Intelligent Recommendation Method, Device, Equipment and Medium for Cardiovascular-Related Monitoring Information

Abstract

The present invention discloses an intelligent recommendation method, device, equipment and medium for cardiovascular-related monitoring information. First, the current target user information and the current target time interval are obtained. If it is determined that the current target user information includes a cardiovascular disease patient label, the current target user and the current monitoring recommendation information are obtained. If it is determined that the current user input information is of the text information type, the corresponding historical text information is obtained, and the current comprehensive input information is formed by the current user input information, the current monitoring recommendation information and the historical text information and input into a pre-trained language model to obtain the current intelligent recommendation information. Through the above method, after combining the patient label of the current target user, the current monitoring recommendation information, the current user input information and the historical text information to form the current comprehensive input information containing more information, it is input into the pre-trained language model for the medical field to obtain more accurate intelligent recommendation information with reference to more context information.

Description

Technical Field

[0001] The present invention relates to the technical field of natural language processing in artificial intelligence, and particularly relates to an intelligent recommendation method, device, equipment and medium for cardiovascular-related monitoring information. Background Art

[0002] Currently, natural language processing technology has been widely applied in intelligent customer service systems. When a user has a conversation with an intelligent customer service system, the intelligent customer service system can give corresponding reply information according to the current text information sent by the user to achieve intelligent recommendation and reply of information. When the intelligent customer service system used by the user is an intelligent system for multiple fields, although it can give corresponding replies to the questions and answers of the user in various technical fields, when the user asks a professional question in a specific field such as the medical field, the replies given are often relatively general and inaccurate reply data, and there is a poor correlation with the context. Summary of the Invention

[0003] Embodiments of the present invention provide an intelligent recommendation method, device, equipment and medium for cardiovascular-related monitoring information, aiming to solve the problem that when a user uses an intelligent customer service system to ask professional questions in a specific field such as the medical field in the existing technical methods, the replies given are often relatively general and inaccurate reply data, and there is a poor correlation with the context.

[0004] In a first aspect, an embodiment of the present invention provides an intelligent recommendation method for cardiovascular-related monitoring information, which is applied to a server and includes:

[0005] In response to an intelligent recommendation instruction of information, obtain the current target user information corresponding to the intelligent recommendation instruction of information, and obtain the current target time interval corresponding to the intelligent recommendation instruction;

[0006] If it is determined that the current target user information includes a cardiovascular disease patient label, obtain the current target user corresponding to the current target user information, and obtain the current monitoring recommendation information corresponding to the current target user and stored in the local storage space;

[0007] If it is detected that the current user enters information, obtain the current information type of the information entered by the current user;

[0008] If it is determined that the current information type is a text information type, obtain the historical text information corresponding to the current target user and with the information generation time within the current target time interval, and form the current comprehensive input information with the information entered by the current user, the current monitoring recommendation information and the historical text information, and input it into a pre-trained language model to obtain the current intelligent recommendation information; wherein, the pre-trained language model is a pre-trained language model for the medical field;

[0009] Send the current intelligent recommendation information to the current target user terminal corresponding to the current target user;

[0010] If the current feedback label sent by the current target user terminal and corresponding to the current intelligent recommendation information is received, add the current feedback label to the current intelligent recommendation information and store it in the local storage space.

[0011] In a second aspect, an embodiment of the present invention provides an intelligent recommendation device for cardiovascular-related monitoring information, configured in a server, and includes:

[0012] A target user information acquisition unit, configured to, in response to an information intelligent recommendation instruction, acquire current target user information corresponding to the information intelligent recommendation instruction, and acquire a current target time interval corresponding to the intelligent recommendation instruction;

[0013] A monitoring recommendation information acquisition unit, configured to, if it is determined that the current target user information includes a cardiovascular disease patient label, acquire a current target user corresponding to the current target user information, and acquire current monitoring recommendation information corresponding to the current target user and stored in the local storage space;

[0014] An information type acquisition unit, configured to, if it detects that a current user enters information, acquire the current information type of the current user-entered information;

[0015] A current intelligent recommendation information acquisition unit, configured to, if it is determined that the current information type is a text information type, acquire historical text information corresponding to the current target user and with an information generation time within the current target time interval, form current comprehensive input information with the current user-entered information, the current monitoring recommendation information, and the historical text information, and input the current comprehensive input information into a pre-trained language model to obtain current intelligent recommendation information; wherein, the pre-trained language model is a pre-trained language model for the medical field;

[0016] A current intelligent recommendation information sending unit, configured to send the current intelligent recommendation information to the current target user terminal corresponding to the current target user;

[0017] A current feedback label storage unit, configured to, if the current feedback label sent by the current target user terminal and corresponding to the current intelligent recommendation information is received, add the current feedback label to the current intelligent recommendation information and store it in the local storage space.

[0018] In a third aspect, an embodiment of the present invention further provides a computer device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer device executes the computer program, it implements the intelligent recommendation method for cardiovascular-related monitoring information as described in the first aspect above.

[0019] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, in which the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the intelligent recommendation method for cardiovascular-related monitoring information as described in the first aspect above.

[0020] The embodiment of the present invention provides an intelligent recommendation method, device, equipment and medium for cardiovascular-related monitoring information. The method includes: in response to an information intelligent recommendation instruction, obtaining the current target user information corresponding to the information intelligent recommendation instruction, and obtaining the current target time interval corresponding to the intelligent recommendation instruction; if it is determined that the current target user information includes a cardiovascular disease patient label, obtaining the current target user corresponding to the current target user information, and obtaining the current monitoring recommendation information corresponding to the current target user and stored in the local storage space; if it is detected that the current user enters information, obtaining the current information type of the current user-entered information; if it is determined that the current information type is a text information type, obtaining the historical text information corresponding to the current target user and with the information generation time within the current target time interval, forming the current comprehensive input information with the current user-entered information, the current monitoring recommendation information and the historical text information, and inputting the current comprehensive input information into a pre-trained language model to obtain the current intelligent recommendation information; sending the current intelligent recommendation information to the current target user terminal corresponding to the current target user; if receiving the current feedback label sent by the current target user terminal and corresponding to the current intelligent recommendation information, adding the current feedback label to the current intelligent recommendation information and storing it in the local storage space. Through the above method, after combining the patient label, current monitoring recommendation information, current user-entered information and historical text information of the current target user to form the current comprehensive input information containing more information, it is input into the pre-trained language model for the medical field to obtain more accurate intelligent recommendation information with more context information considered. Description of the Drawings

