A method, system and electronic device for predicting health risks of the elderly
The method and system address inefficiencies in current assessment methods by using a health risk assessment decision model and prediction model to dynamically evaluate and predict future health risks in the elderly, improving accuracy and compliance.
Patent Information
- Application Number
- CN202111671510.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-31
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2041-12-31
AI Technical Summary
Current methods for assessing cognitive impairment and frailty in the elderly are inefficient, costly, and lack the ability to predict future risks, with fixed assessment scales leading to low accuracy and poor user compliance.
A method and system that uses a health risk assessment decision model to generate various assessment schemes, incorporating scale, scenario, and direct condition evaluations, and a health risk prediction model to predict future risks based on graph neural networks and decision trees, utilizing health data from cameras, sound collection devices, and questionnaires.
Improves assessment accuracy, reduces costs, and enhances user compliance by automatically evaluating and predicting future health risks of cognitive impairment and frailty in the elderly.
Smart Images

Figure CN114429817B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of health risk prediction, and in particular to a method, system and electronic equipment for predicting health risks for the elderly. Background Art
[0002] The current aging problem is prominent, and the proportion of the elderly population is gradually increasing. The health status of the elderly directly affects the cost of family care. In addition to common diseases, the health of the elderly also includes cognitive impairment and frailty. Once the elderly suffer from cognitive impairment or frailty, it will seriously affect the quality of life and bring great pressure to the family. In this regard, the focus of research is to detect it as early as possible.
[0003] Currently, the findings of cognitive impairment and frailty in the elderly are mostly made using assessment scales, that is, manually recorded through interviews. This method has some disadvantages: for example, the questions in the assessment scales used in the process are generally fixed, and the accuracy of the assessment will be affected due to repeated practice when the elderly are assessed multiple times; the interview process requires medical staff or psychological practitioners to ask questions on the spot, which is inefficient and costly; more importantly, the existing methods are a summary of the current situation and cannot predict future situations, which is very important for cognitive impairment, frailty, etc., to predict as far in advance as possible. Summary of the invention
[0004] In view of this, the purpose of the present invention is to provide a method, system and electronic device for predicting health risks for the elderly. The method can use a health risk assessment decision model to randomly generate various assessment schemes, and automatically evaluate the cognition and frailty status of elderly users through scale evaluation, scenario evaluation, direct condition evaluation, etc., and based on this, predict the future risks of cognitive impairment, frailty, etc. for the elderly, which solves the problems of low assessment accuracy, high assessment cost, and inability to predict future risks of traditional methods, and improves user compliance and efficiency.
[0005] In a first aspect, an embodiment of the present invention provides a method for predicting health risks of the elderly, the method comprising:
[0006] Obtaining the user's health data; wherein the health data includes at least: an image of the user taken by a camera, the user's voice data obtained by a sound collection device, and the user's answer data obtained by answering questionnaires;
[0007] The health data is input into a preset health risk assessment decision model, and the health risk assessment scheme of the user is determined from the output results of the health risk assessment decision model; wherein the health risk assessment scheme includes at least: a scale assessment scheme, a scenario assessment scheme, and a direct condition assessment scheme; the scale assessment scheme assesses the user's cognitive state and frailty state through questions generated by voice and text; the scenario assessment scheme assesses the user's cognitive state and frailty state through the completion of scenario interaction questions; the direct condition assessment scheme assesses the user's cognitive state and frailty state through the user's interaction results obtained by sensors;
[0008] The health risk assessment scheme is used to determine the user's health risk assessment data, and after the health risk assessment data is input into a preset health risk prediction model, the user's health risk prediction result is determined from the output result of the health risk prediction model.
[0009] In some embodiments, the process of inputting health data into a preset health risk assessment decision model includes:
[0010] The health data is input into the trained health risk assessment decision model, and the health risk assessment decision model makes statistical decisions on the health data based on the built-in assessment method selection template; wherein the assessment method selection template is a hierarchical tree structure, including: an assessment target layer, an assessment dimension layer, an assessment method layer, an assessment implementation layer, and an assessment content layer; the assessment content layer is used to load the assessment content corresponding to the health data;
[0011] Based on the assessment target layer, determine the assessment targets corresponding to the health data; wherein the assessment targets include: cognitive assessment and / or frailty assessment;
[0012] Based on the evaluation dimension layer, determine the corresponding evaluation dimensions under the evaluation target; the evaluation dimensions include: attention, transaction processing speed, language interaction, visual spatial positioning, memory status, functional execution, mental arithmetic ability, weight loss, strength, physical fatigue, walking difficulty, physical activity, disease status, nutritional status, anxiety and depression, sleep status, and the evaluation process of one or more of the above dimensions;
[0013] Based on the evaluation method layer, determine the evaluation objectives and the corresponding evaluation methods under the evaluation dimensions; the evaluation methods at least include: scale evaluation, scenario evaluation, and direct situation evaluation;
[0014] Based on the assessment implementation layer, determine the assessment objectives, assessment dimensions, and corresponding assessment implementation data under the assessment method; wherein the assessment implementation data at least includes: scale name, scenario name, and direct situation assessment name.
[0015] In some embodiments, the process of determining the user's health risk assessment scheme from the output of the health risk assessment decision model includes:
[0016] Obtaining the output results of the health risk assessment decision model; wherein the health risk assessment decision model includes rules and statistical learning models; wherein the statistical learning model integrates random forests and neural networks; the health risk assessment decision model makes decisions on the user's health data, and obtains the assessment objectives, assessment dimensions, assessment methods, and assessment implementation corresponding to the health data from the output results of the health risk assessment decision model;
[0017] Determine the user's health risk assessment plan based on the assessment objectives, assessment dimensions, assessment methods, and assessment implementation corresponding to the user's health data.
[0018] In some embodiments, if the health risk assessment scheme includes a scale assessment scheme, the process of determining the health risk assessment data of the user using the health risk assessment scheme includes:
[0019] Select an assessment scale based on the health risk assessment scheme;
[0020] Obtain the content of the scale questions in the scale assessment plan;
[0021] Use the logical order of the scale questions to determine the order of the scale questions;
[0022] Dynamically generate scale questions according to the order of questions, and dynamically display the scale questions to users;
[0023] The user's answer results to the scale questions and the process parameters of the answers are determined as the user's health risk assessment data.
