A remote medical health management system based on the Internet

Abstract

The present application provides an Internet-based telemedicine health management system, which relates to the field of intelligent health management. It first collects basic user information including user age, gender, disease type and health status, and then analyzes the basic user information in combination with symptom information in the database and sample information of healthy people to calculate subjective and objective state estimation values, and finally combines these two estimation values ​​to generate health status values ​​and feeds them back to medical staff to support making appropriate medical decisions. In this way, it helps to achieve more accurate and personalized health management, and then provide high-quality telemedicine services.

Description

Technical Field

[0001] The present application relates to the field of intelligent health management, and more specifically, to an Internet-based remote medical health management system. Background Art

[0002] In recent years, with the rapid development of Internet technology and mobile communication technology, telemedicine services have been rapidly promoted. Especially in remote areas and places with insufficient medical resources, telemedicine has become an effective means to solve the problem of unequal medical services.

[0003] Patent CN117637091A proposes a remote medical health management system, which includes three ports: information collection, analysis and feedback. It is used to collect basic information and real-time physiological data of users, and evaluate the health status value and rehabilitation expectation value of users by comparing the health standards and rehabilitation cases in the database, so as to assist medical staff in optimizing treatment plans.

[0004] In the above patent, the subjective state value is calculated by a specific formula, which takes into account the number of characters in the user description that match the symptom information text, the total number of characters in the symptom information text, and the number of punctuation marks in the text. However, directly using character matching to calculate the subjective state value tends to ignore the complexity and ambiguity of language, which may lead to misunderstanding of the user's true intentions. For example, "dizziness" may refer to mild discomfort or severe dizziness, and simple character matching cannot distinguish such semantic nuances. In addition, although the use of specific formulas can provide quantitative indicators, due to the high personalization of health perceptions, the formula may be too general and unable to fully reflect individual differences and the complexity of specific diseases, resulting in inaccurate evaluation results or low correlation, affecting the reliability and validity of the subjective state estimate.

[0005] Therefore, an optimized telemedicine health management system is desired. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a remote medical health management system based on the Internet.

[0007] According to one aspect of the present application, there is provided an Internet-based remote medical health management system, which includes:

[0008] A user information collection module is used to collect basic user information, wherein the basic user information includes age, gender, disease type and health status; a user information storage module is used to store the basic user information; a database module is used to store symptom information of various disease types and sample information of healthy people; a subjective analysis module is used to obtain a subjective state estimation value based on the disease type in the basic user information; an objective analysis module is used to obtain an objective state estimation value based on the health status in the basic user information; a feedback module is used to obtain a health status value based on the subjective state estimation value and the objective state estimation value, and feed the health status value back to medical staff; wherein the subjective analysis module includes:

[0009] A voice description signal acquisition unit, used to acquire a voice description signal of a user's physical condition;

[0010] A body condition semantic coding unit, used for performing speech recognition and semantic coding on the user's speech description signal of the body condition to obtain semantic features of the user's subjective description of the body condition;

[0011] A disease information encoding unit, for extracting a symptom information text matching the disease type based on the disease type in the basic information of the user, and performing sentence processing and semantic encoding on the symptom information text to obtain a sequence of semantic features of the disease information sentence granularity description;

[0012] The user information query encoding and decoding unit is used to perform query encoding optimization under the guidance of a key matrix based on a sequence of semantic features of the subjective description of the user's physical state and semantic features of the granular description of the disease information sentence to obtain a query response representation of the subjective description of the physical state - the objective description of the disease, and obtain the subjective state estimation value.

[0013] In the above-mentioned Internet-based remote medical health management system, the physical condition semantic encoding unit includes: a physical condition voice description signal recognition subunit, which is used to perform voice recognition on the user's voice description signal of the physical condition to obtain a voice-to-text conversion description of the user's physical condition; a physical condition description text semantic encoding subunit, which is used to perform semantic encoding on the user's physical condition voice-to-text conversion description to obtain a semantic encoding vector of the user's physical condition subjective description as the semantic feature of the user's physical condition subjective description.

[0014] In the above-mentioned Internet-based remote medical health management system, the disease information encoding unit includes: a symptom information text extraction subunit, which is used to extract the symptom information text matching the disease type in the database module based on the disease type in the user basic information; a symptom information text sentence processing subunit, which is used to perform sentence processing on the symptom information text to obtain a sequence of disease information sentence granularity text descriptions; a disease information short text semantic encoding subunit, which is used to perform semantic encoding on each disease information sentence granularity text description in the sequence of disease information sentence granularity text descriptions to obtain a sequence of disease information sentence granularity description semantic encoding vectors as a sequence of disease information sentence granularity description semantic features.

[0015] In the above-mentioned Internet-based remote medical health management system, the user information query encoding and decoding unit includes: a user physical state and symptom information query encoding optimization subunit, which is used to perform query coding optimization on the sequence of the user's physical state subjective description semantic coding vector and the symptom information sentence granularity description semantic coding vector under the guidance of a key matrix to obtain the physical state subjective description-symptom objective description query response representation; a subjective state estimation value generation subunit, which is used to obtain the subjective state estimation value based on the physical state subjective description-symptom objective description query response representation.

[0016] In the above-mentioned Internet-based remote medical health management system, the user physical condition and symptom information query coding optimization subunit includes: a symptom information sentence granularity description linear transformation secondary subunit, which is used to use a key embedding matrix to linearly embed the individual symptom information sentence granularity description semantic coding vectors in the sequence of the symptom information sentence granularity description semantic coding vectors to obtain a sequence of symptom information sentence granularity description semantic coding vectors after linear transformation; a symptom information sentence granularity description semantic key matrix construction secondary subunit, which is used to matrix-arrange the sequence of symptom information sentence granularity description semantic coding vectors after linear transformation to obtain a symptom information sentence granularity description semantic key matrix; The information prompt template selects a secondary sub-unit, which is used to extract the maximum value of the symptom information sentence granularity description semantic coding vector after each linear transformation in the symptom information sentence granularity description semantic key matrix to obtain the significant feature vector of the symptom information sentence granularity description semantic key matrix as the symptom information prompt template; the user body state symptom information domain optimization coding secondary sub-unit is used to perform cross-domain optimization coding on the symptom information prompt template, the user body state subjective description semantic coding vector and the symptom information sentence granularity description semantic key matrix to obtain the body state subjective description-symptom objective description query response coding vector as the body state subjective description-symptom objective description query response representation.