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0022] Figure 1Schematic diagram of the application scenario of the intelligent recommendation method for cardiovascular-related monitoring information provided by the embodiments of the present invention;

[0023] Figure 2 Schematic flowchart of the intelligent recommendation method for cardiovascular-related monitoring information provided by the embodiments of the present invention;

[0024] Figure 3 Schematic sub-flowchart of the intelligent recommendation method for cardiovascular-related monitoring information provided by the embodiments of the present invention;

[0025] Figure 4 Schematic sub-flowchart of the intelligent recommendation method for cardiovascular-related monitoring information provided by the embodiments of the present invention;

[0026] Figure 5 Schematic sub-flowchart of the intelligent recommendation method for cardiovascular-related monitoring information provided by the embodiments of the present invention;

[0027] Figure 6 Schematic block diagram of the intelligent recommendation device for cardiovascular-related monitoring information provided by the embodiments of the present invention;

[0028] Figure 7 Schematic block diagram of the computer device provided by the embodiments of the present invention. Detailed implementation manners

[0029] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0030] It should be understood that when used in this specification and the appended claims, the terms "comprises" and "comprising" indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.

[0031] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in this specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an", and "the" are intended to include the plural forms.

[0032] It should be further understood that the term " / and" used in this specification of the present invention and the appended claims refers to any combination and all possible combinations of one or more of the related listed items, and includes these combinations.

[0033] See also Figure 1 and Figure 2 , Figure 1 The embodiment of the present invention provides Figure 1 A schematic diagram of an application scenario of the method for intelligently recommending cardiovascular-related monitoring information provided by an embodiment of the present invention; Figure 2 Schematic diagram of the process of the intelligent recommendation method for cardiovascular-related monitoring information provided by an embodiment of the present invention; the intelligent recommendation method for cardiovascular-related monitoring information is applied to a server 10, which is connected to a user terminal 20 for communication. Figure 2 As shown, the method includes steps S110 to S160.

[0034] S110 . In response to an information intelligent recommendation instruction, obtain current target user information corresponding to the information intelligent recommendation instruction, and obtain current target time interval corresponding to the intelligent recommendation instruction.

[0035] In this embodiment, the technical solution is described using a server as the execution entity. The server houses an intelligent information recommendation system (similar to an intelligent customer service platform, capable of responding to user questions or text messages). After a user logs in to the server's intelligent information recommendation system using a user terminal, they can interact with the system to obtain the required intelligent recommendations.

[0036] Among them, when the user uses the user terminal to log in to the information intelligent recommendation system in the server, under the premise that the user authorizes to obtain his or her user information, the server can respond to the information intelligent recommendation instruction and obtain the current target user information corresponding to the information intelligent recommendation instruction, and the current target user information includes at least user ID, user name, user gender, user age, user patient label and other information. Moreover, the current target time interval corresponding to the current target user information can also be obtained based on the instruction generation time of the intelligent recommendation instruction. The information intelligent recommendation system in this application is a system that focuses more on making more accurate monitoring information intelligent recommendations for users with various patient labels, and can provide more intelligent and accurate replies to information feedback sent by users in the information intelligent recommendation system.

[0037] In one embodiment, if Figure 3 As shown, step S110 includes:

[0038] S111. Acquire current user classification label information in the current target user information, and determine a current acquisition period based on the current user classification label information and a preset label-period mapping relationship;

[0039] S112. Obtain the instruction generation time of the information intelligent recommendation instruction, determine the start time of the current target time interval based on the difference between the instruction generation time and the current acquisition period, and use the instruction generation time as the end time of the current target time interval;

[0040] S113. Determine the current target time interval according to the start time of the current target time interval and the end time of the current target time interval.

[0041] In this embodiment, first, the current user classification label information can be obtained from the current target user information, such as high-risk user classification label, medium-risk user classification label, or low-risk user classification label, etc. Moreover, a mapping relationship between each user classification label information and the acquisition period duration is set in the server, such as setting that the high-risk user classification label corresponds to the acquisition period T1, the medium-risk user classification label corresponds to the acquisition period T2, and the low-risk user classification label corresponds to the acquisition period T3 (where T1, T2, and T3 are all positive numbers), and it is determined that the current user classification label information corresponds to the medium-risk user classification label. At this time, the current acquisition period can be determined as the acquisition period T2 according to the label and period mapping relationship.

[0042] After that, combined with the instruction generation time of the information intelligent recommendation instruction and the current acquisition period (such as the acquisition period T2 in the above example), specifically, the start time of the current target time interval can be determined based on the difference between the instruction generation time and the current acquisition period, and the instruction generation time is used as the end time of the current target time interval. Finally, after determining the start time point and the end time point of the current target time interval, the determination of the current target time interval can be completed. Moreover, the obtained current target time interval is used to determine the communication history data (which can also be understood as historical context data) corresponding to the current target user information.

[0043] S120. If it is determined that the current target user information includes a cardiovascular disease patient label, obtain the current target user corresponding to the current target user information, and obtain the current monitoring recommendation information corresponding to the current target user and stored in the local storage space.

[0044] In this embodiment, since the current target user information includes information of the user patient label, after parsing and determining that it is a cardiovascular disease patient label, the current target user corresponding to the current target user information can be further obtained, and the current monitoring recommendation information corresponding to the current target user and stored in the local storage space can be obtained.

[0045] Among them, the current guardianship recommendation information is uploaded to the server by other business servers and edited and uploaded by doctor users for the current target user. For example, the current guardianship recommendation information includes medical advice information for the current target user within a specified time period (which can be understood as the information effective time period corresponding to the current guardianship recommendation information). More specifically, it can include recommended diet information, medication information, and daily rehabilitation exercise information, etc. It can also include the monitoring thresholds of key physical sign indicators for the current target user within the current specified time period (for example, once the key physical sign indicators of the current target user exceed the monitoring thresholds of the key physical sign indicators multiple times within the current specified time period, it means that a reexamination needs to be carried out in a timely manner). Moreover, after the user logs in to the information intelligent recommendation system of the server using the user terminal, the current guardianship recommendation information corresponding to the current target user can either be hidden information and not displayed on the display interface of the information intelligent recommendation system (such as a chat box-style display interface), or be non-hidden information directly displayed on the display interface of the information intelligent recommendation system.