[0024] In some embodiments, if the health risk assessment scheme includes a scenario assessment scheme, the process of determining the health risk assessment data of the user using the health risk assessment scheme includes:
[0025] Select the assessment scenario according to the health risk assessment plan;
[0026] Obtain the scene interaction content in the scene evaluation plan and display the scene interaction content;
[0027] Using the scene interaction content to randomly determine the scene interaction questions, and present the scene interaction questions to the user;
[0028] The user's interaction results on the scene interaction questions and the interaction process parameters are determined as the user's health risk assessment data.
[0029] In some embodiments, if the health risk assessment protocol includes a direct condition assessment protocol, the process of determining the health risk assessment data of the user includes:
[0030] Select direct condition assessment projects based on the health risk assessment plan;
[0031] Obtain and determine the content of the condition assessment project in the direct condition assessment plan, and display the content of the condition assessment project to the user;
[0032] Using preset sensors to obtain sensor data generated by users interacting with the situation assessment project;
[0033] Utilize sensor data to determine health risk assessment data for users.
[0034] In some embodiments, after the health risk assessment data is input into a preset health risk prediction model, the health risk prediction result of the user is determined from the output result of the health risk prediction model, including:
[0035] After unifying the dimensions and dimensions of the health risk assessment data, calculate the current deviation value and progress deviation value of the health risk assessment data;
[0036] Input the health risk assessment data, current deviation value and progress deviation value into the trained health risk prediction model, and determine the probability value of the user developing cognitive impairment and frailty in the future according to the output result of the health risk prediction model; wherein the health risk prediction model is based on graph neural network and decision tree, and is trained using historical data and expert input data;
[0037] The user's health risk prediction result is determined based on the probability value of the user developing cognitive impairment and frailty in the future output by the health risk prediction model.
[0038] In some embodiments, the process of determining the health risk prediction result of the user further includes:
[0039] Determine the user's intervention items based on the health risk prediction results; the intervention items at least include: scenario intervention, exercise intervention, nutritional intervention, and psychological intervention;
[0040] Input the health risk prediction results and health risk assessment data into the trained intervention selection model, and determine the recommended intervention items for the user based on the output results of the intervention selection model; the intervention selection model is based on the non-monotonic inference algorithm of the graph neural network and is trained using historical data and expert input data;
[0041] Add recommended intervention items to the user's health risk prediction results.
[0042] In a second aspect, an embodiment of the present invention provides a system for predicting health risks of the elderly, the system comprising:
[0043] A health data input module is used to obtain the user's health data; wherein the health data at least includes: an image of the user taken by a camera, the user's voice data obtained by a sound collection device, and the user's answer data obtained by answering questions in a questionnaire;
[0044] A health risk assessment scheme generation module is used to input health data into a preset health risk assessment decision model, and determine the user's health risk assessment scheme from the output results of the health risk assessment decision model; wherein the health risk assessment scheme includes at least: a scale assessment scheme, a scenario assessment scheme, and a direct condition assessment scheme; the scale assessment scheme assesses the user's cognitive state and frailty state through questions generated by voice and text; the scenario assessment scheme assesses the user's cognitive state and frailty state through the completion of scenario interaction questions; the direct condition assessment scheme assesses the user's cognitive state and frailty state through the user's interaction results obtained by sensors;
[0045] The health risk prediction module is used to determine the user's health risk assessment data using a health risk assessment plan, and after inputting the health risk assessment data into a preset health risk prediction model, determine the user's health risk prediction result from the output result of the health risk prediction model.
[0046] In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein when the processor executes the computer program, the steps of the method for predicting health risks for the elderly mentioned in the first aspect are implemented.
[0047] In a fourth aspect, an embodiment of the present invention further provides a computer-readable medium having a non-volatile program code executable by a processor, wherein the program code enables the processor to execute the steps of the method for predicting health risks for the elderly mentioned in the first aspect above.
[0048] The embodiments of the present invention bring at least the following beneficial effects:
[0049] The present invention provides a method, system and electronic device for predicting health risks of the elderly. The method first obtains the health data of the user; wherein the health data at least includes: an image of the user taken by a camera, the sound data of the user obtained by a sound collection device, and the answer data of the user obtained by answering a questionnaire; then the health data is input into a preset health risk assessment decision model, and the health risk assessment scheme of the user is determined from the output result of the health risk assessment decision model; wherein the health risk assessment scheme at least includes: a scale assessment scheme, a scenario assessment scheme, and a direct condition assessment scheme; the scale assessment scheme evaluates the cognitive state and frail state of the user through questions generated by voice and text; the scenario assessment scheme evaluates the cognitive state and frail state of the user through the completion of scenario interaction questions; the direct condition assessment scheme evaluates the cognitive state and frail state of the user through the interaction results of the user obtained by sensors; finally, the health risk assessment data of the user is determined by the health risk assessment scheme, and after the health risk assessment data is input into the preset health risk prediction model, the health risk prediction result of the user is determined from the output result of the health risk prediction model. This method can use the evaluation decision model to randomly generate various evaluation schemes. It can automatically evaluate the cognitive, frail and other conditions of elderly users through scale evaluation, scenario evaluation, direct condition evaluation and other methods, and based on this, predict the future risks of cognitive impairment, frailty and other conditions of the elderly. It solves the problems of low evaluation accuracy, high evaluation cost, and inability to predict future risks of traditional methods, and improves user compliance and efficiency.
[0050] Other features and advantages of the present invention will be described in the following description, or some features and advantages can be inferred or determined without doubt from the description, or can be learned by implementing the above-mentioned technology of the present invention.