[0017] In the above-mentioned Internet-based remote medical health management system, the second-level sub-unit of the linear transformation of the symptom information sentence granularity description is used to: perform vector-matrix multiplication of the symptom information sentence granularity description semantic coding vector and the key embedding matrix, and add the obtained symptom information sentence granularity description semantic embedding vector and the key bias vector by position to obtain the symptom information sentence granularity description semantic coding vector after the linear transformation.

[0018] In the above-mentioned Internet-based remote medical health management system, the user's physical state and symptom information domain optimization coding secondary subunit includes: a user's physical state subjective description semantic embedding coding tertiary subunit, which is used to use a query embedding matrix and a value embedding matrix to linearly embed the user's physical state subjective description semantic coding vector to obtain a user's physical state subjective description semantic query vector and a user's physical state subjective description semantic value vector; a physical state subjective description-symptom objective description cross-domain optimization tertiary subunit, which is used to embed the user's physical state subjective description semantic query vector and the user's physical state subjective description semantic value vector into a user's physical state subjective description semantic query vector and a user's physical state subjective description semantic value vector. The meaning value vector, the semantic coding vector of the symptom information sentence granularity description after each linear transformation in the semantic key matrix of the symptom information sentence granularity description, and the symptom information prompt template are respectively input into the heterogeneous converter structure based on template prompt optimization to obtain a sequence of cross-domain optimized query coding vectors of subjective description of body state-objective description of symptom; the query response coding vector of subjective description of body state-objective description of symptom generates a three-level sub-unit, which is used to calculate the positional mean vector of the sequence of cross-domain optimized query coding vectors of subjective description of body state-objective description of symptom to obtain the query response coding vector of subjective description of body state-objective description of symptom.

[0019] In the above-mentioned Internet-based remote medical health management system, the third-level sub-unit of the semantic embedding coding of the subjective description of the user's physical state is used to: multiply the semantic coding vector of the subjective description of the user's physical state with the query embedding matrix and then add it positionally with the query bias vector to obtain the semantic query vector of the subjective description of the user's physical state; multiply the semantic coding vector of the subjective description of the user's physical state with the value embedding matrix and then add it positionally with the value bias vector to obtain the semantic value vector of the subjective description of the user's physical state.

[0020] In the above-mentioned Internet-based remote medical health management system, the three-level sub-unit of cross-domain optimization of subjective description of physical state and objective description of symptoms is used to: multiply the semantic query vector of the user's subjective description of physical state with the transposed vector of the semantic coding vector of the symptom information sentence granularity description after the linear transformation, and divide the obtained user body-symptom information semantic association matrix by the bi-norm of the symptom information prompt template by position to obtain the user body-symptom information semantic association modulation matrix; use the softmax function to process the user body-symptom information semantic association modulation matrix, and perform matrix-vector multiplication of the obtained user body-symptom information semantic association activation matrix and the symptom information prompt template to obtain the user body-symptom information semantic weight vector; calculate the positional point multiplication of the user body-symptom information semantic weight vector and the user's body state subjective description semantic value vector to obtain the cross-domain optimized query coding vector of the subjective description of physical state-objective description of symptoms.

[0021] In the above-mentioned Internet-based remote medical health management system, the subjective state estimation value generation subunit is used to: input the subjective description of the body state-objective description of the condition query response coding vector into the decoder-based subjective state estimation module to obtain the subjective state estimation value.

[0022] Compared with the prior art, the Internet-based telemedicine health management system provided by the present application adopts artificial intelligence-based speech processing and data analysis technology to perform speech recognition and semantic coding on the user's speech description signal of the physical condition, and extracts the matching information text from the database module based on the type of illness in the basic information of the user, and then performs sentence processing and semantic coding on the matching symptom information text, so as to intelligently obtain the subjective state estimation value according to the query response characteristics between the semantic features of the subjective description of the user's physical state and the semantic features of the granular description of the symptom information sentence. In this way, the physical condition information expressed by the user can be analyzed more deeply, so as to more accurately understand the user's actual intention. In addition, the semantic coding technology can capture the unique characteristics of the individual based on the user's specific description, thereby providing a more personalized evaluation result, which is better adapted to the description style and health status of different users. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0024] Figure 14 is a system block diagram of an Internet-based remote medical health management system according to an embodiment of the present application.

[0025] Figure 2 4 is a block diagram of a subjective analysis module in an Internet-based remote medical health management system according to an embodiment of the present application.

[0026] Figure 3 Schematic diagram of data flow of a subjective analysis module in an Internet-based remote medical health management system according to an embodiment of the present application.

[0027] Figure 4 It is a block diagram of a physical condition semantic encoding unit in an Internet-based remote medical health management system according to an embodiment of the present application.

[0028] Figure 5 It is a block diagram of a disease information encoding unit in an Internet-based remote medical health management system according to an embodiment of the present application.

[0029] Figure 6 The present invention is a block diagram of a user information query encoding and decoding unit in an Internet-based remote medical health management system according to an embodiment of the present application. DETAILED DESCRIPTION

[0030] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0031] In particular, in this application, all actions to obtain signals, information or data are carried out in compliance with the relevant data protection laws and policies of the country where they are located and with the authorization given by the owner of the corresponding device.

[0032] With the rapid development of the Internet and mobile communication technologies, telemedicine services have gradually become popular, especially in remote areas where medical resources are scarce. Telemedicine has become a key way to solve the problem of unbalanced medical resources.

[0033] Patent CN117637091A proposes a remote medical health management system, which consists of three ports: information collection, analysis and feedback. The system aims to collect basic information and real-time physiological data of users, and compare them with health standards and rehabilitation cases stored in the database to evaluate the health status value and rehabilitation expectation of users, thereby assisting medical personnel in optimizing treatment plans.

[0034] The above patent proposes a specific formula for calculating the subjective state value, which takes into account the number of characters in the user description that match the symptom information text, the total number of characters in the symptom information text, and the number of punctuation marks in the text. However, this character matching-based method may ignore the complexity and ambiguity of language, leading to misunderstanding of the user's true intentions. For example, "dizziness" may mean mild discomfort or severe dizziness, and simple character matching cannot capture such subtle semantic differences. In addition, although the specific formula provides a quantitative health indicator, due to the high degree of personalization of health perception, the formula may be overly simplified and fail to fully consider individual differences and the complexity of specific diseases, which may reduce the accuracy and relevance of the evaluation results, thereby affecting the reliability and validity of the subjective state estimation value.