[0046] S130. If the current user input information is detected, obtain the current information type of the current user input information.

[0047] In this embodiment, after the user logs in to the information intelligent recommendation system using the intelligent terminal, one of the common usage methods is to input the current user input information. For example, it can be text input information (such as Chinese text mainly composed of Chinese characters, English text mainly composed of English characters, digital text mainly composed of digital characters, etc.), voice input information, image input information, etc. However, no matter in what way the user inputs the relevant information, the current information type of the current user input information can be obtained first.

[0048] In one embodiment, as Figure 4 shown, step S130 includes:

[0049] S131. Obtain the file format of the current user input information;

[0050] S132. If it is determined that the file format corresponds to the text file format, use the text information type as the current information type;

[0051] S133. If it is determined that the file format corresponds to the picture file format, use the picture information type as the current information type;

[0052] S134. If it is determined that the file format corresponds to the voice file format, use the voice information type as the current information type.

[0053] In this embodiment, when determining the current information type of the currently entered information of the user, the fastest implementation method is to obtain its file format, more specifically, to obtain the file suffix name of the user-entered information. After quickly determining the current information type in combination with the file suffix name, corresponding responses can be further made in combination with the current information type and the specific information content.

[0054] Specifically, if it is determined that the file format corresponds to the text file format, such as txt, then the text information type is used as the current information type; if it is determined that the file format corresponds to the picture file format, such as jpg, then the picture information type is used as the current information type; if it is determined that the file format corresponds to the voice file format, such as wma, then the voice information type is used as the current information type. It can be seen that the current information type can be quickly determined through the above method.

[0055] S140. If it is determined that the current information type is the text information type, then obtain the historical text information corresponding to the current target user and whose information generation time is within the current target time interval, and form the current comprehensive input information with the currently entered information of the user, the current guardianship recommendation information, and the historical text information, and input it into the pre-trained language model to obtain the current intelligent recommendation information.

[0056] Among them, the pre-trained language model is a pre-trained language model for the medical field.

[0057] In this embodiment, if it is determined that the current information type is the text information type, it can further be determined whether it is a Chinese text mainly composed of Chinese characters, an English text mainly composed of English characters, or a digital text mainly composed of digital characters. For example, if it is determined that the currently entered information of the user is mainly input of Chinese characters (that is, the proportion of Chinese characters in the currently entered information of the user exceeds the first character proportion threshold), it means that the current target user is mainly asking questions. The historical text information corresponding to the current target user and whose information generation time is within the current target time interval and the current guardianship recommendation information can be combined, and the above three types of information are used to form a current comprehensive input information and input it into the pre-trained language model (such as the BioBERT model) to obtain the current intelligent recommendation information output by the pre-trained language model.

[0058] In one embodiment, as Figure 5 shown, in step S140, forming the current comprehensive input information with the currently entered information of the user, the current guardianship recommendation information, and the historical text information and inputting it into the pre-trained language model to obtain the current intelligent recommendation information includes:

[0059] S141. Combine the current guardianship recommendation information, the historical text information, and the current user input information in the order of the current guardianship recommendation information, the historical text information, and the current user input information to form the current comprehensive input information.

[0060] S142. Input the current input vector corresponding to the current comprehensive input information into the pre-trained language model to obtain the current intelligent recommendation information.

[0061] In this embodiment, in order to input more accurate current intelligent recommendation information by combining more context information, the directly displayed or hidden current guardianship recommendation information can be used as the context data ranked first in the current comprehensive input information, the historical text information can be used as the context data ranked in the middle of the current comprehensive input information, and combined with the current user input information, a current comprehensive input information that fully considers the current guardianship recommendation information, historical question information, system historical reply information, and other context data of the current target user can be formed. After performing text tokenization processing (i.e., Tokenization, to split the text into tokens) and text encoding processing (to map the tokens to a low-dimensional vector space) on it to obtain the current input vector, and then inputting it into the pre-trained language model, the current intelligent recommendation information can be obtained.

[0062] For example, the current guardianship recommendation information is "Measure blood pressure or heart rate values every 2 hours, and avoid strenuous exercise and emotional excitement". The historical text information includes multiple question information of the user and the corresponding system intelligent recommendation information, such as one of them is "Feeling a bit dizzy currently, how to deal with it?" and "It is recommended to measure blood pressure again after sitting still for 10 minutes". The current user input information is "The currently measured blood pressure value is XX1". By combining the above texts to form a current comprehensive input information containing rich context information and inputting it into the pre-trained language model, the current intelligent recommendation information can be obtained.

[0063] In one embodiment, after step S142, it further includes:

[0064] Obtain the privacy protection degree of the current intelligent recommendation information based on a preset verification strategy; where the formula corresponding to obtaining the privacy protection degree of the intelligent recommendation information in the verification strategy is the privacy protection degree of the information , M is the total number of privacy entities included in the intelligent recommendation information, Sen i represents the privacy entity sensitivity value of the i-th privacy entity in the intelligent recommendation information, and Id i represents the identifiability value of the i-th privacy entity in the intelligent recommendation information.

[0065] If it is determined that the privacy protection level of the current intelligent recommendation information is greater than or equal to a preset privacy protection level threshold, then a trusted protection label is added to the current intelligent recommendation information correspondingly to update the current intelligent recommendation information.