[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are specifically listed below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0053] Figure 1 A flowchart of a method for predicting health risks of the elderly provided by an embodiment of the present invention;
[0054] Figure 2A flowchart of a process of inputting health data into a preset health risk assessment decision model in a method for predicting health risks of the elderly provided by an embodiment of the present invention;
[0055] Figure 3 A flowchart of determining a health risk assessment scheme for a user from an output result of a health risk assessment decision model in a method for predicting health risks for the elderly provided by an embodiment of the present invention;
[0056] Figure 4 In a method for predicting health risks of the elderly provided by an embodiment of the present invention, if the health risk assessment scheme includes a scale assessment scheme, a flow chart of determining health risk assessment data of a user using the health risk assessment scheme;
[0057] Figure 5 In a method for predicting health risks of the elderly provided by an embodiment of the present invention, if the health risk assessment scheme includes a scenario assessment scheme, a flow chart of determining the health risk assessment data of the user using the health risk assessment scheme;
[0058] Figure 6 In a method for predicting health risks of the elderly provided by an embodiment of the present invention, if the health risk assessment scheme includes a direct condition assessment scheme, a flow chart of determining health risk assessment data of a user;
[0059] Figure 7 A flowchart of determining a user's health risk prediction result from an output result of a health risk prediction model in a method for predicting health risks of the elderly provided by an embodiment of the present invention;
[0060] Figure 8 A flowchart of determining a health risk prediction result of a user in a method for predicting health risks of the elderly provided by an embodiment of the present invention;
[0061] Fig. 9 A schematic diagram of the structure of a health risk prediction system for the elderly provided by an embodiment of the present invention;
[0062] Fig.10 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention.
[0063] icon:
[0064] 910-health data input module; 920-health risk assessment program generation module; 930-health risk prediction module;
[0065] 101 - processor; 102 - memory; 103 - bus; 104 - communication interface. DETAILED DESCRIPTION
[0066] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0067] At present, aging is becoming a global problem, and my country is also facing a relatively severe situation. my country is about to enter a deep aging stage, and the number of people over 60 years old will exceed 300 million, accounting for more than 20% of the total population.
[0068] As people age, especially those who are entering old age, they will face many health problems, such as dementia, frailty, Alzheimer's disease, Parkinson's disease, etc. Among them, dementia has become the main cause of disability and dependence on others in the elderly. Similarly, frailty is also a multidimensional syndrome of the elderly, which may include physical, psychological, cognitive damage, and is an important cause of falls, disability, hospitalization and death in the elderly. These health problems faced by the elderly will not only affect the quality of life of individuals, but also bring tremendous pressure to families and society.
[0069] But so far, although there are some hypotheses about the health problems faced by these elderly people, the exact pathogenesis has not been clearly analyzed. Once the elderly develop cognitive impairment and frailty, it will seriously affect the quality of life and bring great pressure to the family. For these problems, the current practice is to detect mild cognitive impairment (MCI) and pre-frailty in the elderly as early as possible, and intervene in time to avoid further dementia, frailty, disability, etc. This is of great significance to the elderly themselves, their families and the entire society.
[0070] At present, the assessment scale is mostly used to detect cognitive impairment and frailty in the elderly. These methods mainly use interviews and record with paper and pen during implementation. These methods have some disadvantages: they need to be operated and executed by medical staff or psychological practitioners, which is time-consuming and economically costly; because the scale questions are fixed or basically fixed, especially in multiple assessments, it will bring the influence of repeated practice, affecting the accuracy of the assessment results; assessment by scale may have answer bias, especially overestimation results; the results of the assessment are represented by the scale, and there are inaccurate representations, etc. Due to the above problems in the current assessment methods, the compliance of elderly assessment users is relatively poor, which affects the smooth progress of the assessment, and then leads to the inability to conduct timely and accurate assessments, abnormal findings and corresponding interventions. More importantly, the existing methods are a summary of the current situation and cannot predict future situations. For cognitive impairment, frailty, etc., it is very important to predict as far as possible.
[0071] Based on this, an embodiment of the present invention provides a method, system and electronic device for predicting health risks for the elderly. The method can use an evaluation decision model to randomly generate various evaluation schemes, and can automatically evaluate the cognitive, frail and other states of elderly users through scale evaluation, scenario evaluation, direct condition evaluation and other methods, and based on this, predict the future risks of cognitive impairment, frailty and the like of the elderly, solving the problems of low evaluation accuracy, high evaluation cost, and inability to predict future risks of traditional methods, and improving user compliance and efficiency.
[0072] To facilitate understanding of this embodiment, a method for predicting health risks of the elderly disclosed in an embodiment of the present invention is first described in detail. The flowchart of the method is as follows: Figure 1 As shown, including:
[0073] Step S101, obtaining the user's health data; wherein the health data at least includes: an image of the user taken by a camera, the user's voice data obtained by a sound collection device, and the user's answer data obtained by answering questions in a questionnaire.
[0074] The user's health data can be obtained through inquiries from relevant personnel, or guided by the user to enter it with the support of relevant hardware devices. These hardware devices are equipped with components such as cameras, microphones, display screens, keyboards, etc., and corresponding robots or devices similar to ATMs can also be used to guide users to enter their health data.
[0075] In specific scenarios, the user's health data can be communicated and exchanged through a reception robot similar to a bank teller machine. The reception robot is equipped with relevant modules such as a camera, a sound collection device, and a display screen. The camera is used to obtain the user's image, and then obtain the user's mood, complexion, mobility and other related health data; the sound collection device can also be used to obtain the user's tone and answer results, and the sound collection device can be used to achieve language communication with the user; the display screen can also be used to provide relevant questionnaires, forms, etc., to guide users to enter their health data in the form of answering questions. These health data include basic information such as the user's name, age, gender, education level, occupation, medical history, etc.; it also includes various real-time situations such as the user's mood, complexion, tone, mobility, etc.; it also includes the purpose of this interaction, etc.
[0076] It is worth mentioning that the appearance of the reception robot needs to give priority to the specific usage of the elderly group. For example, the shape of the reception robot should be as consistent as possible with the aesthetic tastes of the elderly, the speaking speed should be appropriately slowed down during voice interaction, and the font size on the display screen should be appropriately increased.
[0077] Step S102, input the health data into a preset health risk assessment decision model, and determine the user's health risk assessment plan from the output results of the health risk assessment decision model; wherein the health risk assessment plan includes at least: a scale assessment plan, a scenario assessment plan, and a direct condition assessment plan.