[0035] Based on this, this application proposes an Internet-based remote medical health management system. Figure 1 FIG. 1 is a system block diagram of an Internet-based remote medical health management system according to an embodiment of the present application. Figure 1 As shown, in the Internet-based remote medical health management system 100, it includes: a user information collection module 110, which is used to collect basic user information, and the user basic information includes age, gender, disease type and health status; a user information storage module 120, which is used to store the user basic information; a database module 130, which is used to store symptom information of various disease types and sample information of healthy people; a subjective analysis module 140, which is used to obtain a subjective state estimation value based on the disease type in the user basic information; an objective analysis module 150, which is used to obtain an objective state estimation value based on the health status in the user basic information; a feedback module 160, which is used to obtain a health status value based on the subjective state estimation value and the objective state estimation value, and feed the health status value back to medical staff.

[0036] In the embodiment of the present application, the user information collection module 110 is used to collect basic user information, which includes age, gender, disease type and health status. It should be understood that age is a key factor in assessing health status, and the health problems and disease risks of people of different age groups are different; gender differences can affect the incidence of certain diseases, such as sex-linked genetic diseases, and gender information helps to provide a more accurate health assessment for users; the disease type is a direct factor in assessing the health status of users, mainly including the user's voice description of the physical condition, which is often an important basis for medical diagnosis; the health status specifically includes objective information such as the user's heart rate information, blood pressure information, and body temperature information. In general, the user's age, gender, disease type and health status together constitute a comprehensive user health profile, and the comprehensive evaluation of this information helps to more accurately calculate the user's health status value, thereby helping medical staff to better understand the user's health status, thereby providing more accurate and personalized medical services. In particular, in the embodiment of the present application, the user can enter some basic information (such as age, gender and disease type) on the user page of the system through a device connected to the Internet, and obtain the user's health status through a smart wearable device, and automatically upload the user's health status information to the system through the network.

[0037] In the embodiment of the present application, the user information storage module 120 is used to store the basic information of the user. It should be understood that by storing the basic information of the user in the database of the system, it is convenient for the algorithm to analyze it to identify the health risks of the user and provide relevant medical services. At the same time, medical staff can quickly access the basic information of the user, which is convenient for reviewing the health status and health assessment of the user, which is conducive to saving time and improving the efficiency of medical services.

[0038] In the embodiment of the present application, the database module 130 is used to store symptom information of various disease types and sample information of healthy people. It should be understood that storing symptom information of various disease types in the database enables the subsequent system to quickly identify the symptoms described by the user and match them with the known symptoms in the database. This matching can help to make a preliminary health diagnosis faster. Moreover, by analyzing the sample information of healthy people stored in the database, the system can establish a baseline of normal health status, which provides a reference standard for the subsequent calculation of objective state estimation values, and thus can better judge the health status of the user.

[0039] In the embodiment of the present application, the subjective analysis module 140 is used to obtain a subjective state estimation value based on the disease type in the basic information of the user.

[0040] Correspondingly, in the subjective analysis module, the technical concept of the present application is to use artificial intelligence-based speech processing and data analysis technology to perform speech recognition and semantic coding on the user's speech description signal of the physical condition, and based on the disease type in the user's basic information, extract the matching information text from the database module, and then perform sentence processing and semantic coding on the matching symptom information text, so as to intelligently obtain the subjective state estimation value based on the query response characteristics between the semantic features of the user's subjective description of the physical state and the semantic features of the granular description of the symptom information sentence. In this way, the physical condition information expressed by the user can be analyzed more deeply, so as to more accurately understand the user's actual intention. In addition, the semantic coding technology can capture the unique characteristics of the individual based on the user's specific description, thereby providing a more personalized evaluation result, which is better adapted to the description style and health status of different users.

[0041] Specifically, Figure 2 4 is a block diagram of a subjective analysis module in an Internet-based remote medical health management system according to an embodiment of the present application. Figure 3 FIG. 1 is a schematic diagram of data flow of a subjective analysis module in an Internet-based remote medical health management system according to an embodiment of the present application. Figure 2 and Figure 3 As shown, the subjective analysis module 140 includes: a speech description signal acquisition unit 141, which is used to acquire the user's speech description signal of the physical condition; a physical condition semantic encoding unit 142, which is used to perform speech recognition and semantic encoding on the user's speech description signal of the physical condition to obtain the semantic features of the user's subjective description of the physical condition; a disease information encoding unit 143, which is used to extract the symptom information text matching the disease type based on the disease type in the user's basic information, and perform sentence processing and semantic encoding on the symptom information text to obtain a sequence of semantic features of the granularity description of the disease information sentence; a user information query encoding and decoding unit 144, which is used to perform query encoding optimization under the guidance of a key matrix based on the sequence of the semantic features of the user's subjective description of the physical condition and the semantic features of the granularity description of the disease information sentence to obtain a query response representation of the subjective description of the physical condition-objective description of the disease condition, and obtain the subjective state estimation value.

[0042] In the embodiment of the present application, the speech description signal acquisition unit 141 is used to acquire the speech description signal of the user's physical condition. It should be understood that the semantic description signal of the user's physical condition specifically includes the specific physical symptom information that the user wants to express (such as pain, fatigue, cough, fever, etc.), the severity information of the symptom (the intensity or frequency of the symptom), the occurrence time and duration of the symptom, and other information. By processing and analyzing the semantic description signal, the user's health status can be understood, and effective health management and treatment suggestions can be provided.

[0043] In the embodiment of the present application, the body condition semantic encoding unit 142 is used to perform speech recognition and semantic encoding on the user's speech description signal of the body condition to obtain the semantic features of the user's subjective description of the body condition. Specifically, Figure 4 FIG. 1 is a block diagram of a physical condition semantic encoding unit in an Internet-based remote medical health management system according to an embodiment of the present application. Figure 4 As shown, the body condition semantic encoding unit 142 includes: a body condition speech description signal recognition subunit 1421, used for performing speech recognition on the user's speech description signal of the physical condition to obtain a speech-to-text conversion description of the user's body condition; a body condition description text semantic encoding subunit 1422, used for performing semantic encoding on the user's body condition speech-to-text conversion description to obtain a semantic encoding vector of the user's body condition subjective description as the semantic feature of the user's body condition subjective description.

