Old people ability comprehensive assessment method based on artificial intelligence

Abstract

The invention discloses a comprehensive assessment method for the ability of old people based on artificial intelligence, and relates to the technical field of pension services, and the method integrates community electronic health archives and regional health information platform data, and real-time data of wearable devices and sensors, flexibly generates risk probabilities through nonlinear mapping, and carries out the classification. In the aspect of subjective information and objective data fusion, by means of a natural language processing NLP technology and a knowledge graph, a fusion analysis model is constructed, subjective information such as psychological states, family support and community adaptive capacity is vectorized and fused with a health state feature matrix, the structured expression ability of the subjective information and the comprehensiveness of an evaluation result are greatly enhanced, and the evaluation accuracy of the subjective information is improved. Besides, an immersive scene is constructed through a virtual reality VR platform, the students simulate the action ability of the elderly and the emergency performance of emergencies in a virtual community, behavior parameters are adjusted in real time, and a dynamic simulation evaluation result is generated, so that the evaluation is closer to an actual application scene.

Description

Technical Field

[0001] The present invention relates to the technical field of elderly care services, and in particular to an artificial intelligence-based comprehensive ability assessment method for the elderly. Background Art

[0002] In the comprehensive assessment of the elderly's abilities, common assessment methods include questionnaires, observation records, and standardized testing tools to provide information on the elderly's health status and living abilities.

[0003] However, in the actual assessment process, such assessment methods rely more on the subjective descriptions of the elderly and their families, which can easily lead to inaccurate information due to memory bias or personal understanding differences, and have greater deviations in open scenarios such as communities. At the same time, the design of the assessment tools is too standardized, making it difficult to flexibly adapt to the different living conditions of the elderly in the community. The elderly in the community involve a variety of life patterns such as living alone, family care, and neighborhood interaction, but traditional assessment methods find it difficult to assess the real problems in these scenarios.

[0004] In addition, traditional assessment methods are mostly static and one-time, lacking dynamic monitoring. For the elderly in the community, potential problems will continue to exist and may easily lead to greater hidden dangers. Therefore, how to break through the limitations of traditional assessment methods and achieve a more efficient and actual scenario assessment method has become an urgent problem that needs to be solved. Summary of the invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides an artificial intelligence-based comprehensive ability assessment method for the elderly to solve the problem that traditional assessment methods rely heavily on subjective descriptions of the elderly and their families, which can easily lead to inaccurate information due to memory bias or personal understanding differences, and have large deviations in open scenarios such as communities; at the same time, standardized assessment tools are difficult to truly reflect the different living conditions of the elderly in the community.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] The embodiment of the present invention provides a comprehensive assessment method of the ability of the elderly based on artificial intelligence, which includes:

[0009] Step S1, extracting personal health data of the elderly from the medical health data system, establishing a personal profile, and constructing a personal health status model of the elderly based on the personal profile;

[0010] Step S2, using a questionnaire to supplement subjective information, integrating the subjective information with the state model, using natural language processing (NLP) and knowledge graph to build a fusion analysis model, and the analysis model analyzes the fused data to obtain an evaluation report;

[0011] Step S3, construct an immersive assessment scenario based on a virtual reality VR platform, import the assessment report into the VR platform, and students simulate community scenarios through virtual assessment scenarios to observe the mobility of virtual elderly people in different environments and their emergency response performance in emergencies;

[0012] Step S4, after completing the simulation assessment, an operation feedback report is generated based on the student's operation path and assessment strategy, and a scoring report is generated based on the preset standard answers.

[0013] As a preferred solution of the comprehensive assessment method of the ability of the elderly based on artificial intelligence described in the present invention, wherein: the source of the personal health data includes the personal electronic health records of the community and the personal information in the regional health information platform;

[0014] The data in the personal profile includes medical history, medication use records, and daily activity monitoring data, and the personal profile is regularly updated through wearable devices and sensors.