[0078] The health risk assessment decision model is constructed by combining rule-based and statistical learning methods. The model constructed by the statistical learning method includes the integration of random forests and neural networks. The input of the health risk assessment decision model is the user's basic situation, immediate situation, interaction purpose and other health data, and the output is the assessment target, assessment dimension, assessment method, assessment implementation and other contents. The main purpose of the health risk assessment decision model is to select appropriate assessment targets, assessment dimensions, assessment methods, assessment implementation and other related contents based on the user's health data, while improving the assessment accuracy and reducing the assessment time.
[0079] The health risk assessment plan for the user is obtained through the health risk assessment decision model. Specifically, the health risk assessment plan includes at least: a scale assessment plan, a scenario assessment plan, and a direct condition assessment plan. Among them, the scale assessment plan evaluates the user's cognitive state and frailty state through questions generated by voice and text; by generating corresponding scale questions and presenting them to the user in the form of voice and text, the user's cognitive state and frailty state are evaluated.
[0080] The scenario assessment solution evaluates the user's cognitive state and debilitation state by completing the scenario interaction questions; by selecting the corresponding scenario and presenting it to the user in combination with the content of the scenario. The user interacts with the scenario to complete the tasks specified in the scenario, thereby evaluating the user's cognitive state and debilitation state.
[0081] The direct condition assessment solution assesses the user's cognitive state and debilitating state through the user's interaction results obtained by sensors; by determining the items of direct condition assessment and their assessment content, the data collection related to the assessment content is completed through relevant sensors, and the user's cognitive state and debilitating state are assessed based on the collected data.
[0082] Step S103, using the health risk assessment scheme to determine the user's health risk assessment data, and after inputting the health risk assessment data into a preset health risk prediction model, determining the user's health risk prediction result from the output result of the health risk prediction model.
[0083] After the health risk assessment plan is determined, the user completes the specific items of the health risk assessment plan to obtain the corresponding health risk assessment data. Since the dimensions and size ranges of these health risk assessment data are not unified, it is also necessary to perform a corresponding data integration process on the health risk assessment data. The health risk assessment data is integrated by standardization, feature calculation, and data fusion to obtain health risk assessment data with unified dimensions.
[0084] After the health risk assessment data is determined, it is input into the preset health risk prediction model, and the health risk assessment data is processed according to the health risk prediction model. During the construction process, the health risk prediction model can use historical data and domain knowledge data to train the model through the corresponding graph neural network and decision tree. The trained health risk prediction model inputs the health risk assessment data, outputs the category and probability value of the user's health risk after relevant processing within the model, and determines the user's health risk prediction result based on the category and probability value of the user's health risk output by the health risk prediction model.
[0085] In some embodiments, the process of inputting health data into a preset health risk assessment decision model, such as Figure 2 As shown, including:
[0086] Step S201, inputting health data into a trained health risk assessment decision model, and the health risk assessment decision model performs statistical decision on the health data based on a built-in assessment method selection template.
[0087] The evaluation method selection template is a hierarchical tree structure, including: evaluation target layer, evaluation dimension layer, evaluation method layer, evaluation implementation layer and evaluation content layer; specifically, the evaluation method selection template is a multi-connected hierarchical tree structure, which is divided into evaluation target layer, evaluation dimension layer, evaluation method layer, evaluation implementation layer and evaluation content layer. It is worth mentioning that the evaluation content layer is used to load the evaluation content corresponding to the health data.
[0088] Step S202, based on the assessment target layer, determining the assessment targets corresponding to the health data; wherein the assessment targets include: cognitive assessment and / or frailty assessment.
[0089] Specifically, the content of the assessment objectives includes: cognitive assessment, frailty assessment, cognitive and / or frailty assessment, etc.
[0090] Step S203, based on the evaluation dimension layer, determine the corresponding evaluation dimension under the evaluation target.
[0091] The content of the evaluation dimension mainly represents the cognitive dimension of the user, specifically, the evaluation of various interactive processes that require cognitive behavior in actual scenarios. Among them, the content of the evaluation dimension includes: attention, transaction processing speed, language interaction, visual spatial positioning, memory status, functional execution, mental arithmetic ability, weight loss, strength, physical fatigue, walking difficulty, physical activity, disease status, nutritional status, anxiety and depression, sleep status, and the evaluation process of one or more of the above dimensions.
[0092] Step S204, based on the evaluation method layer, determine the evaluation target and the corresponding evaluation method under the evaluation dimension; wherein the evaluation method at least includes: scale evaluation, scenario evaluation, and direct situation evaluation.
[0093] Scale assessment, scenario assessment, and direct condition assessment correspond one-to-one to the above three types of schemes included in the health risk assessment scheme. Different assessment strategies will be selected according to different assessment methods in the future.
[0094] Step S205, based on the evaluation implementation layer, determine the evaluation objectives, evaluation dimensions and corresponding evaluation implementation data under the evaluation method; wherein the evaluation implementation data at least includes: scale name, scenario name, direct situation evaluation name.
[0095] In actual scenarios, the scale name can be: MMSE (Mini-mental State Examination) scale, etc.; the scene name can be a phone call test, etc.; the direct situation assessment name can be a repeated language test, etc.
[0096] After the health data is input into the health risk assessment decision model, in some embodiments, the process of determining the user's health risk assessment scheme from the output of the health risk assessment decision model is as follows: Figure 3 As shown, including:
[0097] Step S301, obtain the output result of the health risk assessment decision model; wherein the health risk assessment decision model includes rules and statistical learning models; wherein the statistical learning model integrates random forests and neural networks; the health risk assessment decision model makes decisions on the user's health data, and obtains the assessment objectives, assessment dimensions, assessment methods, and assessment implementation corresponding to the health data from the output result of the health risk assessment decision model.
[0098] By utilizing the user's health data, with the goal of maximizing assessment accuracy and minimizing assessment time, etc., appropriate assessment objectives, assessment dimensions, assessment methods, assessment implementation, etc. are selected as the output results of the health risk assessment decision model.