[0044] In an embodiment of the present application, the physical condition voice description signal recognition subunit 1421 is used to perform voice recognition on the user's voice description signal of the physical condition to obtain a voice-to-text description of the user's physical condition. It should be understood that the voice description signal contains a large amount of information that is desired to be expressed and conveyed when describing the physical condition, which contains rich semantic information. Based on this, in order to more accurately and clearly understand and analyze the semantic information and content expressed by the user, in the technical solution of the present application, voice recognition is performed on the user's voice description signal of the physical condition to obtain a voice-to-text description of the user's physical condition, thereby laying the foundation for subsequent semantic understanding.

[0045] In an embodiment of the present application, the body state description text semantic encoding subunit 1422 is used to semantically encode the user's body state speech-to-text conversion description to obtain a semantic encoding vector of the user's body state subjective description as the semantic feature of the user's body state subjective description. Accordingly, in order to further understand and analyze the health status information expressed by the user, thereby more deeply understanding and analyzing the user's health status, and then providing more accurate and personalized health management services, in the technical solution of the present application, the user's body state speech-to-text conversion description is semantically encoded to obtain a semantic encoding vector of the user's body state subjective description. In particular, in a specific implementable embodiment of the present application, a converter-based semantic encoder can be used to semantically encode the user's body state speech-to-text conversion description to obtain a semantic encoding vector of the user's body state subjective description.

[0046] In the embodiment of the present application, the symptom information encoding unit 143 is used to extract the symptom information text matching the disease type based on the disease type in the user basic information, and perform sentence processing and semantic encoding on the symptom information text to obtain a sequence of semantic features of the symptom information sentence granularity description. Specifically, Figure 5 FIG. 1 is a block diagram of a disease information encoding unit in an Internet-based remote medical health management system according to an embodiment of the present application. Figure 5 As shown, the disease information encoding unit 143 includes: a symptom information text extraction subunit 1431, which is used to extract the symptom information text matching the disease type in the database module based on the disease type in the user basic information; a symptom information text sentence processing subunit 1432, which is used to perform sentence processing on the symptom information text to obtain a sequence of disease information sentence granularity text descriptions; a disease information short text semantic encoding subunit 1433, which is used to perform semantic encoding on each disease information sentence granularity text description in the sequence of disease information sentence granularity text descriptions to obtain a sequence of disease information sentence granularity description semantic encoding vectors as a sequence of disease information sentence granularity description semantic features.

[0047] In an embodiment of the present application, the symptom information text extraction subunit 1431 is used to extract the symptom information text matching the disease type in the database module based on the disease type in the user's basic information. It should be understood that the symptom information of various disease types in the database module expresses a set of typical or common symptoms related to a specific disease, and this information is usually summarized based on medical literature, clinical guidelines, historical cases and scientific research. Therefore, in order to confirm whether the symptoms reported by the user meet the known characteristics of the disease, so as to more accurately assess the health status, in the technical solution of the present application, based on the disease type in the user's basic information, the symptom information text matching the disease type is extracted in the database module.

[0048] In an embodiment of the present application, the symptom information text sentence processing subunit 1432 and the disease information short text semantic encoding subunit 1433 are respectively used to perform sentence processing on the symptom information text to obtain a sequence of disease information sentence granularity text descriptions and to perform semantic encoding on each disease information sentence granularity text description in the sequence of disease information sentence granularity text descriptions to obtain a sequence of disease information sentence granularity description semantic encoding vectors as a sequence of disease information sentence granularity description semantic features. Accordingly, considering that the symptom information text is composed of multiple independent symptom sentences, and each independent sentence contains the semantic association relationship between the contexts, based on this, in order to more accurately understand the specific meaning of each symptom to identify key semantic information, in the technical solution of the present application, the symptom information text is sentence-processed to obtain a sequence of symptom information sentence granularity text descriptions, and each symptom information sentence granularity text description in the sequence of symptom information sentence granularity text descriptions is semantically encoded to obtain a sequence of symptom information sentence granularity description semantic encoding vectors. In this way, the obtained symptom information sentence granularity description semantic encoding vector expresses the semantics of each sentence, including details such as the nature, severity, and duration of the symptom, which can be used for subsequent matching with the physical condition described by the user to more accurately describe the user's physical health status.

[0049] In the embodiment of the present application, the user information query encoding and decoding unit 144 is used to perform query encoding optimization under the guidance of a key matrix based on the sequence of the semantic features of the subjective description of the user's physical state and the semantic features of the granular description of the symptom information sentence to obtain the subjective description of the physical state-objective description of the symptom query response representation, and obtain the subjective state estimation value. Specifically, Figure 6 FIG. 1 is a block diagram of a user information query encoding and decoding unit in an Internet-based remote medical health management system according to an embodiment of the present application. Figure 6 As shown, the user information query encoding and decoding unit 144 includes: a user body state and pathology information query encoding optimization subunit 1441, which is used to perform query encoding optimization guided by a key matrix on the sequence of the user's body state subjective description semantic encoding vector and the pathology information sentence granularity description semantic encoding vector to obtain the body state subjective description-pathology objective description query response representation; a subjective state estimation value generation subunit 1442, which is used to obtain the subjective state estimation value based on the body state subjective description-pathology objective description query response representation.

[0050] In an embodiment of the present application, the user's physical state and symptom information query coding optimization subunit 1441 is used to perform query coding optimization under the guidance of a key matrix on the sequence of the user's physical state subjective description semantic coding vector and the symptom information sentence granularity description semantic coding vector to obtain the physical state subjective description-symptom objective description query response representation. Accordingly, considering that the user's physical state subjective description semantic coding vector reflects the user's intuitive understanding of his own health and the semantic information of the description of the symptoms. The sequence of the symptom information sentence granularity description semantic coding vector expresses the semantic information of each symptom description sentence granularity. Therefore, in order to query and encode the user's physical state subjective description semantic features based on the symptom information sentence granularity description semantic features, so as to establish an efficient mechanism to evaluate the user's health status and provide a more accurate estimate of the subjective health status in the future, in the technical solution of the present application, the user's physical state subjective description semantic coding vector and the symptom information sentence granularity description semantic coding vector are subjected to query coding optimization under the guidance of a key matrix to obtain the physical state subjective description-symptom objective description query response coding vector as the physical state subjective description-symptom objective description query response representation.