[0015] As a preferred solution of the method for comprehensive assessment of the ability of the elderly based on artificial intelligence described in the present invention, the step of constructing a personal health status model of the elderly based on personal files is as follows:

[0016] Define the data set and represent the health data of the elderly as a set D = {d1, d2, ..., d n}, where each data point contains a disease history H i , Drug use records i , and daily activity monitoring data A i ,

[0017] The matrix form of the feature is expressed as:

[0018]

[0019] Among them, X represents the feature matrix containing all the health data of the elderly, H i represents the disease history characteristics of the i-th individual, M i represents the drug use record characteristics of the ith individual, A i represents the daily activity monitoring data characteristics of the i-th individual;

[0020] The health data is standardized, and the feature matrix X is standardized to X′. The processing formula is:

[0021]

[0022] Among them, X′ is the standardized feature matrix, μ is the feature mean vector, and σ is the feature standard deviation vector;

[0023] The health status model is constructed by weighted summation of eigenvalues. The model formula is:

[0024] S i =αH i +βM i +γA i ,

[0025] Among them, S i represents the health status score of the ith individual, α, β, and γ represent the weight parameters of disease history, medication record, and daily activity characteristics, respectively;

[0026] The health score is mapped to the risk probability through a nonlinear function. The mapping formula is:

[0027]

[0028] Among them, R i represents the health risk probability of the ith individual, S i represents the corresponding health status score, and e represents the base of the natural logarithm;

[0029] Classify health status and set risk threshold T based on R i Classify health status as normal or at risk by:

[0030] If R i >T, then C i =1, otherwise C i =0,

[0031] Among them, C i represents the health classification result of the i-th individual, 1 represents high risk, 0 represents normal, T represents the risk threshold, R i represents the health risk probability of the ith individual.

[0032] As a preferred solution of the artificial intelligence-based comprehensive ability assessment method for the elderly described in the present invention, the subjective information includes the elderly's psychological state, family support status and community adaptability.

[0033] As a preferred solution of the comprehensive assessment method of the elderly's ability based on artificial intelligence described in the present invention, wherein: the fusion analysis model is constructed by using natural language processing NLP and knowledge graph, and the steps of analyzing the fused data by the analysis model are as follows:

[0034] Vectorize the subjective information and define it as Q = {q1, q2, …, m}. The features include the psychological state P i , Family support status i 、Community AdaptabilityC i , the vectorized representation is:

[0035]

[0036] Among them, Y represents the subjective information feature matrix, P i represents the psychological state characteristics of the i-th individual, F i represents the family support characteristics of the ith individual, C i represents the community adaptability characteristics of the i-th individual,

[0037] The health status feature matrix X′ and the subjective information Y are combined into a comprehensive matrix Z:

[0038] Z=[X′|Y],

[0039] Among them, Z represents the integrated feature matrix after fusion, X′ represents the standardized health status feature matrix, Y represents the subjective information feature matrix, and | represents the horizontal concatenation operation of the matrix;

[0040] Embed natural language data to transform the subjective information T of the questionnaire text into a semantic vector E T , expressed using the embedded function, the formula is:

[0041] E T =Embed(T),

[0042] Among them, E T represents the semantic vector of subjective text data, T represents the questionnaire text data, and Embed(·) represents the word embedding function;

[0043] Using the semantic vector E T Match the knowledge graph nodes and construct the relationship matrix G:

[0044] G={g1,g2,…,g k},

[0045] Among them, G represents the knowledge graph node set, g i represents the i-th node in the knowledge graph, E T Semantic vectors representing subjective textual data;

[0046] Comprehensive health status feature Z, semantic vector E T And the knowledge graph node G, build a multimodal parsing model, the parsing model is defined as:

[0047] O=f(Z,G,E T ),

[0048] Among them, O represents the output of the fusion analysis model, including the evaluation results, f represents the multimodal fusion function, specifically a neural network model based on deep learning, Z represents the comprehensive feature matrix, G represents the knowledge graph node set, E represents the multimodal fusion function, and G represents the multimodal fusion function.T Semantic vectors representing subjective textual data;

[0049] Parse the output of the fusion model O into a set of evaluation indicators R eval :

[0050] R eval =Parse(O),

[0051] Among them, R eval represents the set of evaluation indicators after parsing, Parse(·) represents the specific method of the parsing operation, the evaluation indicators are extracted through the activation layer output by the model, and O represents the output result of the fusion parsing model;

[0052] Using R eval The extracted indicators generate a comprehensive ability assessment report for the elderly:

[0053] R report =Summarize(R eval ),

[0054] Among them, R report represents the evaluation report, Summarize(·) represents the process of generating a structured report based on the evaluation indicator set, R eval Represents the parsed set of evaluation metrics.