[0099] In the process of constructing the health risk assessment decision model, a dynamic construction method based on a combination of rules and statistical learning can be adopted: first, adopt the idea of evidence-based literature, and establish rules for patient needs and related assessment goals, assessment dimensions, assessment methods, and assessment implementation contents that have clear and high-quality evidence support. These rules contain domain knowledge provided by assessment experts, and selection decisions are made based on this. Based on the relevant data in the assessment process, statistical learning methods are used to construct relevant models of user health data and assessment goals, assessment dimensions, assessment methods, and assessment implementation contents. It is worth mentioning that the training data comes from the annotations of assessment experts. Then the above rules are optimized and replaced, and the effect of the model needs to meet the rule indicators during replacement. In the above process, the algorithms used include integrated algorithms based on random forests and neural networks.
[0100] Step S302, determining the user's health risk assessment plan based on the assessment objectives, assessment dimensions, assessment methods, and assessment implementation corresponding to the user's health data.
[0101] It is worth mentioning that in the process of selecting evaluation methods and evaluation dimensions, try to choose as many methods and dimensions as possible.
[0102] In some embodiments, if the health risk assessment scheme includes a scale assessment scheme, the process of determining the health risk assessment data of the user using the health risk assessment scheme is as follows: Figure 4 As shown, including:
[0103] Step S401, selecting an assessment scale according to a health risk assessment scheme.
[0104] Specifically, the scale is determined by the specific content of the assessment implementation corresponding to the health risk assessment program.
[0105] Step S402, obtaining the content of the scale questions in the scale evaluation scheme.
[0106] In the process of obtaining the health risk assessment data of the user from the scale assessment scheme, the corresponding scale is first selected according to the health risk assessment scheme to determine the content of the scale questions for use in subsequent steps. In the specific implementation process, the content of the scale questions can also be optimized according to relevant literature.
[0107] Step S403, using the logical order of the contents of the scale questions, determine the order of asking the scale questions.
[0108] In this step, the order of questions is determined according to the rules constructed in advance; for example, questions with no logical order can be randomly determined, and questions with a logical order need to appear in a determined order, etc.
[0109] Step S404, dynamically generate scale questions according to the question-asking sequence, and dynamically display the scale questions to the user.
[0110] Each question is displayed according to the determined question order. In this process, the question generation method is adopted to ensure the diversity of questions. Specifically, the text generation method based on seq2seq can be adopted, but the key parts are limited by rules. It is obtained by extracting specific entities and relationships, and the extracted entities and relationships are expanded by synonym replacement, so as to ensure the diversity of questions and the consistency of semantics.
[0111] Step S405, determining the user's answer results to the scale questions and the process parameters of the answering questions as the user's health risk assessment data.
[0112] In the actual use scenario of the reception robot, for each question displayed, while displaying it on the reception robot display interface, the robot also interacts with the user through voice, and can also interact through the touch screen. Then the process and results of the user's answering the scale questions are collected, including: the identity of the respondent, the length of time the user answers the scale questions, the length of time between answering questions, the emotions when answering questions, the results of answering questions, the number of times the results are modified, etc. Among them, the results of each question are scored and calculated according to the predetermined rules. In this way, the relevant health risk assessment data of the scale assessment is obtained.
[0113] In some embodiments, if the health risk assessment scheme includes a scenario assessment scheme, the process of determining the health risk assessment data of the user using the health risk assessment scheme is as follows: Figure 5 As shown, including:
[0114] Step S501, selecting an assessment scenario according to a health risk assessment scheme.
[0115] Specifically, the assessment scenario is determined by the relevant content such as the assessment implementation corresponding to the health risk assessment plan.
[0116] Step S502, obtaining the scene interaction content in the scene evaluation scheme, and displaying the scene interaction content.
[0117] Scene interaction content may include: preparing food, doing housework, washing clothes, handling finances and other specific scenes. Different scenes will correspond to different interactive content, and these interactive scenes can be displayed through corresponding display devices.
[0118] Step S503: randomly determine scene interaction problems using the scene interaction content, and present the scene interaction problems to the user.
[0119] After the scene interaction content is determined, the corresponding scene interaction questions can be determined based on the interaction content. For example, the food preparation scene can be divided into sub-scenes such as preparing breakfast, lunch, and dinner. In this process, the sub-scenes are determined for the patient in a random manner; in the selected sub-scenes, the specific elements of the scene are determined, such as rice and steamed buns in lunch, and utensils for preparing food, and the corresponding scene interaction questions are randomly generated through these specific elements of the scene.
[0120] Step S504: Determine the user's interaction results for the scene interaction questions and the interaction process parameters as the user's health risk assessment data.
[0121] The selected scene content is displayed on a pre-set screen, and the user interacts with the scene content on the screen through relevant interactive devices to complete the scene-specified tasks. When the user is performing the scene task, the interaction process and its corresponding data will be collected in real time, including the results of the user's completion of the scene task, the time spent on the scene task, the interval time between task links, the number of attempts in each link, the number of errors in each link, and the emotions during the scene task. In this way, relevant evaluation data for the scene evaluation is obtained.
[0122] In some embodiments, if the health risk assessment protocol includes a direct condition assessment protocol, the process of determining the health risk assessment data of the user is as follows: Figure 6 As shown, including:
[0123] Step S601, selecting a direct condition assessment item according to a health risk assessment scheme.
[0124] Specifically, direct condition assessment items are determined by the relevant contents of the assessment implementation corresponding to the health risk assessment plan.
[0125] Step S602, obtaining and determining the content of the situation assessment project in the direct situation assessment solution, and presenting the content of the situation assessment project to the user.
[0126] Direct condition assessment mainly evaluates the user's language, weight, strength, and actions. When the health risk assessment plan is determined as a direct condition assessment plan, the condition assessment items can be determined from the direct condition assessment plan. For example, language assessment can be divided into repetitive sentences and storytelling, and strength assessment can be divided into grip strength and pressure. In this process, the subdivided ability assessment items are determined for the patient in a random manner.
[0127] Step S603: using a preset sensor to obtain sensor data generated when the user interacts under the situation assessment project.
[0128] For the selected condition assessment items, they are displayed on a pre-set screen, and the user's relevant information is collected through voice sensors, visual cameras, tactile sensors, etc., or tested through pressure sensors, visual cameras, voice sensors, etc. carried by the reception robot to complete the capacity assessment.
[0129] Step S604, using the sensor data to determine the user's health risk assessment data.