[0051] In detail, first, in order to capture the relationship between different vectors and ensure that they can be effectively compared and operated in the same space to better represent the complex patterns in the data, the sequence of the semantic coding vectors of the granular description of the pathological information sentence is linearly transformed. Then, the sequence of the semantic coding vectors of the granular description of the pathological information sentence obtained after the linear transformation is matrix-arranged to arrange them in order to form a semantic key matrix of the granular description of the pathological information sentence, so as to facilitate parallel processing and improve computational efficiency. After that, the maximum value of each column (i.e., each granular description semantic coding vector of the granular description of the pathological information sentence after the linear transformation) is selected from the semantic key matrix of the granular description of the pathological information sentence to obtain the significant feature vector of the semantic key matrix of the granular description of the pathological information sentence as the symptom information prompt template. In other words, by identifying the maximum value in each column in the key matrix, the most representative semantic features in each sentence can be determined, which usually reflect the most critical and core information in the sentence. Subsequently, the semantic encoding vector of the subjective description of the user's physical state is embedded and encoded by using a query embedding matrix and a value embedding matrix to extract the feature information of different aspects respectively, thereby obtaining a semantic query vector of the subjective description of the user's physical state and a semantic value vector of the subjective description of the user's physical state. It should be understood that the query vector is used to identify which symptom information sentence granularity description semantic information is the focus, while the value vector contains the actual symptom information sentence granularity description semantic content. Through the interaction between the query vector and the value vector, the system can more accurately identify the key semantic information. Next, the query vector, the value vector, the semantic key matrix, and the symptom information prompt template are jointly input into the heterogeneous converter structure based on template prompt optimization, so as to utilize the multi-attention mechanism within the module, automatically adjust the focus according to the symptom information prompt template, and capture the connection between sentences from multiple angles and details, thereby optimizing the query vector and realizing cross-domain optimization encoding of the symptom information sentence granularity. Finally, the positional mean of the optimized cross-domain optimized query encoding vector of subjective description of body state - objective description of disease symptoms is calculated to obtain the final query response encoding vector of subjective description of body state - objective description of disease symptoms. In this way, the positional information of the entire sequence can be retained, and the most critical and core features can be obtained.

[0052] Specifically, in the embodiment of the present application, the user physical condition symptom information query coding optimization subunit includes: a symptom information sentence granularity description linear transformation secondary subunit, which is used to use a key embedding matrix to linearly embed the semantic coding vectors of each symptom information sentence granularity description in the sequence of the symptom information sentence granularity description semantic coding vectors to obtain a sequence of symptom information sentence granularity description semantic coding vectors after linear transformation; a symptom information sentence granularity description semantic key matrix construction secondary subunit, which is used to matrix-arrange the sequence of symptom information sentence granularity description semantic coding vectors after linear transformation to obtain a symptom information sentence granularity description semantic key matrix; symptom information prompt The template selection secondary sub-unit is used to extract the maximum value of the symptom information sentence granularity description semantic coding vector after each linear transformation in the symptom information sentence granularity description semantic key matrix to obtain the significant feature vector of the symptom information sentence granularity description semantic key matrix as the symptom information prompt template; the user body state symptom information domain optimization coding secondary sub-unit is used to perform cross-domain optimization coding on the symptom information prompt template, the user body state subjective description semantic coding vector and the symptom information sentence granularity description semantic key matrix to obtain the body state subjective description-symptom objective description query response coding vector as the body state subjective description-symptom objective description query response representation.

[0053] More specifically, in an embodiment of the present application, the pathology information sentence granularity description linear transformation secondary subunit is used to: perform vector-matrix multiplication of the pathology information sentence granularity description semantic encoding vector and the key embedding matrix, and positionally add the obtained pathology information sentence granularity description semantic embedding vector and the key bias vector to obtain the linearly transformed pathology information sentence granularity description semantic encoding vector.

[0054] More specifically, in an embodiment of the present application, the user body state and symptom information domain optimization encoding secondary sub-unit includes: a user body state subjective description semantic embedding encoding tertiary sub-unit, which is used to use a query embedding matrix and a value embedding matrix to perform linear embedding encoding on the user body state subjective description semantic encoding vector to obtain a user body state subjective description semantic query vector and a user body state subjective description semantic value vector; a body state subjective description-symptom objective description cross-domain optimization tertiary sub-unit, which is used to input the user body state subjective description semantic query vector, the user body state subjective description semantic value vector, the symptom information sentence granularity description semantic encoding vector after each linear transformation in the symptom information sentence granularity description semantic key matrix, and the symptom information prompt template into a heterogeneous converter structure based on template prompt optimization to obtain a sequence of body state subjective description-symptom objective description cross-domain optimization query encoding vectors; a body state subjective description-symptom objective description query response encoding vector generation tertiary sub-unit, which is used to calculate the positional mean vector of the sequence of body state subjective description-symptom objective description cross-domain optimization query encoding vectors to obtain the body state subjective description-symptom objective description query response encoding vector.

[0055] More specifically, in an embodiment of the present application, the third-level sub-unit of the semantic embedding coding of the subjective description of the user's body state is used to: multiply the semantic coding vector of the subjective description of the user's body state with the query embedding matrix, and then add the result by position with the query bias vector to obtain the semantic query vector of the subjective description of the user's body state; multiply the semantic coding vector of the subjective description of the user's body state with the value embedding matrix, and then add the result by position with the value bias vector to obtain the semantic value vector of the subjective description of the user's body state.

[0056] More specifically, in an embodiment of the present application, the three-level sub-unit for cross-domain optimization of subjective description of body state and objective description of symptoms is used to: multiply the semantic query vector of the user's subjective description of body state with the transposed vector of the semantic encoding vector of the granular description of the symptom information sentence after the linear transformation, and divide the obtained user body-symptom information semantic association matrix by the bi-norm of the symptom information prompt template by position to obtain the user body-symptom information semantic association modulation matrix; use the softmax function to process the user body-symptom information semantic association modulation matrix, and perform matrix-vector multiplication of the obtained user body-symptom information semantic association activation matrix and the symptom information prompt template to obtain the user body-symptom information semantic weight vector; calculate the positional point multiplication of the user body-symptom information semantic weight vector and the semantic value vector of the user's subjective description of body state to obtain the cross-domain optimization query encoding vector of the subjective description of body state and the objective description of symptoms.