[0066] In this embodiment, in order to ensure that the current intelligent recommendation information sent to the user terminal by the information intelligent recommendation system has good privacy protection, the privacy protection level of the current intelligent recommendation information can also be obtained by a verification policy pre-deployed by the information intelligent recommendation system. The formula corresponding to obtaining the privacy protection level of the intelligent recommendation information in the verification policy is the privacy protection level of the information , M is the total number of privacy entities included in the intelligent recommendation information, Sen i represents the privacy entity sensitivity value of the i-th privacy entity in the intelligent recommendation information, and Id i represents the identifiability value of the i-th privacy entity in the intelligent recommendation information. More specifically, when determining the total number of privacy entities in the current intelligent recommendation information, the current intelligent recommendation information can be compared and matched with multiple privacy entities included in the preset privacy entity set (for example, the preset privacy entity set includes privacy entities such as name, ID number, etc.), and the total number of privacy entities existing in the preset privacy entity set included in the current intelligent recommendation information is determined. Moreover, a privacy entity sensitivity initial value is preset for each of the multiple privacy entities included in the preset privacy entity set. When the comparison and matching between the current intelligent recommendation information and the multiple privacy entities included in the preset privacy entity set are completed, the privacy entity sensitivity initial value corresponding to each privacy entity in the current intelligent recommendation information can also be initially determined. Then, the privacy entity sensitivity value can be adjusted correspondingly based on the number of occurrences of each privacy entity in the text of the current intelligent recommendation information. For example, each time a privacy entity appears more in the text of the current intelligent recommendation information, its current privacy entity sensitivity value is subtracted, that is, the final privacy entity sensitivity value corresponding to the i-th privacy entity in the current intelligent recommendation information = the i-th privacy entity sensitivity initial value - k i × the preset privacy entity sensitivity adjustment value, where k i is the total number of occurrences of the i-th privacy entity in the current intelligent recommendation information. When determining the identifiability value of the i-th privacy entity in the current intelligent recommendation information, the i-th privacy entity can be compared and matched with the preset identifiable entity set (the preset identifiable entity set includes multiple identifiable entities, and each identifiable entity corresponds to an identifiability value), so as to determine the identifiable entity with the same entity name as the i-th privacy entity and the corresponding identifiability value in the preset identifiable entity set, and use the identifiability value corresponding to the identifiable entity as the identifiability value of the i-th privacy entity.

[0067] It can be seen that through the above process, the privacy protection level of the current intelligent recommendation information can be quickly determined by the verification strategy of the information intelligent recommendation system. When it is determined that the privacy protection level of the current intelligent recommendation information is greater than or equal to the preset privacy protection level threshold, it indicates that the current intelligent recommendation information has good privacy protection. At this time, a trusted protection label can be added to the current intelligent recommendation information to update the current intelligent recommendation information. For the current intelligent recommendation information with a trusted protection label, the information intelligent recommendation system can send it to the current target user terminal corresponding to the current target user without any interception or processing measures and release it.

[0068] In one embodiment, after step S130, it further includes:

[0069] If it is determined that the current information type is an image information type, then based on the image text recognition model, obtain the image recognition text information corresponding to the information entered by the current user;

[0070] If it is determined that the image recognition text information is physical sign detection data information, then obtain the previous physical sign detection data information corresponding to the current target user, and display the previous physical sign detection data information and the image recognition text information side by side.

[0071] In this embodiment, if it is determined that the current information type is an image information type, it means that the current target user needs to intelligently recognize the text content in the image information, such as physical sign detection data information, and then perform corresponding processing. At this time, it can be based on an image text recognition model such as a convolutional recurrent neural network to obtain the image recognition text information corresponding to the information entered by the current user. When it is determined that it is physical sign detection data information and includes several physical sign detection data values, then obtain the previous physical sign detection data information corresponding to the current target user (for example, the time interval between the information generation time of the previous physical sign detection data information corresponding to the current target user and the input time of the information entered by the current user is less than the time interval between the information generation time of other physical sign detection data information and the input time of the information entered by the current user). Finally, display the previous physical sign detection data information and the image recognition text information side by side, and display this information in a side-by-side comparison manner so that the user can more intuitively view the data change trend.

[0072] S150. Send the current intelligent recommendation information to the current target user terminal corresponding to the current target user.

[0073] In this embodiment, after the current intelligent recommendation information corresponding to the current user input information sent by the current target user is obtained in the information intelligent recommendation system of the server, the server sends the current intelligent recommendation information to the current target user terminal corresponding to the current target user, so that the current target user can view it in time.

[0074] S160. If the current feedback label sent by the current target user terminal and corresponding to the current intelligent recommendation information is received, add the current feedback label to the current intelligent recommendation information and store it in the local storage space.

[0075] In this embodiment, in order to implement the in-flight training (i.e., continuous training) of the pre-trained language model, the current feedback label of the user for each intelligent recommendation information (such as a label indicating accurate reply, inaccurate reply, or unable to judge accuracy, etc.) can be continuously obtained. For example, the current feedback label corresponding to the current intelligent recommendation information is obtained, and the current feedback label is added to the current intelligent recommendation information and stored in the local storage space as a data set for subsequent continuous in-flight training of the pre-trained language model.

[0076] In one embodiment, after step S160, it further includes:

[0077] If it is detected that the current system time meets the preset model online training condition, obtain a training set composed of multiple stored intelligent recommendation information including feedback labels, and perform model training on the pre-trained language model until the preset model training stop condition is met, so as to update the pre-trained language model.

[0078] In this embodiment, for example, the preset model online training condition is set such that the time interval between the current system time and the last model online training is equal to the preset model online training period. When it is detected that the current system time meets the preset model online training condition, it means that a training set composed of multiple stored intelligent recommendation information including feedback labels can be obtained currently (wherein, the first three sentences of the context information of each intelligent recommendation information in the training set are also known). Perform model training on the pre-trained language model until the preset model training stop condition (such as the convergence of the loss function, etc.) is met, so as to update the pre-trained language model, thereby realizing the continuous update of the pre-trained language model and making it have better model performance.

[0079] It can be seen that the embodiment of implementing this method can combine the patient label of the current target user, the current monitoring recommendation information, the current user input information, and the historical text information to form the current comprehensive input information containing more information, and then input it into the pre-trained language model for the medical field to obtain more accurate intelligent recommendation information that references more context information.

[0080] An embodiment of the present invention further provides an intelligent recommendation device for cardiovascular-related monitoring information. The intelligent recommendation device for cardiovascular-related monitoring information can be configured in a server and is used to execute any one of the foregoing embodiments of the intelligent recommendation method for cardiovascular-related monitoring information. Specifically, please refer to Figure 6 , Figure 6 which is a schematic block diagram of the intelligent recommendation device for cardiovascular-related monitoring information provided by an embodiment of the present invention. As Figure 6 shown, the intelligent recommendation device 100 for cardiovascular-related monitoring information includes a target user information acquisition unit 110, a monitoring recommendation information acquisition unit 120, an information type acquisition unit 130, a current intelligent recommendation information acquisition unit 140, a current intelligent recommendation information sending unit 150, and a current feedback tag storage unit 160.

[0081] The target user information acquisition unit 110 is configured to, in response to an information intelligent recommendation instruction, acquire current target user information corresponding to the information intelligent recommendation instruction, and acquire a current target time interval corresponding to the intelligent recommendation instruction.