[0130] During the user's direct situation assessment and after the assessment is completed, process and result information is collected: for example, for strength assessment, mainly record the strength size (maximum, minimum), strength duration, emotional state (happy, indifferent, painful, etc.); for action assessment, mainly record the action distance, speed (maximum, minimum), gait, action time, emotional state (happy, indifferent, painful, etc.); for language assessment, mainly record voice signals (speech speed, interval, fluency, etc.), language information (characters, words, sentences, etc.), semantics (clear, logical), grammar (correct or not), emotion (positive, neutral, negative, etc.), etc. In this way, relevant assessment data for direct situation assessment is obtained.
[0131] The scale assessment data, scenario assessment data, and direct condition assessment data obtained through the above steps need to be integrated to determine the health risk prediction result of the user. In some implementations, after the health risk assessment data is input into a preset health risk prediction model, the health risk prediction result of the user is determined from the output result of the health risk prediction model, such as Figure 7 As shown, including:
[0132] Step S701, after unifying the dimensions and dimensions of the health risk assessment data, calculate the current deviation value and the progress deviation value of the health risk assessment data.
[0133] The process of unifying the dimensions and dimensions of health risk assessment data can be achieved through the steps of standardization, feature calculation, and data fusion. Since the above data have different size ranges and dimensions, the data needs to be standardized for subsequent calculations. The specific methods used may include z-score, maximum and minimum methods, etc. Feature calculation is aimed at overall assessment, cognitive ability assessment, frailty assessment, etc. The Delphi method is used to determine the relative weights between the above assessment dimensions, assessment methods, and assessment implementation, and the relevant intermediate values are calculated based on this, such as language ability assessment scores and scale assessment scores. Data fusion is to splice the above-mentioned health risk assessment data values, calculated intermediate result values, etc. to form assessment input data of unified dimensions.
[0134] The current status deviation value is obtained based on a pre-established data template; the data template is calculated in advance through the above steps for the assessment data of healthy people with normal cognitive and frail status. The specific calculation process includes data standardization, feature calculation, data fusion, data dimension reduction, mean calculation, etc. The data template is represented by the mean vector after dimension reduction. There are many types of data templates, including those established according to different dimensions (overall, cognitive, frailty, etc.), age, gender, and medical history. The calculation process includes: first, the assessment data is subjected to the same dimensionality reduction process as the data template, and then the corresponding template data is selected according to the user characteristics, and then the degree of deviation between the assessment data and the template data is calculated again, which is the current status deviation value.
[0135] The process of obtaining the progress deviation value is aimed at users who regularly make health risk predictions. The comparison benchmark of the progress deviation value is specific historical data. The calculation principle is similar to that of the current status deviation value, which will not be repeated here.
[0136] Step S702, input the health risk assessment data, current deviation value, and progress deviation value into the trained health risk prediction model, and determine the probability of the user developing cognitive impairment and frailty in the future based on the output results of the health risk prediction model; wherein the health risk prediction model is based on a graph neural network and a decision tree, and is trained using historical data and expert input data.
[0137] Specifically, the health risk assessment data includes user characteristic data; if it is a regular risk prediction, it includes previous historical health risk assessment data. After the health risk assessment data, current deviation value, and progress deviation value are input into the corresponding health risk prediction model, the probability value of the user's future cognitive impairment and frailty is obtained according to the prediction results, including: the probability values of cognition and frailty-related MCI (Mild Cognitive Impairment) pre-stage, MCI, dementia, pre-frailty, frailty, disability, etc.
[0138] In the process of constructing the health risk prediction model, the data source can be existing health risk assessment data and historical health risk assessment data, and can also include domain knowledge input by experts. The model algorithm is graph neural network and decision tree, and the model output is the corresponding risk category and its probability value.
[0139] Step S703, determining the user's health risk prediction result according to the probability value of the user developing cognitive impairment and frailty in the future output by the health risk prediction model.
[0140] The health risk prediction results include the types of cognitive impairment and frailty that the user will develop in the future and their probability values. Based on the risk prediction results, the user can take subsequent intervention measures to slow down the process of cognitive impairment and frailty. These intervention measures can also be displayed in the health risk prediction results. Specifically, in some embodiments, the process of determining the health risk prediction results of the user is as follows: Figure 8 As shown, it also includes:
[0141] Step S801, determining the user's intervention items according to the health risk prediction results; wherein the intervention items at least include: scenario intervention, exercise intervention, nutritional intervention, and psychological intervention.
[0142] According to the health risk prediction results, intervention suggestion results are provided to patients who need intervention. The intervention selection suggestion is determined by the intervention selection model, which is specifically implemented through step S802.
[0143] Step S802: input the health risk prediction results and health risk assessment data into the trained intervention selection model, and determine the user's recommended intervention items based on the output results of the intervention selection model; wherein the intervention selection model is based on the non-monotonic inference algorithm of the graph neural network, and is trained using historical data and expert input data.
[0144] The output of the intervention selection model is intervention options, including scenario intervention, exercise intervention, nutritional intervention (the patient's nutrition-related data needs to be obtained during the intervention process), psychological intervention (the patient's psychological condition and other relevant data need to be obtained during the intervention process), etc.
[0145] Step S803: adding the user's recommended intervention items to the user's health risk prediction results.
[0146] It can be seen from the health risk prediction method for the elderly in the above embodiment that the method can use the health risk assessment decision model to randomly generate various assessment schemes, and can automatically evaluate the cognitive, frail and other states of elderly users through scale evaluation, scenario evaluation, direct condition evaluation and other methods, and based on this, predict the future risks of cognitive impairment, frailty and the like of the elderly, which solves the problems of low assessment accuracy, high assessment cost, and inability to predict future risks of traditional methods, and improves user compliance and efficiency.
