-dimensional embedding coding vector are used as a sequence of message propagation data low-dimensional embedding coding vectors, and the information transmission dynamic energy spatiotemporal transfer weight of each message propagation data low-dimensional embedding coding vector in the sequence of message propagation data low-dimensional embedding coding vectors relative to the target data low-dimensional embedding coding vector is calculated to obtain a sequence of information transmission dynamic energy spatiotemporal transfer weights;

Based on the sequence of information transfer dynamic energy spatiotemporal transfer weights, the training physiological data low-dimensional embedded coding vector, the training medication data low-dimensional embedded coding vector and the training inspection data low-dimensional embedded coding vector are fused, and then the obtained node information fusion representation is passed to the training drug concentration monitoring data low-dimensional embedded coding vector to obtain the pharmacokinetic timing coding vector.

5. The method for automatically optimizing the parameters of the mathematical model of personalized medication according to claim 4, characterized in that the information transfer dynamic energy spatiotemporal transfer weights of each message propagation data low-dimensional embedded coding vector in the sequence of message propagation data low-dimensional embedded coding vectors relative to the target data low-dimensional embedded coding vector are calculated to obtain a sequence of information transfer dynamic energy spatiotemporal transfer weights, comprising:

Calculating a data low-dimensional embedding coding feature static energy value of each message propagation data low-dimensional embedding coding vector in the sequence of message propagation data low-dimensional embedding coding vectors;

Determining a data low-dimensional embedding coding feature energy spatial attenuation factor based on a spatial span coefficient between the low-dimensional embedding coding vectors of the respective message propagation data and the low-dimensional embedding coding vectors of the target data;

Determine a data low-dimensional embedding coding feature energy time decay factor based on a time span coefficient between the low-dimensional embedding coding vector of each message propagation data and the low-dimensional embedding coding vector of the target data;

Based on the data low-dimensional embedding coding feature energy space decay factor and the data low-dimensional embedding coding feature energy time decay factor, the data low-dimensional embedding coding feature static energy value of each message propagation data low-dimensional embedding coding vector is temporally modulated to obtain the data low-dimensional embedding coding feature dynamic energy space-time transfer value of each message propagation data low-dimensional embedding coding vector relative to the target data low-dimensional embedding coding vector;

The data low-dimensional embedding coding feature dynamic energy spatiotemporal transfer value of each message propagation data low-dimensional embedding coding vector is input into the energy transfer gating module based on the mask function to obtain the sequence of the information transfer dynamic energy spatiotemporal transfer weights.

6. The method for automatically optimizing the parameters of the mathematical model of personalized medication according to claim 5, characterized in that the static energy value of the low-dimensional embedding coding feature of each message propagation data low-dimensional embedding coding vector in the sequence of message propagation data low-dimensional embedding coding vectors is calculated, comprising:

Respectively calculating the mean and standard deviation of the low-dimensional embedding coding vector of the message propagation data to obtain the mean of the low-dimensional embedding coding feature of the message propagation data and the standard deviation of the low-dimensional embedding coding of the message propagation data;

After calculating the fourth power of the positional difference between the low-dimensional embedding coding vector of the message propagation data and the mean value of the low-dimensional embedding coding feature of the message propagation data, the expected value of the fourth-power modulated low-dimensional embedding coding offset vector of the message propagation data is calculated to obtain the expected factor of the data low-dimensional embedding coding feature offset;

After calculating the division between the expected factor of the data low-dimensional embedding coding feature offset and the fourth power of the standard deviation of the message propagation data low-dimensional embedding coding, it is added to the maximum eigenvalue in the message propagation data low-dimensional embedding coding vector to obtain the data low-dimensional embedding coding feature static energy value.

7. The method for automatically optimizing the parameters of the mathematical model of personalized medication according to claim 6, characterized in that, based on the sequence of the information transfer dynamic energy spatiotemporal transfer weights, the low-dimensional embedded coding vector of the training physiological data, the low-dimensional embedded coding vector of the training medication data and the low-dimensional embedded coding vector of the training inspection data are fused, and then the obtained node information fusion representation is transmitted to the low-dimensional embedded coding vector of the training drug concentration monitoring data to obtain the pharmacokinetic time series coding vector, comprising:

Based on the sequence of the information transfer dynamic energy spatiotemporal transfer weights, a weighted sum of the training physiological data low-dimensional embedding coding vector, the training medication data low-dimensional embedding coding vector and the training inspection data low-dimensional embedding coding vector is calculated to obtain a message propagation data low-dimensional embedding coding node feature transfer aggregation modulation vector as the node information fusion representation;

The position-wise sum of the message propagation data low-dimensional embedded coding node feature transfer aggregation modulation vector and the training drug concentration monitoring data low-dimensional embedded coding vector is calculated to obtain the pharmacokinetic time series coding vector.

8. An automatic optimization system for parameters of a mathematical model for personalized medication, characterized by comprising:

A data acquisition module, used to acquire physiological data, medication data, drug concentration monitoring data and test data of training patient subjects;

A data processing module, for processing the physiological data, medication data, drug concentration monitoring data and test data of the training patient subject using a population pharmacokinetic model to obtain a predicted value of blood drug concentration;

A model parameter training module, for automatically optimizing the model parameters of the population pharmacokinetic model based on the error information between the predicted value of the blood drug concentration and the actual measured blood drug concentration value to obtain a trained population pharmacokinetic model;

A prediction result generating module, used to process the physiological data, medication data, drug concentration monitoring data and test data of the test patient subject using the trained population pharmacokinetic model to obtain a prediction result, wherein the prediction result is a predicted value of the blood drug concentration of the test patient subject;

Wherein, the data processing module includes:

A physiological data embedding coding unit, configured to embed the physiological data of the training patient object using a physiological data embedding coding matrix to obtain a low-dimensional embedding coding vector of the training physiological data;

A medication data embedding coding unit, used for embedding and coding the medication data of the training patient object using a medication data embedding coding matrix to obtain a low-dimensional embedding coding vector for the training medication data;

A drug concentration monitoring data embedding coding unit, used to embed the drug concentration monitoring data of the training patient object using a drug concentration monitoring data embedding coding matrix to obtain a low-dimensional embedding coding vector of the training drug concentration monitoring data;

A test data embedding coding unit, used for embedding coding the test data of the training patient object using a test data embedding coding matrix to obtain a low-dimensional embedding coding vector of the training test data;

a pharmacokinetic encoding unit, configured to use the population pharmacokinetic model to perform pharmacokinetic encoding on the training physiological data low-dimensional embedded coding vector, the training medication data low-dimensional embedded coding vector, the training drug concentration monitoring data low-dimensional embedded coding vector, and the training test data low-dimensional embedded coding vector to obtain a pharmacokinetic time series coding vector;

A vector decoding unit, used for decoding the pharmacokinetic time-series coding vector to obtain a predicted value of the blood drug concentration;