[0082] In this embodiment, the technical solution is described with the server as the execution subject. An information intelligent recommendation system (similar to an intelligent customer service platform that can reply to user questions or texts sent) is deployed in the server. When a user logs in to the information intelligent recommendation system in the server using a user terminal, the user can interact with the information intelligent recommendation system in the server to obtain the required intelligent recommendation information.

[0083] Among them, when a user logs in to the information intelligent recommendation system in the server using a user terminal, on the premise that the user authorizes the acquisition of their user information, the server can, in response to an information intelligent recommendation instruction, acquire current target user information corresponding to the information intelligent recommendation instruction, and at least include information such as user ID, user name, user gender, user age, and user patient label in the current target user information. Moreover, a current target time interval corresponding to the current target user information can also be obtained based on the instruction generation time of the intelligent recommendation instruction. The information intelligent recommendation system in this application is a system that focuses more on more accurate monitoring information intelligent recommendation for users with various patient labels and can provide more intelligent and accurate replies to the information feedback sent by users in the information intelligent recommendation system.

[0084] In one embodiment, the target user information acquisition unit 110 is specifically configured to:

[0085] Acquire current user classification label information in the current target user information, and determine a current acquisition period based on the current user classification label information and a preset label and period mapping relationship;

[0086] Obtain the instruction generation time of the information intelligent recommendation instruction, determine the start time of the current target time interval based on the difference between the instruction generation time and the current acquisition period, and use the instruction generation time as the end time of the current target time interval;

[0087] Determine the current target time interval according to the start time of the current target time interval and the end time of the current target time interval.

[0088] In this embodiment, first, the current user classification label information can be obtained from the current target user information, such as a high-risk user classification label, a medium-risk user classification label, or a low-risk user classification label, etc. Moreover, a mapping relationship between each user classification label information and the acquisition period duration is also set in the server. For example, it is set that the high-risk user classification label corresponds to the acquisition period T1, the medium-risk user classification label corresponds to the acquisition period T2, and the low-risk user classification label corresponds to the acquisition period T3 (where T1, T2, and T3 are all positive numbers), and it is determined that the current user classification label information corresponds to the medium-risk user classification label. At this time, the current acquisition period can be determined as the acquisition period T2 according to the label and period mapping relationship.

[0089] After that, combined with the instruction generation time of the information intelligent recommendation instruction and the current acquisition period (such as the acquisition period T2 in the above example), specifically, the start time of the current target time interval can be determined based on the difference between the instruction generation time and the current acquisition period, and the instruction generation time is used as the end time of the current target time interval. Finally, after determining the start time point and end time point of the current target time interval, the determination of the current target time interval can be completed. Moreover, the obtained current target time interval is used to determine the communication history data (which can also be understood as historical context data) corresponding to the current target user information.

[0090] The guardianship recommendation information acquisition unit 120 is used to, if it is determined that the current target user information includes a cardiovascular disease patient label, obtain the current target user corresponding to the current target user information, and obtain the current guardianship recommendation information corresponding to the current target user and stored in the local storage space.

[0091] In this embodiment, since the current target user information includes information of a user patient label, after parsing and determining that it is a cardiovascular disease patient label, the current target user corresponding to the current target user information can be further obtained, and the current guardianship recommendation information corresponding to the current target user and stored in the local storage space can be obtained.

[0092] Among them, the current guardianship recommendation information is uploaded to the server by other business servers and edited and uploaded by doctor users for the current target user. For example, the current guardianship recommendation information includes doctor's advice information for the current target user within a specified time period (this specified time period can be understood as the information effective time period corresponding to the current guardianship recommendation information). More specifically, it can include recommended diet information, medication information, daily rehabilitation exercise information, etc. It can also include the monitoring thresholds of key physical sign indicators for the current target user within the current specified time period (for example, once the key physical sign indicators of the current target user exceed the monitoring thresholds of the key physical sign indicators multiple times within the current specified time period, it means that a reexamination needs to be carried out in a timely manner). Moreover, after the user logs in to the information intelligent recommendation system of the server using the user terminal, the current guardianship recommendation information corresponding to the current target user can either be hidden information and not displayed on the display interface of the information intelligent recommendation system (such as a chat box-style display interface), or be non-hidden information directly displayed on the display interface of the information intelligent recommendation system.

[0093] The information type acquisition unit 130 is configured to acquire the current information type of the current user input information if it detects that the current user inputs information.

[0094] In this embodiment, after the user logs in to the information intelligent recommendation system using the intelligent terminal, one of the common usage methods is to input the current user input information. For example, it can be text input information (such as Chinese text mainly composed of Chinese characters, English text mainly composed of English characters, digital text mainly composed of digital characters, etc.), voice input information, image input information, etc. However, no matter in what way the user inputs the relevant information, the current information type of the current user input information can be acquired first.

[0095] In one embodiment, the information type acquisition unit 130 is specifically configured to:

[0096] Acquire the file format of the current user input information;

[0097] If it is determined that the file format corresponds to a text file format, use the text information type as the current information type;

[0098] If it is determined that the file format corresponds to a picture file format, use the picture information type as the current information type;

[0099] If it is determined that the file format corresponds to a voice file format, use the voice information type as the current information type.

[0100] In this embodiment, when determining the current information type of the currently entered information by the user, the fastest implementation method is to obtain its file format, more specifically, to obtain the file suffix name of the user-entered information. After quickly determining the current information type in combination with the file suffix name, corresponding responses can be further made in combination with the current information type and the specific information content.

[0101] Specifically, if it is determined that the file format corresponds to the text file format, such as txt, then the text information type is used as the current information type; if it is determined that the file format corresponds to the picture file format, such as jpg, then the picture information type is used as the current information type; if it is determined that the file format corresponds to the voice file format, such as wma, then the voice information type is used as the current information type. It can be seen that the current information type can be quickly determined through the above method.

[0102] The current intelligent recommendation information acquisition unit 140 is configured to, if it is determined that the current information type is the text information type, obtain historical text information corresponding to the current target user and with the information generation time within the current target time interval, and use the currently entered information by the user, the current guardianship recommendation information, and the historical text information to form the current comprehensive entered information and input it into the pre-trained language model to obtain the current intelligent recommendation information.

[0103] Among them, the pre-trained language model is a pre-trained language model for the medical field.