[0057] In an embodiment of the present application, the user physical state and symptom information query coding optimization subunit is used to perform query coding optimization processing guided by a key matrix on the sequence of the user physical state subjective description semantic coding vector and the symptom information sentence granular description semantic coding vector according to the following formula; wherein the formula is:

[0058]

[0059]

[0060]

[0061]

[0062]

[0063]

[0064]

[0065]

[0066] in, is a sequence of semantic encoding vectors describing the granularity of the disease information sentence, is the first, second, ..., first in the sequence of the semantic encoding vector of the granular description of the disease information sentence ,..., The semantic encoding vector is described at the granularity of disease information sentences. is the first, second, ..., first in the sequence of semantic encoding vectors of the granular description of the disease information sentence after linear transformation. ,..., The granularity description semantic encoding vector of the disease information sentence after linear transformation, is the semantic key matrix of the granular description of the disease information sentence, and are the key embedding matrix and key bias vector respectively, For Perform linear transformation, To extract the maximum value in a vector, is the significant feature vector of the semantic key matrix describing the granularity of the disease information sentence, is the semantic encoding vector of the subjective description of the user's physical state, and are the query embedding matrix and the query bias vector respectively, is the semantic query vector of the user's subjective description of his / her physical state, and are the value embedding matrix and value bias vector respectively, is the semantic value vector of the subjective description of the user's physical state, for The transposed vector of is the two-norm of the vector, yes function, is the matrix-vector multiplication, It is the point multiplication by position. is the first in the sequence of the cross-domain optimization query encoding vector of the subjective description of the physical state and the objective description of the disease. A subjective description of the body state-objective description of the disease condition cross-domain optimization query encoding vector, is the number of vectors in the sequence of the cross-domain optimized query encoding vector of the subjective description of the physical state and the objective description of the symptom, It is the encoding vector of the query response of subjective description of physical state-objective description of disease condition.

[0067] In an embodiment of the present application, the subjective state estimation value generation subunit 1442 is used to obtain the subjective state estimation value based on the subjective description of the body state-objective description of the symptom query response representation. Specifically, in an embodiment of the present application, the subjective state estimation value generation subunit is used to: input the subjective description of the body state-objective description of the symptom query response encoding vector into the decoder-based subjective state estimation module to obtain the subjective state estimation value. That is, the subjective description of the body state-objective description of the symptom query response obtained by querying the sequence of the semantic encoding vector of the user's body state subjective description and the semantic encoding vector of the symptom information sentence granularity description is decoded to intelligently obtain the subjective state estimation value. In this way, the physical condition information expressed by the user can be analyzed more deeply, so as to more accurately understand the user's actual intention. In addition, the semantic encoding technology can capture the unique characteristics of the individual according to the user's specific description, thereby providing a more personalized evaluation result, which is better adapted to the description style and health status of different users.

[0068] In particular, the present application takes into account that the sequence of the semantic coding vector of the subjective description of the user's body state and the semantic coding vector of the sentence granularity description of the disease information respectively represent the contextual semantic coding features of the speech-to-text conversion description of the user's body state and the sentence granularity semantic coding features of the sentence granularity text description of the disease information. When performing query encoding optimization guided by the key matrix, the granularity difference of the contextual semantic features will make the query response encoding vector of the subjective description of the body state-objective description of the disease have a complex spatial structure due to the difference in attention weight distribution of cross-domain and hetero-granularity queries, thereby expecting to improve the decoding regression convergence and generalization effects under the complex spatial structure.

[0069] Preferably, in an example of the present application, passing the body state subjective description - pathology objective description query response coding vector through a decoder-based subjective state estimation module to obtain a subjective state estimation value includes: calculating the sum of the absolute values ​​of each eigenvalue of the body state subjective description - pathology objective description query response coding vector to obtain a first subjective description - objective description query response space structure value, and calculating the square root of the sum of the squares of each eigenvalue of the body state subjective description - pathology objective description query response coding vector to obtain a second subjective description - objective description query response space structure value; multiplying each eigenvalue of the body state subjective description - pathology objective description query response coding vector with the first subjective description - objective description query response space structure value and the second subjective description - objective description query response space structure value to obtain a first subjective description - objective description query response structure reference value and a second subjective description - objective description query response structure reference value corresponding to each eigenvalue; multiplying each eigenvalue of the body state subjective description - pathology objective description query response coding vector with the length of the body state subjective description - pathology objective description query response coding vector and the length of the body state subjective description - pathology objective description query response coding vector. The square root of is multiplied to obtain a first subjective description-objective description query response scale transformation value and a second subjective description-objective description query response scale transformation value corresponding to each eigenvalue; the first subjective description-objective description query response structure reference value is divided by the difference between the first subjective description-objective description query response space structure value and the first subjective description-objective description query response scale transformation value to obtain a first subjective description-objective description query response transformation adjustment value; the second subjective description-objective description query response structure reference value is divided by the difference between the second subjective description-objective description query response space structure value and the second subjective description-objective description query response scale transformation value to obtain a second subjective description-objective description query response transformation adjustment value; the weighted sum of the first subjective description-objective description query response transformation adjustment value and the second subjective description-objective description query response transformation adjustment value is calculated to obtain each eigenvalue of the optimized body state subjective description-pathology objective description query response coding vector; the optimized body state subjective description-pathology objective description query response coding vector is passed through the decoder-based subjective state estimation module to obtain the subjective state estimation value.

[0070] Here, the optimization expression of the query response encoding vector of the subjective description of the physical state-objective description of the disease condition is:

[0071]

[0072]

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079] in, represents the query response encoding vector of subjective description of physical state-objective description of disease condition, represents the length of the encoding vector of the query response of the subjective description of the physical state and the objective description of the symptom, The subjective description of the physical state - objective description of the disease condition query response encoding vector represents the first The eigenvalues ​​at the positions, represents the first subjective description-objective description query response space structure value, represents the second subjective description-objective description query response space structure value, represents the first subjective description-objective description query response transformation adjustment vector The first subjective description-objective description query response transformation adjustment value of the position, represents the first in the second subjective description-objective description query response transformation adjustment vector The second subjective description-objective description query response transformation adjustment value of the position, represents the first subjective description-objective description query response transformation adjustment vector, represents the second subjective description-objective description query response transformation adjustment vector, Indicates adding by position point. It means point multiplication by position. represents the weighted hyperparameter, The optimized subjective description of body state-objective description of disease condition query response encoding vector is represented.