[0055] As a preferred solution of the comprehensive assessment method of the ability of the elderly based on artificial intelligence described in the present invention, wherein: the mobility includes gait stability and independent activity ability;

[0056] During the simulation phase of the community scenario, students can dynamically adjust the behavioral parameters and health status of virtual individuals to generate real-time simulation evaluation results.

[0057] As a preferred solution of the comprehensive assessment method of the elderly's ability based on artificial intelligence described in the present invention, the step of constructing an immersive assessment scene based on a virtual reality VR platform is as follows:

[0058] Initialize the scene, build the physical environment E and virtual character parameters V of the VR evaluation scene, the environmental elements include buildings, roads and community interaction points, and the character parameters include the action ability index A v and health status index H v According to the evaluation report R report , the health data is converted into virtual character parameters through the mapping function, and the conversion method is:

[0059] V = g(R report ),

[0060] Where V represents the set of behavior parameters of the virtual character, g(·) represents the mapping function from the evaluation report to the virtual character parameters, and R report Indicates an assessment report.

[0061] As a preferred solution of the method for comprehensive assessment of the abilities of the elderly based on artificial intelligence described in the present invention, the step of adjusting the behavior parameters and health status of the virtual individual to generate a real-time simulation assessment result is as follows:

[0062] Students adjust the character parameters V through the interactive interface for dynamic adjustment and real-time simulation, and generate new simulation results S in real time. sim , the simulation process is expressed as;

[0063] S sim =h(E,V),

[0064] Among them, S sim represents the real-time simulation result, h(·) represents the simulation generation function, which combines the environment E and the character parameters V. E represents the physical environment of the VR scene, and V represents the set of virtual character parameters.

[0065] As a preferred solution of the comprehensive assessment method of the elderly ability based on artificial intelligence described in the present invention, in step S4, the operation feedback report is generated in the following manner:

[0066] A rule-based AI expert system and reinforcement learning model are used to analyze students’ operation paths. The expert system presets operation specifications, evaluates the logic of students’ operation paths, and uses a reinforcement learning model to identify whether students deviate from standard operation procedures.

[0067] Use decision tree models to analyze students’ nodes and strategy deviations in the evaluation process and identify whether some necessary links are missed;

[0068] The operation results are compared with the standard answers, and the students' performance is scored using deviation detection technology.

[0069] As a preferred solution of the comprehensive assessment method of the elderly's ability based on artificial intelligence described in the present invention, the steps of generating an operation feedback report based on the student's operation path and assessment strategy and generating a scoring report based on a preset standard answer are as follows:

[0070] For the logical analysis of the operation path, let the student's operation path be P = {p1, p2, ..., p k}, the rule engine analyzes the path logic and obtains the logic score L. The analysis process is expressed as:

[0071] L=LogicCheck(P),

[0072] Where L represents the logic score of the operation path, LogicCheck(·) represents the logic evaluation function based on the rule engine, and P represents the student's operation path set.

[0073] The reinforcement learning model is used to evaluate the deviation D of the operation path. The evaluation formula is:

[0074] D=|P-π|,

[0075] Where D represents the deviation value of the student's operation path, P represents the set of students' operation paths, π represents the standard operation path, and |·| represents the path difference measurement function;

[0076] The decision tree model T is used to analyze the nodes and strategy deviations in the operation path. The analysis formula is:

[0077] Δ=T(P),

[0078] Among them, Δ represents the path parsing result, including strategy deviation and node status, T represents the node parsing model based on decision tree, P represents the student's operation path set,

[0079] Compare the path analysis result with the standard answer A to generate the scoring result S score :

[0080] S score =Evaluate(Δ,A),

[0081] Among them, S score represents the final scoring result, Evaluate(·) represents the scoring function, Δ represents the path resolution result, and A represents the standard answer.