[0147] Corresponding to the above-mentioned embodiment of the method for predicting health risks of the elderly, the embodiment of the present invention provides a system for predicting health risks of the elderly, such as Fig. 9 As shown, the system includes:
[0148] The health data input module 910 is used to obtain the health data of the user; wherein the health data at least includes: an image of the user taken by a camera, the sound data of the user obtained by a sound collection device, and the answer data of the user obtained by answering questions in a questionnaire;
[0149] The health risk assessment scheme generation module 920 is used to input health data into a preset health risk assessment decision model, and determine the user's health risk assessment scheme from the output results of the health risk assessment decision model; wherein the health risk assessment scheme at least includes: a scale assessment scheme, a scenario assessment scheme, and a direct condition assessment scheme; the scale assessment scheme assesses the user's cognitive state and frailty state through questions generated by voice and text; the scenario assessment scheme assesses the user's cognitive state and frailty state through the completion of scenario interaction questions; the direct condition assessment scheme assesses the user's cognitive state and frailty state through the user's interaction results obtained by sensors;
[0150] The health risk prediction module 930 is used to determine the user's health risk assessment data using a health risk assessment scheme, and after inputting the health risk assessment data into a preset health risk prediction model, determine the user's health risk prediction result from the output result of the health risk prediction model.
[0151] The elderly health risk prediction system provided in the embodiment of the present invention has the same implementation principle and technical effects as the aforementioned elderly health risk prediction method embodiment. For the sake of brief description, for matters not mentioned in the embodiment part, reference can be made to the corresponding contents in the aforementioned method embodiment.
[0152] This embodiment also provides an electronic device, which is a structural diagram of the electronic device as shown in Fig.10As shown, the device includes a processor 101 and a memory 102; wherein the memory 102 is used to store one or more computer instructions, and the one or more computer instructions are executed by the processor to implement the above-mentioned method for predicting health risks for the elderly.
[0153] Fig.10 The electronic device shown further includes a bus 103 and a communication interface 104 , and the processor 101 , the communication interface 104 and the memory 102 are connected via the bus 103 .
[0154] The memory 102 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk storage. The bus 103 may be an ISA bus, a PCI bus, or an EISA bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.10 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.
[0155] The communication interface 104 is used to connect to at least one user terminal and other network units through a network interface, and send the encapsulated IPv4 message or IPv6 message to the user terminal through the network interface.
[0156] The processor 101 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 101. The above processor 101 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, steps and logic block diagrams in the embodiments of the present disclosure can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present disclosure can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module may be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 102, and the processor 101 reads the information in the memory 102 and completes the steps of the method of the above embodiment in combination with its hardware.
[0157] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method of the above embodiment are executed.
[0158] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0159] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0160] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0161] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present invention can essentially or in other words, the part that contributes to the prior art or the 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 enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.
[0162] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.
Claims
1. A method for predicting health risks of the elderly, characterized in that: The method comprises: Acquire the user's health data; wherein the health data at least includes: an image of the user taken by a camera, the user's voice data acquired by a sound acquisition device, and the user's answer data acquired by answering a questionnaire; The health data is input into a preset health risk assessment decision model, and the health risk assessment scheme of the user is determined from the output result of the health risk assessment decision model; wherein the health risk assessment scheme includes at least: a scale assessment scheme, a scenario assessment scheme, and a direct condition assessment scheme; the scale assessment scheme assesses the cognitive state and frailty state of the user through questions generated by voice and text; the scenario assessment scheme assesses the cognitive state and frailty state of the user through the completion of scenario interaction questions; the direct condition assessment scheme assesses the cognitive state and frailty state of the user through the interaction results of the user obtained by sensors; Determining the health risk assessment data of the user using the health risk assessment scheme, and after inputting the health risk assessment data into a preset health risk prediction model, determining the health risk prediction result of the user from the output result of the health risk prediction model; The process of inputting the health data into a preset health risk assessment decision model includes: The health data is input into the trained health risk assessment decision model, and the health risk assessment decision model performs statistical decision on the health data based on a built-in assessment method selection template; wherein the assessment method selection template is a hierarchical tree structure, including: an assessment target layer, an assessment dimension layer, an assessment method layer, an assessment implementation layer, and an assessment content layer; the assessment content layer is used to load the assessment content corresponding to the health data; Based on the assessment target layer, determining the assessment target corresponding to the health data; wherein the assessment target includes: cognitive assessment and / or frailty assessment; Based on the evaluation dimension layer, determine the evaluation dimensions corresponding to the evaluation target; wherein the evaluation dimensions include: attention, transaction processing speed, language interaction, visual spatial positioning, memory status, functional execution, mental arithmetic ability, weight loss, strength, physical fatigue, walking difficulty, physical activity, disease status, nutritional status, anxiety and depression, and sleep status; Based on the evaluation method layer, determining the evaluation target and the corresponding evaluation method under the evaluation dimension; wherein the evaluation method at least includes: scale evaluation, scenario evaluation, and direct situation evaluation; Based on the assessment implementation layer, determining the assessment target, the assessment dimension and the corresponding assessment implementation data under the assessment method; wherein the assessment implementation data at least includes: a scale name, a scenario name, and a direct condition assessment name; The process of determining the health risk assessment scheme of the user from the output result of the health risk assessment decision model includes: Obtaining the output result of the health risk assessment decision model; wherein the health risk assessment decision model includes rules and statistical learning models; wherein the statistical learning model integrates random forests and neural networks; the health risk assessment decision model makes decisions on the health data of the user, and obtains the assessment target, assessment dimension, assessment method, and assessment implementation corresponding to the health data from the output result of the health risk assessment decision model; Determine a health risk assessment plan for the user according to the assessment objectives, assessment dimensions, assessment methods, and assessment implementation corresponding to the health data of the user; After the health risk assessment data is input into a preset health risk prediction model, the health risk prediction result of the user is determined from the output result of the health risk prediction model, including: After unifying the dimensions and dimensions of the health risk assessment data, calculating the current deviation value and the progress deviation value of the health risk assessment data; Inputting the health risk assessment data, the current status deviation value and the progress deviation value into a trained health risk prediction model, and determining the probability value of the user developing cognitive impairment and frailty in the future according to the output result of the health risk prediction model; wherein the health risk prediction model is based on a graph neural network and a decision tree, and is trained using historical data and expert input data; The health risk prediction result of the user is determined according to the probability value of the user developing cognitive impairment and frailty in the future output by the health risk prediction model.