[0080] That is, in the preferred example, for the spatial structure information of the feature set of the subjective description of the body state-objective description of the condition query response encoding vector in the high-dimensional space, the scale-based frame transformation of each feature value of the subjective description of the body state-objective description of the condition query response encoding vector is performed by taking the class norm space structured representation of the subjective description of the body state-objective description of the condition query response encoding vector as the reference window, and the spatial structure-based frame attention weight adjustment of each feature value of the subjective description of the body state-objective description of the condition query response encoding vector is realized to ensure the spatial transformation (translation, scaling and rotation) invariance of the subjective description of the body state-objective description of the condition query response encoding vector under the interaction of the feature space, thereby improving the convergence and generalization effect of the decoding regression of the feature set of the subjective description of the body state-objective description of the condition query response encoding vector under the complex spatial structure representation, and improving the accuracy of the subjective state estimation value obtained by passing it through the decoder-based subjective state estimation module. In this way, the physical condition information expressed by the user can be analyzed more deeply, so as to more accurately understand the actual intention of the user. In addition, semantic coding technology can capture the unique characteristics of an individual based on the user's specific description, thereby providing more personalized assessment results and better adapting to the description style and health status of different users.

[0081] Therefore, the subjective analysis module 140 is explained clearly, which uses artificial intelligence-based speech processing and data analysis technology to perform speech recognition and semantic coding on the user's speech description signal of the physical condition, and extracts the matching information text from the database module based on the disease type in the user's basic information, and then performs sentence processing and semantic coding on the matching symptom information text, so as to intelligently obtain the subjective state estimation value according to the query response characteristics between the semantic features of the user's subjective description of the physical state and the semantic features of the granular description of the symptom information sentence. In this way, the physical condition information expressed by the user can be analyzed more deeply, so as to more accurately understand the user's actual intention. In addition, the semantic coding technology can capture the unique characteristics of the individual according to the user's specific description, so as to provide a more personalized evaluation result, which is better adapted to the description style and health status of different users.

[0082] In the embodiment of the present application, the objective analysis module 150 is used to obtain an objective state estimation value based on the health state in the basic information of the user. It should be understood that the subjective description of the user is important for judging the health state of the user, but it may be affected by personal emotions and expression ability. In order to provide a more scientific basis for user health assessment and reduce subjective bias, in the technical solution of the present application, it is necessary to obtain an objective state estimation value based on the health state in the basic information of the user. The objective state estimation value is obtained based on the specific user's physical health state data, which can assist in judging the user's health status, thereby improving the accuracy of health assessment. Specifically, in the embodiment of the present application, a feasible implementation method for obtaining an objective state estimation value based on the health state in the basic information of the user can be: a. extract the heart rate information, blood pressure information, and body temperature information contained in the health state in the basic information of the user, and search the database module for sample information of healthy people who match the user's age and gender at the same time, so as to obtain the heart rate range, blood pressure range, and body temperature range of healthy people of the same age and gender as the user; b. calculate the objective state estimation value, and the calculation formula of the objective state estimation value is:

[0083]

[0084] in, The heart rate information of the user is beyond the heart rate range of healthy people of the same age and gender. The median heart rate range of healthy people of the same age and gender as the user; The blood pressure information of the user is beyond the blood pressure range of healthy people of the same age and gender. The median blood pressure range of healthy people of the same age and gender as the user; The user's body temperature information is out of the range of healthy people of the same age and gender. is the median of the body temperature range of healthy people of the same age and gender as the user. is the objective state estimation value.

[0085] In an embodiment of the present application, the feedback module 160 is used to obtain a health status value based on the subjective state estimation value and the objective state estimation value, and to feed back the health status value to medical staff. It should be understood that the evaluation of the user's health status should be comprehensive, and it is necessary to include both objective physiological indicators and the user's subjective feelings. In order to more comprehensively reflect the patient's health status, in the technical solution of the present application, a health status value is obtained based on the subjective state estimation value and the objective state estimation value. In this way, the user's health status can be comprehensively evaluated and the accuracy of the evaluation can be improved. At the same time, the health status value is fed back to medical staff, which can provide them with important decision support information, help them better understand the patient's health status, and thus develop a more effective treatment plan. Specifically, in an embodiment of the present application, a feasible implementation method for obtaining a health status value based on the subjective state estimation value and the objective state estimation value can be:

[0086]

[0087]

[0088] in, is the health status value, and are the subjective state estimation value and the objective state estimation value, respectively, and are respectively the proportion coefficient of the subjective state estimation value and the proportion coefficient of the objective state estimation value.

[0089] In summary, the Internet-based telemedicine health management system 100 according to the embodiment of the present application is explained, which first collects basic user information including user age, gender, disease type and health status, and then analyzes the basic user information in combination with symptom information in the database and sample information of healthy people to calculate the subjective and objective state estimation values, and finally combines the two estimation values ​​to generate the health state value and feed it back to the medical staff to support making appropriate medical decisions. In this way, it helps to achieve more accurate and personalized health management, and then provide high-quality telemedicine services.

[0090] As described above, the Internet-based remote medical health management system 100 according to the embodiment of the present application can be implemented in various terminal devices, such as a server for Internet-based remote medical health management. In one example, the Internet-based remote medical health management system 100 according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the Internet-based remote medical health management system 100 can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the Internet-based remote medical health management system 100 can also be one of the many hardware modules of the terminal device.

[0091] Alternatively, in another example, the Internet-based remote medical health management system 100 and the terminal device may also be separate devices, and the Internet-based remote medical health management system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0092] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details of the above application are only for the purpose of illustration and ease of understanding, rather than limitation, and the above details do not limit the present application to the use of the above specific details to be implemented.