[0082] The beneficial effects of the present invention are as follows: the present invention integrates community electronic health records and regional health information platform data, as well as real-time data from wearable devices and sensors, comprehensively reflects the personalized changes in the health status of the elderly, and flexibly generates and classifies risk probabilities through nonlinear mapping; in the fusion of subjective information and objective data, with the help of natural language processing (NLP) technology and knowledge graphs, a fusion analysis model is constructed to vectorize subjective information such as psychological state, family support and community adaptability, and integrate it with the health status feature matrix. The parsed semantic vector establishes semantic associations between data by matching with knowledge graph nodes, greatly enhancing the structured expression ability of subjective information and the comprehensiveness of evaluation results.

[0083] The present invention constructs an immersive scene through a virtual reality VR platform. Students simulate the mobility and emergency response performance of the elderly in a virtual community, adjust behavioral parameters in real time and generate dynamic simulation evaluation results, so that the evaluation is closer to the actual application scenario; and combines rule-based AI expert systems, reinforcement learning models and decision tree technology to perform logical analysis and deviation detection on students' operation paths and strategies, generate operation feedback reports and scoring reports, and comprehensively improve the scientificity and intelligence level of the evaluation process, providing a more efficient way for comprehensive ability evaluation of the elderly and student learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0085] Figure 1 This is a flow chart of a comprehensive assessment method for the abilities of the elderly based on artificial intelligence in Example 1. DETAILED DESCRIPTION

[0086] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0087] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0088] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0089] Example 1, reference Figure 1 , this embodiment provides a comprehensive assessment method for the abilities of the elderly based on artificial intelligence, comprising the following steps:

[0090] Step S1, extracting personal health data of the elderly from the medical health data system, establishing a personal profile, and constructing a personal health status model of the elderly based on the personal profile;

[0091] Sources of personal health data include personal electronic health records in the community and personal information in the regional health information platform;

[0092] The data in the personal profile includes medical history, medication use records, and daily activity monitoring data, and the personal profile is regularly updated through wearable devices and sensors;

[0093] The steps to construct a personal health status model for the elderly based on personal profiles are:

[0094] Define the data set and represent the health data of the elderly as a set D = {d1, d2, ..., d n}, where each data point contains a disease history H i , Drug use records i , and daily activity monitoring data A i ,

[0095] The matrix form of the feature is expressed as:

[0096]

[0097] Among them, X represents the feature matrix containing all the health data of the elderly, H i represents the disease history characteristics of the i-th individual, M i represents the drug use record characteristics of the ith individual, A i represents the daily activity monitoring data characteristics of the i-th individual;

[0098] The health data is standardized, and the feature matrix X is standardized to X′. The processing formula is:

[0099]

[0100] Among them, X′ is the standardized feature matrix, μ is the feature mean vector, and σ is the feature standard deviation vector;

[0101] The health status model is constructed by weighted summation of eigenvalues. The model formula is:

[0102] S i =αH i +βM i +γA i ,

[0103] Among them, S i represents the health status score of the ith individual, α, β, and γ represent the weight parameters of disease history, medication record, and daily activity characteristics, respectively;

[0104] The health score is mapped to the risk probability through a nonlinear function. The mapping formula is:

[0105]

[0106] Among them, R i represents the health risk probability of the ith individual, S i represents the corresponding health status score, and e represents the base of the natural logarithm;

[0107] Classify health status and set risk threshold T based on R i Classify health status as normal or at risk by:

[0108] If R i >T, then C i =1, otherwise C i =0,

[0109] Among them, C i represents the health classification result of the i-th individual, 1 represents high risk, 0 represents normal, T represents the risk threshold, R i represents the health risk probability of the i-th individual;

[0110] Specifically, through the above steps, the health status model is used to start from structured data, and after feature extraction, standardization processing and weighted calculation, the health status of the elderly is quantified, and risk classification is achieved through nonlinear mapping, which can flexibly adapt to the health assessment needs of different elderly groups.