2. The method for predicting health risks of the elderly according to claim 1, characterized in that: If the health risk assessment scheme includes the scale assessment scheme, the process of determining the health risk assessment data of the user using the health risk assessment scheme includes: According to the health risk assessment scheme, select an assessment scale; Obtaining the content of the scale questions in the scale assessment scheme; Determine the order of asking the questions in the scale by using the logical order of the contents of the questions in the scale; Dynamically generate the scale questions according to the question-posing order, and dynamically display the scale questions to the user; The user's answer results to the scale questions and the process parameters of the answer are determined as the user's health risk assessment data.
3. The method for predicting health risks of the elderly according to claim 1, characterized in that: If the health risk assessment scheme includes the scenario assessment scheme, the process of determining the health risk assessment data of the user using the health risk assessment scheme includes: Selecting an assessment scenario according to the health risk assessment scheme; Obtaining scene interaction content in the scene evaluation scheme, and displaying the scene interaction content; Randomly determine scene interaction problems using the scene interaction content, and present the scene interaction problems to the user; The user's interaction results on the scene interaction problem and the interaction process parameters are determined as the user's health risk assessment data.
4. The method for predicting health risks of the elderly according to claim 1, characterized in that: If the health risk assessment scheme includes the direct condition assessment scheme, the process of determining the health risk assessment data of the user includes: Select direct condition assessment items based on the health risk assessment plan; Acquire and determine the content of the condition assessment items in the direct condition assessment scheme, and display the content of the condition assessment items to the user; Using a preset sensor to obtain sensor data generated when the user interacts under the condition assessment project; The sensor data is used to determine health risk assessment data of the user.
5. The method for predicting health risks of the elderly according to claim 1, characterized in that: The process of determining the health risk prediction result of the user also includes: Determine the intervention items for the user according to the health risk prediction result; wherein the intervention items at least include: scene intervention, exercise intervention, nutritional intervention, and psychological intervention; Inputting the health risk prediction result and the health risk assessment data into a trained intervention selection model, and determining the recommended intervention items for the user according to the output result of the intervention selection model; wherein the intervention selection model is based on a non-monotonic inference algorithm of a graph neural network and is trained using historical data and expert input data; The recommended intervention item is added to the health risk prediction result of the user.
6. A health risk prediction system for the elderly, characterized in that: The system comprises: A health data input module, used to obtain the user's health data; wherein the health data at least includes: an image of the user taken by a camera, the user's voice data obtained by a sound collection device, and the user's answer data obtained by answering questions in a questionnaire; A health risk assessment scheme generation module is used to input the health data into a preset health risk assessment decision model, and determine the health risk assessment scheme of the user from the output results of the health risk assessment decision model; wherein the health risk assessment scheme includes at least: a scale assessment scheme, a scenario assessment scheme, and a direct condition assessment scheme; the scale assessment scheme assesses the user's cognitive state and frailty state through questions generated by voice and text; the scenario assessment scheme assesses the user's cognitive state and frailty state through the completion of scenario interaction questions; the direct condition assessment scheme assesses the user's cognitive state and frailty state through the user's interaction results obtained by sensors; A health risk prediction module, which uses the health risk assessment scheme to determine the health risk assessment data of the user, and after inputting the health risk assessment data into a preset health risk prediction model, determines the health risk prediction result of the user from the output result of the health risk prediction model; The health risk assessment scheme generation module, in the process of inputting the health data into the preset health risk assessment decision model, is also used to: input the health data into the health risk assessment decision model that has completed training, and the health risk assessment decision model performs statistical decisions on the health data based on the built-in assessment method selection template; wherein the assessment method selection template is a hierarchical tree structure, including: an assessment target layer, an assessment dimension layer, an assessment method layer, an assessment implementation layer and an assessment content layer; the assessment content layer is used to load the assessment content corresponding to the health data; based on the assessment target layer, determine the assessment target corresponding to the health data; wherein the assessment target includes: cognitive assessment and / or frailty assessment; based on the assessment dimension layer, determine the assessment target The corresponding evaluation dimensions under the evaluation method layer are as follows; wherein the evaluation dimensions include: attention, transaction processing speed, language interaction, visual spatial positioning, memory status, functional execution, mental arithmetic ability, weight loss, strength, physical fatigue, difficulty in walking, physical activity, disease status, nutritional status, anxiety and depression, and sleep status; based on the evaluation method layer, the evaluation target and the corresponding evaluation method under the evaluation dimension are determined; wherein the evaluation method at least includes: scale evaluation, scenario evaluation, and direct situation evaluation; based on the evaluation implementation layer, the evaluation target, the evaluation dimension and the corresponding evaluation implementation data under the evaluation method are determined; wherein the evaluation implementation data at least includes: scale name, scenario name, and direct situation evaluation name; The health risk assessment scheme generation module, in the process of determining the health risk assessment scheme of the user from the output results of the health risk assessment decision model, is also used to: obtain the output results of the health risk assessment decision model; wherein the health risk assessment decision model includes rules and statistical learning models; wherein the statistical learning model integrates random forests and neural networks; the health risk assessment decision model makes decisions on the health data of the user, and obtains the assessment objectives, assessment dimensions, assessment methods, and assessment implementation corresponding to the health data from the output results of the health risk assessment decision model; and determines the health risk assessment scheme of the user according to the assessment objectives, assessment dimensions, assessment methods, and assessment implementation corresponding to the health data of the user; The health risk prediction module, after inputting the health risk assessment data into a preset health risk prediction model, is further used in the process of determining the health risk prediction result of the user from the output result of the health risk prediction model: after unifying the dimensions and dimensions of the health risk assessment data, calculate the current deviation value and the progress deviation value of the health risk assessment data; input the health risk assessment data, the current deviation value and the progress deviation value into the trained health risk prediction model, and determine the probability value of the user developing cognitive impairment and frailty in the future according to the output result of the health risk prediction model; wherein the health risk prediction model is based on a graph neural network and a decision tree, and is trained using historical data and expert input data; and determines the health risk prediction result of the user according to the probability value of the user developing cognitive impairment and frailty in the future output by the health risk prediction model.
7. An electronic device, characterized in that: include: processor and storage device; The storage device stores a computer program, which, when executed by the processor, implements the steps of the method for predicting health risks for the elderly as described in any one of claims 1 to 5.