[0104] In this embodiment, if it is determined that the current information type is the text information type, it can be further determined whether it is a Chinese text mainly composed of Chinese characters, an English text mainly composed of English characters, or a digital text mainly composed of digital characters. For example, if it is determined that the currently entered information by the user is mainly Chinese character input (that is, the proportion of Chinese characters in the currently entered information by the user exceeds the first character proportion threshold), it means that the current target user is mainly for consultation and questions. The historical text information corresponding to the current target user and with the information generation time within the current target time interval and the current guardianship recommendation information can be combined, and the above three types of information are used to form a current comprehensive entered information and input it into the pre-trained language model (such as the BioBERT model) to obtain the current intelligent recommendation information output by the pre-trained language model.

[0105] In one embodiment, the current intelligent recommendation information acquisition unit 140 is specifically configured to:

[0106] In the order of the current guardianship recommendation information, the historical text information, and the currently entered information by the user, the current guardianship recommendation information, the historical text information, and the currently entered information by the user are used to form the current comprehensive entered information;

[0107] Input the current input vector corresponding to the current comprehensive input information into the pre-trained language model to obtain the current intelligent recommendation information.

[0108] In this embodiment, in order to input more accurate current intelligent recommendation information by combining more context information, the currently displayed or hidden current guardianship recommendation information can be used as the context data ranked first in the current comprehensive input information, and the historical text information can be used as the context data ranked in the middle of the current comprehensive input information. Combining the current user input information, a current comprehensive input information that fully considers the context data such as the current guardianship recommendation information, historical question information, and system historical reply information of the current target user can be formed. After performing text tokenization processing (i.e., Tokenization, to split the text into tokens) and text encoding processing (to map the tokens to a low-dimensional vector space) on it to obtain the current input vector and then inputting it into the pre-trained language model, the current intelligent recommendation information can be obtained.

[0109] For example, the current guardianship recommendation information is "Measure blood pressure or heart rate values every 2 hours, and avoid strenuous exercise and emotional excitement". The historical text information includes multiple question information of the user and corresponding system intelligent recommendation information, such as one of them is "Feeling a bit dizzy currently, how to deal with it?" and "It is recommended to measure blood pressure again after sitting still for 10 minutes". The current user input information is "The currently measured blood pressure value is XX1". By synthesizing the above texts to form a current comprehensive input information containing rich context information and inputting it into the pre-trained language model, the current intelligent recommendation information can be obtained.

[0110] In one embodiment, the current intelligent recommendation information acquisition unit 140 is further configured to:

[0111] Obtain the privacy protection degree of the current intelligent recommendation information based on a preset verification policy; where the formula corresponding to obtaining the privacy protection degree of the intelligent recommendation information in the verification policy is the privacy protection degree of the information , M is the total number of privacy entities included in the intelligent recommendation information, Sen i represents the privacy entity sensitivity value of the i-th privacy entity in the intelligent recommendation information, and Id i represents the identifiability value of the i-th privacy entity in the intelligent recommendation information;

[0112] If it is determined that the privacy protection degree of the current intelligent recommendation information is greater than or equal to a preset privacy protection degree threshold, then add a trusted protection label to the current intelligent recommendation information to update the current intelligent recommendation information.

[0113] In this embodiment, in order to ensure that the current intelligent recommendation information sent to the user terminal by the information intelligent recommendation system has better privacy protection, the privacy protection degree of the current intelligent recommendation information can also be obtained by the verification policy pre-deployed by the information intelligent recommendation system. The formula corresponding to obtaining the privacy protection degree of the intelligent recommendation information in the verification policy is the privacy protection degree of the information , M is the total number of privacy entities included in the intelligent recommendation information, Sen i represents the privacy entity sensitivity value of the i-th privacy entity in the intelligent recommendation information, and Id i represents the identifiability value of the i-th privacy entity in the intelligent recommendation information. More specifically, when determining the total number of privacy entities in the current intelligent recommendation information, the current intelligent recommendation information can be compared and matched with multiple privacy entities included in the preset privacy entity set (for example, the preset privacy entity set includes privacy entities such as names and ID numbers), and the total number of privacy entities existing in the preset privacy entity set included in the current intelligent recommendation information is determined. Moreover, each of the multiple privacy entities included in the preset privacy entity set is preset with an initial privacy entity sensitivity value. When the comparison and matching between the current intelligent recommendation information and the multiple privacy entities included in the preset privacy entity set are completed, the initial privacy entity sensitivity value corresponding to each privacy entity in the current intelligent recommendation information can also be initially determined. Then, the privacy entity sensitivity value can be adjusted correspondingly based on the number of occurrences of each privacy entity in the text of this article. For example, for each time a privacy entity appears more in the text of the current intelligent recommendation information, its current privacy entity sensitivity value is subtracted, that is, the final privacy entity sensitivity value corresponding to the i-th privacy entity in the current intelligent recommendation information = the initial privacy entity sensitivity value of the i-th privacy entity - k i × the preset privacy entity sensitivity adjustment value, where k i is the total number of occurrences of the i-th privacy entity in the current intelligent recommendation information. When determining the identifiability value of the i-th privacy entity in the current intelligent recommendation information, the i-th privacy entity can be compared and matched with the preset identifiability entity set (the preset identifiability entity set includes multiple identifiability entities, and each identifiability entity corresponds to an identifiability value), so as to determine the identifiability entity with the same entity name as the i-th privacy entity and the corresponding identifiability value in the preset identifiability entity set, and use the identifiability value corresponding to this identifiability entity as the identifiability value of the i-th privacy entity.

[0114] It can be seen that through the above process, the privacy protection degree of the current intelligent recommendation information can be quickly determined by the verification strategy of the information intelligent recommendation system. When it is determined that the privacy protection degree of the current intelligent recommendation information is greater than or equal to the preset privacy protection degree threshold, it indicates that the current intelligent recommendation information has good privacy protection. At this time, a trusted protection label can be added to the current intelligent recommendation information to update the current intelligent recommendation information. The information intelligent recommendation system can send the current intelligent recommendation information with the trusted protection label to the current target user terminal corresponding to the current target user without any interception or processing measures and release it.