[0111] Step S2, using a questionnaire to supplement subjective information, integrating the subjective information with the state model, using natural language processing (NLP) and knowledge graph to build a fusion analysis model, and the analysis model analyzes the fused data to obtain an evaluation report;

[0112] Subjective information includes the older adult’s psychological state, family support status, and community adaptation ability;

[0113] The fusion parsing model is constructed by using natural language processing (NLP) and knowledge graph. The steps of parsing the fused data by the parsing model are as follows:

[0114] Vectorize the subjective information and define it as Q = {q1,q2,…,q m}, features include psychological state P i , Family support status i 、Community Adaptability C i , the vectorized representation is:

[0115]

[0116] Among them, Y represents the subjective information feature matrix, P i represents the psychological state characteristics of the i-th individual, F irepresents the family support characteristics of the ith individual, C i represents the community adaptability characteristics of the i-th individual,

[0117] The health status feature matrix X′ and the subjective information Y are combined into a comprehensive matrix Z:

[0118] Z=[X′|Y],

[0119] Among them, Z represents the integrated feature matrix after fusion, X′ represents the standardized health status feature matrix, Y represents the subjective information feature matrix, and | represents the horizontal concatenation operation of the matrix;

[0120] Embed natural language data to transform the subjective information T of the questionnaire text into a semantic vector E T , expressed using the embedded function, the formula is:

[0121] E T =Embed(T),

[0122] Among them, E T represents the semantic vector of subjective text data, T represents the questionnaire text data, and Embed(·) represents the word embedding function;

[0123] Using the semantic vector E T Match the knowledge graph nodes and construct the relationship matrix G:

[0124] G={g1,g2,…,g k},

[0125] Among them, G represents the knowledge graph node set, g i represents the i-th node in the knowledge graph, E T Semantic vectors representing subjective textual data;

[0126] Comprehensive health status feature Z, semantic vector E T And the knowledge graph node G, build a multimodal parsing model, the parsing model is defined as:

[0127] O=f(Z,G,E T ),

[0128] Among them, O represents the output of the fusion analysis model, including the evaluation results, f represents the multimodal fusion function, specifically a neural network model based on deep learning, Z represents the comprehensive feature matrix, G represents the knowledge graph node set, E represents the multimodal fusion function, and G represents the multimodal fusion function. T Semantic vectors representing subjective textual data;

[0129] Parse the output of the fusion model O into a set of evaluation indicators R eval :

[0130] Reval =Parse(O),

[0131] Among them, R eval represents the set of evaluation indicators after parsing, Parse(·) represents the specific method of the parsing operation, the evaluation indicators are extracted through the activation layer output by the model, and O represents the output result of the fusion parsing model;

[0132] Using R eval The extracted indicators generate a comprehensive ability assessment report for the elderly:

[0133] R report =Summarize(R eval ),

[0134] Among them, R report represents the evaluation report, Summarize(·) represents the process of generating a structured report based on the evaluation indicator set, R eval Represents the parsed set of evaluation indicators;

[0135] Specifically, through multimodal fusion analysis of subjective information and state models, combined with NLP technology and knowledge graphs, intelligent conversion from multidimensional data to evaluation results is achieved. Knowledge graphs provide semantic association capabilities, and embedding technology further enhances the structured expression capabilities of subjective data, ultimately forming an evaluation report.

[0136] Step S3, constructing an immersive assessment scenario based on a virtual reality VR platform, importing the assessment report into the VR platform, and allowing students to simulate community scenarios through virtual assessment scenarios, observing the mobility of virtual elderly people in different environments and their emergency response performance in emergencies;

[0137] Mobility including gait stability and independent movement;

[0138] During the simulation phase of the community scenario, students can dynamically adjust the behavior parameters and health status of virtual individuals to generate real-time simulation evaluation results;

[0139] The steps to build an immersive evaluation scenario based on the virtual reality VR platform are:

[0140] Initialize the scene, build the physical environment E and virtual character parameters V of the VR evaluation scene, the environmental elements include buildings, roads and community interaction points, and the character parameters include the action ability index A v and health status index H v According to the evaluation report R report , the health data is converted into virtual character parameters through the mapping function, and the conversion method is:

[0141] V = g(R report ),

[0142] Where V represents the set of behavior parameters of the virtual character, g(·) represents the mapping function from the evaluation report to the virtual character parameters, and R report It means evaluation report;

[0143] The steps to adjust the behavior parameters and health status of the virtual individual and generate real-time simulation evaluation results are:

[0144] Students adjust the character parameters V through the interactive interface for dynamic adjustment and real-time simulation, and generate new simulation results S in real time. sim , the simulation process is expressed as;

[0145] S sim =h(E,V),

[0146] Among them, S sim represents the real-time simulation result, h(·) represents the simulation generation function, which combines the environment E and the character parameters V, where E represents the physical environment of the VR scene and V represents the set of virtual character parameters;

[0147] Specifically, through the immersive scenes and dynamic adjustment functions of the VR platform, students can observe the performance of virtual individuals during interaction and observe the impact of parameter changes on simulation results.

[0148] Step S4, after completing the simulation assessment, generating an operation feedback report based on the student's operation path and assessment strategy, and generating a scoring report based on the preset standard answers;

[0149] In step S4, the operation feedback report is generated by:

[0150] A rule-based AI expert system and reinforcement learning model are used to analyze students’ operation paths. The expert system presets operation specifications, evaluates the logic of students’ operation paths, and uses a reinforcement learning model to identify whether students deviate from standard operation procedures.

[0151] Use decision tree models to analyze students’ nodes and strategy deviations in the evaluation process and identify whether some necessary links are missed;

[0152] The operation results are compared with the standard answers, and the students' performance is scored using deviation detection technology;

[0153] The steps for generating an operation feedback report based on the student's operation path and evaluation strategy and generating a scoring report based on the preset standard answers are as follows:

[0154] For the logical analysis of the operation path, let the student's operation path be P = {p1, p2, ..., p k}, the rule engine analyzes the path logic and obtains the logic score L. The analysis process is expressed as:

[0155] L=LogicCheck(P),

[0156] Where L represents the logic score of the operation path, LogicCheck(·) represents the logic evaluation function based on the rule engine, and P represents the student's operation path set.

[0157] The reinforcement learning model is used to evaluate the deviation D of the operation path. The evaluation formula is:

[0158] D=|P-π|,

[0159] Where D represents the deviation value of the student's operation path, P represents the set of students' operation paths, π represents the standard operation path, and |·| represents the path difference measurement function;

[0160] The decision tree model T is used to analyze the nodes and strategy deviations in the operation path. The analysis formula is:

[0161] Δ=T(P),

[0162] Among them, Δ represents the path parsing result, including strategy deviation and node status, T represents the node parsing model based on decision tree, P represents the student's operation path set,

[0163] Compare the path analysis result with the standard answer A to generate the scoring result S score :

[0164] S score =Evaluate(Δ,A),

[0165] Among them, S score represents the final scoring result, Evaluate(·) represents the scoring function, Δ represents the path resolution result, and A represents the standard answer;

[0166] Specifically, multiple analyses conducted through rule engines, reinforcement learning, and decision trees to generate operation feedback reports and scoring reports can effectively reflect students' operation paths and strategy deviations; automated analysis in the scoring process will help to comprehensively improve the fairness and accuracy of the assessment.

[0167] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A comprehensive assessment method for the abilities of the elderly based on artificial intelligence, characterized by: include, Step S1, extracting personal health data of the elderly from the medical health data system, establishing a personal profile, and constructing a personal health status model of the elderly based on the personal profile; Step S2, using a questionnaire to supplement subjective information, integrating the subjective information with the state model, using natural language processing (NLP) and knowledge graph to build a fusion analysis model, and the analysis model analyzes the fused data to obtain an evaluation report; Step S3, construct an immersive assessment scenario based on a virtual reality VR platform, import the assessment report into the VR platform, and students simulate community scenarios through virtual assessment scenarios to observe the mobility of virtual elderly people in different environments and their emergency response performance in emergencies; Step S4, after completing the simulation assessment, an operation feedback report is generated based on the student's operation path and assessment strategy, and a scoring report is generated based on the preset standard answers.