Claims

1. A remote medical health management system based on the Internet, characterized in that: include: A user information collection module, used to collect basic user information, including age, gender, disease type and health status; A user information storage module, used to store the basic information of the user; A database module is used to store symptom information of various disease types and sample information of healthy people; A subjective analysis module, used to obtain a subjective state estimation value based on the disease type; An objective analysis module, configured to obtain an objective state estimation value based on the health state; A feedback module, configured to obtain a health status value based on the subjective state estimation value and the objective state estimation value, and feed back the health status value to medical staff; The subjective analysis module comprises: A voice description signal acquisition unit, used to acquire a voice description signal of a user's physical condition; A body condition semantic coding unit, used for performing speech recognition and semantic coding on the user's speech description signal of the body condition to obtain semantic features of the user's subjective description of the body condition; A symptom information encoding unit, used for extracting symptom information text matching the disease type, and performing sentence processing and semantic encoding on the symptom information text to obtain a sequence of semantic features of symptom information sentence granularity description; A user information query encoding and decoding unit, configured to obtain a subjective description of the user's physical state-objective description of the physical state query response representation obtained by performing query encoding optimization under the guidance of a key matrix based on a sequence of semantic features of the subjective description of the user's physical state and semantic features of the granular description of the symptom information sentence, and obtain the subjective state estimation value; Wherein, the user information query encoding and decoding unit includes: a user physical condition and disease information query encoding optimization subunit; The user's physical condition and symptom information query coding optimization subunit includes: The symptom information sentence granularity description linear transformation secondary subunit is used to use the key embedding matrix to perform linear embedding coding on each symptom information sentence granularity description semantic coding vector in the sequence of symptom information sentence granularity description semantic coding vectors to obtain a sequence of symptom information sentence granularity description semantic coding vectors after linear transformation; The symptom information sentence granularity description semantic key matrix constructs a secondary subunit, which is used to perform matrix arrangement on the sequence of the symptom information sentence granularity description semantic coding vectors after the linear transformation to obtain the symptom information sentence granularity description semantic key matrix; The symptom information prompt template selects a secondary subunit, which is used to extract the maximum value of the symptom information sentence granularity description semantic encoding vector after each linear transformation in the symptom information sentence granularity description semantic key matrix to obtain the salient feature vector of the symptom information sentence granularity description semantic key matrix as the symptom information prompt template; The user body state and symptom information domain optimization encoding secondary sub-unit is used to perform cross-domain optimization encoding on the symptom information prompt template, the semantic features of the user body state subjective description and the semantic key matrix of the symptom information sentence granularity description to obtain a body state subjective description-symptom objective description query response encoding vector as the body state subjective description-symptom objective description query response representation.

2. The Internet-based remote medical health management system according to claim 1, characterized in that: The physical condition semantic encoding unit comprises: A body condition voice description signal recognition subunit, used for performing voice recognition on the user's voice description signal of the body condition to obtain a voice-to-text description of the user's body condition; The body state description text semantic coding subunit is used to semantically code the user's body state speech-to-text conversion description to obtain a semantic coding vector of the user's body state subjective description as the semantic feature of the user's body state subjective description.

3. The Internet-based remote medical health management system according to claim 2, characterized in that: The disease information encoding unit comprises: A symptom information text extraction subunit, configured to extract, in the database module, symptom information text matching the disease type based on the disease type in the user basic information; A symptom information text sentence processing subunit, used for performing sentence processing on the symptom information text to obtain a sequence of symptom information sentence granularity text descriptions; The symptom information short text semantic encoding subunit is used to semantically encode each symptom information sentence granularity text description in the sequence of the symptom information sentence granularity text description to obtain a sequence of symptom information sentence granularity description semantic encoding vectors as a sequence of symptom information sentence granularity description semantic features.

4. The Internet-based remote medical health management system according to claim 3, characterized in that: The user information query encoding and decoding unit includes: A user body state and symptom information query coding optimization subunit is used to perform query coding optimization guided by a key matrix on the sequence of the user body state subjective description semantic coding vector and the symptom information sentence granular description semantic coding vector to obtain the body state subjective description-symptom objective description query response representation; The subjective state estimation value generating subunit is used to obtain the subjective state estimation value based on the query response representation of the subjective description of the body state-objective description of the disease condition.

5. The Internet-based remote medical health management system according to claim 4, characterized in that: The second-level subunit of linear transformation of the granularity description of the pathological information sentence is used to: perform vector-matrix multiplication on the semantic coding vector of the granularity description of the pathological information sentence and the key embedding matrix, and add the obtained semantic embedding vector of the granularity description of the pathological information sentence and the key bias vector by position to obtain the semantic coding vector of the granularity description of the pathological information sentence after the linear transformation.

6. The Internet-based remote medical health management system according to claim 5, characterized in that: The user's physical condition information domain optimization encoding secondary subunit includes: The user's body state subjective description semantic embedding coding tertiary sub-unit is used to use the query embedding matrix and the value embedding matrix to perform linear embedding coding on the user's body state subjective description semantic coding vector to obtain the user's body state subjective description semantic query vector and the user's body state subjective description semantic value vector; The subjective description of body state - objective description of symptom cross-domain optimization three-level sub-unit is used to input the semantic query vector of the subjective description of user's body state, the semantic value vector of the subjective description of user's body state, the semantic encoding vector of the symptom information sentence granularity description after each linear transformation in the semantic key matrix of the symptom information sentence granularity description, and the symptom information prompt template into the heterogeneous converter structure based on template prompt optimization to obtain a sequence of cross-domain optimization query encoding vectors of the subjective description of body state - objective description of symptom; The subjective description of the body state - objective description of the disease condition query response coding vector generates a three-level sub-unit, which is used to calculate the positional mean vector of the sequence of the cross-domain optimized query coding vector of the subjective description of the body state - objective description of the disease condition to obtain the query response coding vector of the subjective description of the body state - objective description of the disease condition.

7. The Internet-based remote medical health management system according to claim 6, characterized in that: The user's body state subjective description semantic embedding coding three-level sub-unit is used for: The semantic encoding vector of the subjective description of the user's physical state is multiplied by the query embedding matrix, and then added to the query bias vector by position to obtain the semantic query vector of the subjective description of the user's physical state; The semantic encoding vector of the subjective description of the user's physical state is multiplied by the value embedding matrix and then added to the value bias vector by position to obtain the semantic value vector of the subjective description of the user's physical state.

8. The Internet-based remote medical health management system according to claim 7, characterized in that: The subjective description of physical state and the objective description of disease condition cross-domain optimization three-level subunit are used for: Multiplying the user's body state subjective description semantic query vector by the transposed vector of the granular description semantic encoding vector of the condition information sentence after the linear transformation, and dividing the obtained user body-condition information semantic association matrix by the bi-norm of the condition information prompt template by position to obtain the user body-condition information semantic association modulation matrix; The user body-symptom information semantic association modulation matrix is ​​processed using a softmax function, and the obtained user body-symptom information semantic association activation matrix is ​​matrix-vector multiplied with the symptom information prompt template to obtain a user body-symptom information semantic weight vector; The user's body-symptom information semantic weight vector and the user's body state subjective description semantic value vector are calculated and multiplied by the position point to obtain the body state subjective description-symptom objective description cross-domain optimized query coding vector.

9. The Internet-based remote medical health management system according to claim 8, characterized in that: The subjective state estimation value generating subunit is used to input the body state subjective description-symptom objective description query response coding vector into a decoder-based subjective state estimation module to obtain the subjective state estimation value.

Images