[0115] In one embodiment, the cardiovascular-related monitoring information intelligent recommendation device 100 further includes:

[0116] An image text recognition unit, configured to, if it is determined that the current information type is an image information type, obtain image recognition text information corresponding to the current user input information based on an image text recognition model;

[0117] A data juxtaposition display unit, configured to, if it is determined that the image recognition text information is physical sign detection data information, obtain the previous physical sign detection data information corresponding to the current target user, and juxtapositionally display the previous physical sign detection data information and the image recognition text information.

[0118] In this embodiment, if it is determined that the current information type is an image information type, it means that the current target user needs to intelligently recognize the text content in the image information, such as physical sign detection data information, and then perform corresponding processing. At this time, it can be based on an image text recognition model, such as a convolutional recurrent neural network, to obtain image recognition text information corresponding to the current user input information. When it is determined that it is physical sign detection data information and includes several physical sign detection data values, then obtain the previous physical sign detection data information corresponding to the current target user (for example, the time interval between the information generation time of the previous physical sign detection data information corresponding to the current target user and the input time of the current user input information is less than the time interval between the information generation time of other physical sign detection data information and the input time of the current user input information). Finally, the previous physical sign detection data information and the image recognition text information are juxtapositionally displayed, and this information is displayed in a juxtaposed comparison manner so that the user can more intuitively view the data change trend.

[0119] A current intelligent recommendation information sending unit 150, configured to send the current intelligent recommendation information to the current target user terminal corresponding to the current target user.

[0120] In this embodiment, after the current intelligent recommendation information corresponding to the current user input information sent by the current target user is obtained in the information intelligent recommendation system of the server, the server sends the current intelligent recommendation information to the current target user terminal corresponding to the current target user, so that the current target user can view it in time.

[0121] The current feedback label storage unit 160 is configured to, if receiving the current feedback label sent by the current target user terminal and corresponding to the current intelligent recommendation information, add the current feedback label to the current intelligent recommendation information and store it in the local storage space.

[0122] In this embodiment, in order to implement the in-flight training (i.e., continuous training) of the pre-trained language model, the current feedback label of the user for each intelligent recommendation information (such as a label indicating accurate reply, or inaccurate answer, or a label that cannot judge whether it is accurate, etc.) can also be continuously obtained. For example, the current feedback label corresponding to the current intelligent recommendation information is obtained, and the current feedback label is added to the current intelligent recommendation information and stored in the local storage space as a data set for subsequent continuous in-flight training of the pre-trained language model.

[0123] In one embodiment, the cardiovascular-related monitoring information intelligent recommendation device 100 further includes:

[0124] The model online training unit is configured to, if it detects that the current system time meets the preset model online training condition, obtain a training set composed of multiple stored intelligent recommendation information including feedback labels, and perform model training on the pre-trained language model until the preset model training stop condition is met, so as to update the pre-trained language model.

[0125] In this embodiment, for example, the preset model online training condition is set such that the time interval between the current system time and the last model online training is equal to the preset model online training period. When it is detected that the current system time meets the preset model online training condition, it means that a training set composed of multiple stored intelligent recommendation information including feedback labels can be obtained currently (wherein, the first three sentences of the context information of each intelligent recommendation information in the training set are also known). Perform model training on the pre-trained language model until the preset model training stop condition (such as the convergence of the loss function, etc.) is met, so as to update the pre-trained language model, thereby realizing the continuous update of the pre-trained language model and making it have better model performance.

[0126] It can be seen that when the embodiment of the device is implemented, the patient label of the current target user, the current monitoring recommendation information, the current user input information, and the historical text information can be combined to form the current comprehensive input information containing more information, and then input into the pre-trained language model for the medical field, so as to obtain more accurate intelligent recommendation information with more context information taken into account.

[0127] The above-mentioned intelligent recommendation device for cardiovascular-related monitoring information can be implemented in the form of a computer program, and this computer program can run on a computer device as shown in Figure 7 the following.

[0128] Please refer to Figure 7 , Figure 7 which is a schematic block diagram of a computer device provided by an embodiment of the present invention. This computer device integrates any one of the intelligent recommendation devices for cardiovascular-related monitoring information provided by the embodiments of the present invention.

[0129] Referring to Figure 7 , this computer device 400 includes a processor 402, a memory, and a network interface 405 connected through a system bus 401. Among them, the memory may include a storage medium 403 and an internal memory 404.

[0130] The storage medium 403 can store an operating system 4031 and a computer program 4032. This computer program 4032 includes program instructions, and when these program instructions are executed, the processor 402 can be made to execute the above-mentioned intelligent recommendation method for cardiovascular-related monitoring information.

[0131] The processor 402 is used to provide computing and control capabilities to support the operation of the entire computer device.

[0132] The internal memory 404 provides an environment for the operation of the computer program 4032 in the storage medium 403. When this computer program 4032 is executed by the processor 402, the processor 402 can be made to execute the above-mentioned intelligent recommendation method for cardiovascular-related monitoring information.

[0133] The network interface 405 is used for network communication with other devices. Those skilled in the art can understand that Figure 7 the structure shown in

[0134] is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0135] It should be understood that in the embodiments of the present invention, the processor 402 may be a central processing unit (CPU), and the processor 402 may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0136] Those of ordinary skill in the art can understand that all or part of the processes in the methods of implementing the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program includes program instructions, and the computer program can be stored in a storage medium, and the storage medium is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer device to implement the process steps of the above method embodiments.

[0137] Therefore, the present invention also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, where the computer program includes program instructions. When the program instructions are executed by a processor, the processor is caused to execute the intelligent recommendation method for cardiovascular-related monitoring information as described above.

[0138] The computer-readable storage medium may be various computer-readable storage media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a magnetic disk, or an optical disc that can store program codes.

[0139] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of the present invention.

[0140] In several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of each unit is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed.

[0141] The steps in the method embodiments of the present invention can be adjusted, combined, and deleted according to actual needs. The units in the device embodiments of the present invention can be combined, divided, and deleted according to actual needs. In addition, the functional units in each embodiment of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0142] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present invention.