2. The method for comprehensive assessment of the abilities of the elderly based on artificial intelligence as claimed in claim 1, characterized in that: The sources of personal health data include personal electronic health records in the community and personal information in the regional health information platform; The data in the personal profile includes medical history, medication use records, and daily activity monitoring data, and the personal profile is regularly updated through wearable devices and sensors.

3. The method for comprehensive assessment of the abilities of the elderly based on artificial intelligence as claimed in claim 2, characterized in that: The steps of constructing the personal health status model of the elderly based on the personal profile are: Define the data set and represent the health data of the elderly as a set D = {d1, d2, ..., d n }, where each data point contains a disease history H i , Drug use records i , and daily activity monitoring data A i , The matrix form of the feature is expressed as: Among them, X represents the feature matrix containing all the health data of the elderly, H i represents the disease history characteristics of the i-th individual, M i represents the drug use record characteristics of the ith individual, A i represents the daily activity monitoring data characteristics of the i-th individual; The health data is standardized, and the feature matrix X is standardized to X′. The processing formula is: Among them, X′ is the standardized feature matrix, μ is the feature mean vector, and σ is the feature standard deviation vector; The health status model is constructed by weighted summation of eigenvalues. The model formula is: S i =αH i +βM i +γA i , Among them, S i represents the health status score of the ith individual, α, β, and γ represent the weight parameters of disease history, medication record, and daily activity characteristics, respectively; The health score is mapped to the risk probability through a nonlinear function. The mapping formula is: Among them, R i represents the health risk probability of the ith individual, S i represents the corresponding health status score, and e represents the base of the natural logarithm; Classify health status and set risk threshold T based on R i Classify health status as normal or at risk by: If R i >T, then C i =1, otherwise C i =0, Among them, C i represents the health classification result of the i-th individual, 1 represents high risk, 0 represents normal, T represents the risk threshold, R i represents the health risk probability of the ith individual.

4. The method for comprehensive assessment of the abilities of the elderly based on artificial intelligence as claimed in claim 3, characterized in that: The subjective information includes the elderly's mental state, family support status and community adaptability.

5. The method for comprehensive assessment of the abilities of the elderly based on artificial intelligence as claimed in claim 4, characterized in that: The method uses natural language processing (NLP) and knowledge graph to build a fusion parsing model, and the steps of parsing the fused data by the parsing model are as follows: Vectorize the subjective information and define it as Q = {q1,q2,…,q m }, features include psychological state P i , Family support status i 、Community AdaptabilityC i , the vectorized representation is: Among them, Y represents the subjective information feature matrix, P i represents the psychological state characteristics of the i-th individual, F i represents the family support characteristics of the ith individual, C i represents the community adaptability characteristics of the i-th individual, The health status feature matrix X′ and the subjective information Y are combined into a comprehensive matrix Z: Z=[X′|Y], Among them, Z represents the integrated feature matrix after fusion, X′ represents the standardized health status feature matrix, Y represents the subjective information feature matrix, and | represents the horizontal concatenation operation of the matrix; Embed natural language data to transform the subjective information T of the questionnaire text into a semantic vector E T , expressed using the embedded function, the formula is: E T =Embed(T), Among them, E T represents the semantic vector of subjective text data, T represents the questionnaire text data, and Embed(·) represents the word embedding function; Using the semantic vector E T Match the knowledge graph nodes and construct the relationship matrix G: G={g1,g2,…,g k }, Among them, G represents the knowledge graph node set, g i represents the i-th node in the knowledge graph, E T Semantic vectors representing subjective textual data; Comprehensive health status feature Z, semantic vector E T And the knowledge graph node G, build a multimodal parsing model, the parsing model is defined as: O=f(Z,G,E T ), Among them, O represents the output of the fusion analysis model, including the evaluation results, f represents the multimodal fusion function, specifically a neural network model based on deep learning, Z represents the comprehensive feature matrix, G represents the knowledge graph node set, E represents the multimodal fusion function, and G represents the multimodal fusion function. T Semantic vectors representing subjective textual data; Parse the output of the fusion model O into a set of evaluation indicators R eval : R eval Parse(O) Among them, R eval represents the set of evaluation indicators after parsing, Parse(·) represents the specific method of the parsing operation, the evaluation indicators are extracted through the activation layer output by the model, and O represents the output result of the fusion parsing model; Using R eval The extracted indicators generate a comprehensive ability assessment report for the elderly: R report =Summarize(R eval ), Among them, R report represents the evaluation report, Summarize(·) represents the process of generating a structured report based on the evaluation indicator set, R eval Represents the parsed set of evaluation metrics.