[0143] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. An intelligent recommendation method for cardiovascular-related monitoring information, which is applied to a server, characterized in that, Including: In response to an information intelligent recommendation instruction, obtain the current target user information corresponding to the information intelligent recommendation instruction, and obtain the current target time interval corresponding to the intelligent recommendation instruction; If it is determined that the current target user information includes a cardiovascular disease patient label, obtain the current target user corresponding to the current target user information, and obtain the current guardianship recommendation information corresponding to the current target user and stored in the local storage space; If the current user input information is detected, obtain the current information type of the current user input information; If it is determined that the current information type is a text information type, obtain the historical text information corresponding to the current target user and with the information generation time within the current target time interval, and form the current comprehensive input information with the current user input information, the current guardianship recommendation information, and the historical text information, and input it into the pre-trained language model to obtain the current intelligent recommendation information; wherein, the pre-trained language model is a pre-trained language model for the medical field; Send the current intelligent recommendation information to the current target user terminal corresponding to the current target user; If the current feedback label sent by the current target user terminal and corresponding to the current intelligent recommendation information is received, add the current feedback label to the current intelligent recommendation information and store it in the local storage space; The forming the current comprehensive input information with the current user input information, the current guardianship recommendation information, and the historical text information, and inputting it into the pre-trained language model to obtain the current intelligent recommendation information includes: In the order of the current guardianship recommendation information, the historical text information, and the current user input information, form the current comprehensive input information with the current guardianship recommendation information, the historical text information, and the current user input information; Input the current input vector corresponding to the current comprehensive input information into the pre-trained language model to obtain the current intelligent recommendation information; Among them, the directly displayed or hidden current guardianship recommendation information is used as the context data ranked first in the current comprehensive input information, the historical text information is used as the context data ranked in the middle of the current comprehensive input information, and combined with the current user input information, form the current comprehensive input information, and perform text tokenization processing and text encoding processing on the current comprehensive input information to obtain the current input vector.

2. The intelligent recommendation method for cardiovascular-related monitoring information according to claim 1, wherein The obtaining the current information type of the current user input information includes: Obtain the file format of the current user input information; If it is determined that the file format corresponds to a text file format, use the text information type as the current information type; If it is determined that the file format corresponds to a picture file format, use the picture information type as the current information type; If it is determined that the file format corresponds to a voice file format, use the voice information type as the current information type.

3. The intelligent recommendation method for cardiovascular-related monitoring information according to claim 1, wherein After the step of inputting the current input vector corresponding to the current comprehensive input information into the pre-trained language model to obtain the current intelligent recommendation information, the method further includes: Obtain the privacy protection degree of the current intelligent recommendation information based on a preset verification strategy; wherein, the formula corresponding to obtaining the privacy protection degree of the intelligent recommendation information in the verification strategy is , M is the total number of privacy entities included in the intelligent recommendation information, Seni represents the privacy entity sensitivity value of the i-th privacy entity in the intelligent recommendation information, and Idi represents the identifiability value of the i-th privacy entity in the intelligent recommendation information; If it is determined that the privacy protection degree of the current intelligent recommendation information is greater than or equal to the preset privacy protection degree threshold, a trusted protection label is added to the current intelligent recommendation information to update the current intelligent recommendation information.

4. The intelligent recommendation method for cardiovascular-related monitoring information according to claim 1, wherein After the step of, if it is detected that the current user enters information, obtaining the current information type of the current user-entered information, the method further includes: If it is determined that the current information type is an image information type, image recognition text information corresponding to the current user-entered information is obtained based on an image text recognition model; If it is determined that the image recognition text information is physical sign detection data information, the previous physical sign detection data information corresponding to the current target user is obtained, and the previous physical sign detection data information and the image recognition text information are displayed side by side.

5. The intelligent recommendation method for cardiovascular-related monitoring information according to claim 1, wherein After the step of, if the current feedback label sent by the current target user terminal and corresponding to the current intelligent recommendation information is received, adding the current feedback label to the current intelligent recommendation information and storing it in the local storage space, the method further includes: If it is detected that the current system time meets the preset model online training condition, a training set composed of multiple stored intelligent recommendation information including feedback labels is obtained, and the pre-trained language model is trained until the preset model training stop condition is met to update the pre-trained language model.

6. The intelligent recommendation method for cardiovascular-related monitoring information according to claim 1, wherein, The obtaining the current target user information corresponding to the information intelligent recommendation instruction and obtaining the current target time interval corresponding to the intelligent recommendation instruction includes: Obtaining the current user classification label information in the current target user information, and determining the current acquisition period based on the current user classification label information and the preset label and period mapping relationship; Obtaining the instruction generation time of the information intelligent recommendation instruction, determining the start time of the current target time interval by the difference between the instruction generation time and the current acquisition period, and using the instruction generation time as the end time of the current target time interval; According to the start time of the current target time interval and the end time of the current target time interval, determining the current target time interval.

7. An intelligent recommendation device for cardiovascular-related monitoring information, configured in a server, characterized in that, Including: A target user information acquisition unit, configured to, in response to an information intelligent recommendation instruction, acquire current target user information corresponding to the information intelligent recommendation instruction and acquire a current target time interval corresponding to the intelligent recommendation instruction; A guardianship recommendation information acquisition unit, configured to, if it is determined that the current target user information includes a cardiovascular disease patient label, acquire a current target user corresponding to the current target user information and acquire current guardianship recommendation information corresponding to the current target user and stored in the local storage space; An information type acquisition unit, configured to, if it is detected that the current user enters information, acquire the current information type of the current user-entered information; The current intelligent recommendation information acquisition unit is used to, if it is determined that the current information type is a text information type, acquire historical text information corresponding to the current target user and with an information generation time within the current target time interval, form current comprehensive input information with the current user input information, the current guardianship recommendation information, and the historical text information, and input the current comprehensive input information into a pre-trained language model to obtain current intelligent recommendation information; wherein, the pre-trained language model is a pre-trained language model for the medical field; The current intelligent recommendation information sending unit is used to send the current intelligent recommendation information to the current target user terminal corresponding to the current target user; The current feedback label storage unit is used to, if it receives the current feedback label sent by the current target user terminal and corresponding to the current intelligent recommendation information, add the current feedback label to the current intelligent recommendation information and store it in the local storage space; The current intelligent recommendation information acquisition unit is specifically used for: According to the order of the current guardianship recommendation information, the historical text information, and the current user input information, form the current comprehensive input information with the current guardianship recommendation information, the historical text information, and the current user input information; Input the current input vector corresponding to the current comprehensive input information into the pre-trained language model to obtain the current intelligent recommendation information.

8. A computer device, the computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the computer device executes the computer program, it implements the intelligent recommendation method for cardiovascular-related guardianship information as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the intelligent recommendation method for cardiovascular-related guardianship information as described in any one of claims 1-6.

Images