6. The method for comprehensive assessment of the abilities of the elderly based on artificial intelligence as claimed in claim 5, characterized in that: Said mobility includes gait stability and independent movement ability; During the simulation phase of the community scenario, students can dynamically adjust the behavioral parameters and health status of virtual individuals to generate real-time simulation evaluation results.

7. The method for comprehensive assessment of the abilities of the elderly based on artificial intelligence as claimed in claim 6, characterized in that: The steps of constructing an immersive evaluation scene based on a virtual reality VR platform are: Initialize the scene, build the physical environment E and virtual character parameters V of the VR evaluation scene, the environmental elements include buildings, roads and community interaction points, and the character parameters include the action ability index A v and health status index H v According to the evaluation report R report , the health data is converted into virtual character parameters through the mapping function, and the conversion method is: V=g(R report ), Where V represents the set of behavior parameters of the virtual character, g(·) represents the mapping function from the evaluation report to the virtual character parameters, and R report Indicates an assessment report.

8. The method for comprehensive assessment of the abilities of the elderly based on artificial intelligence as claimed in claim 7, characterized in that: The steps of adjusting the behavior parameters and health status of the virtual individual to generate real-time simulation evaluation results are: Students adjust the character parameters V through the interactive interface for dynamic adjustment and real-time simulation, and generate new simulation results S in real time. sim , the simulation process is expressed as; S sim =h(E,V), Among them, S sim represents the real-time simulation result, h(·) represents the simulation generation function, which combines the environment E and the character parameters V. E represents the physical environment of the VR scene, and V represents the set of virtual character parameters.

9. The method for comprehensive assessment of the abilities of the elderly based on artificial intelligence as claimed in claim 8, characterized in that: In step S4, the operation feedback report is generated by: A rule-based AI expert system and reinforcement learning model are used to analyze students’ operation paths. The expert system presets operation specifications, evaluates the logic of students’ operation paths, and uses a reinforcement learning model to identify whether students deviate from standard operation procedures. Use decision tree models to analyze students’ nodes and strategy deviations in the evaluation process and identify whether some necessary links are missed; The operation results are compared with the standard answers, and the students' performance is scored using deviation detection technology.

10. The method for comprehensive assessment of the abilities of the elderly based on artificial intelligence as claimed in claim 9, characterized in that: The steps of generating an operation feedback report based on the student's operation path and evaluation strategy and generating a scoring report based on a preset standard answer are: For the logical analysis of the operation path, let the student's operation path be P = {p1, p2, ..., p k }, the rule engine analyzes the path logic and obtains the logic score L. The analysis process is expressed as: L=LogicCheck(P), Where L represents the logic score of the operation path, LogicCheck(·) represents the logic evaluation function based on the rule engine, and P represents the student's operation path set. The reinforcement learning model is used to evaluate the deviation D of the operation path. The evaluation formula is: D=|P-π|, Where D represents the deviation value of the student's operation path, P represents the set of students' operation paths, π represents the standard operation path, and |·| represents the path difference measurement function; The decision tree model T is used to analyze the nodes and strategy deviations in the operation path. The analysis formula is: Δ=T(P), Among them, Δ represents the path parsing result, including strategy deviation and node status, T represents the node parsing model based on decision tree, P represents the student's operation path set, Compare the path analysis result with the standard answer A to generate the scoring result S score : S score =Evaluate(Δ,A), Among them, S score represents the final scoring result, Evaluate(·) represents the scoring function, Δ represents the path resolution result, and A represents the standard answer.

Images