Old-age nursing management optimization system based on artificial intelligence

Abstract

The invention discloses an elderly nursing management optimization system based on artificial intelligence. The elderly nursing management optimization system comprises a nursing data management module, a nursing plan enhancement module, a behavior health guidance module, a psychological health nursing module and an elderly nursing optimization module. The invention belongs to the technical field of old-age nursing optimization, and particularly relates to an artificial intelligence-based old-age nursing management optimization system, which adopts a reinforcement learning method combined with a multi-task deep neural network to enhance nursing plans, and performs time sequence modeling through innovative shared feature learning and emotional state. Efficient information sharing among a plurality of tasks (health assessment, chronic disease management and the like) and mutual optimization among the tasks are realized; performing behavior health guidance by adopting a bidirectional long-short-term memory neural network combining time sequence convolution and an improved support vector classifier; and generating an optimization network by adopting an emotion dynamic modeling strategy, and carrying out psychological health care.

Description

Technical Field

[0001] The present invention belongs to the technical field of elderly care optimization, and specifically relates to an elderly care management optimization system based on artificial intelligence. Background Art

[0002] The Elderly Care Management Optimization System is an intelligent system that integrates advanced artificial intelligence technologies (such as machine learning, deep learning, data analysis, and natural language processing) to improve the efficiency and quality of elderly care services. By conducting real-time monitoring and intelligent analysis of multi-dimensional data such as elderly health status, care needs, and lifestyle habits, the system automatically optimizes care resource allocation, personalized health management, and service scheduling, aiming to provide more precise, personalized, and efficient care services for the elderly.

[0003] However, in existing nursing planning enhancement methods, there is a technical problem that multi-task learning cannot efficiently share the characteristics of each task, resulting in insufficient information transmission between multiple tasks. At the same time, the nursing plan lacks the ability to make real-time adjustments to individuals; in existing behavioral health guidance methods, there is a technical problem that the impact of the personalized behavioral habits and emotional changes of the elderly cannot be fully considered; in existing mental health care methods, there is a lack of technical means to monitor and evaluate the emotional state of the elderly in real time. At the same time, the counseling strategies lack personalization and cannot be dynamically adjusted according to the emotional changes of the elderly. Summary of the Invention

[0004] In response to the above technical problems, the present invention provides an artificial intelligence-based elderly care management optimization system. In view of the technical problems in existing nursing planning enhancement methods, multi-task learning cannot efficiently share the characteristics of each task, resulting in insufficient information transmission between multiple tasks. At the same time, the nursing plan lacks the ability to adjust in real time for individuals. This solution creatively adopts a reinforcement learning method combined with a multi-task deep neural network to enhance nursing planning. Through innovative shared feature learning and time series modeling of emotional states, efficient information sharing and mutual optimization between multiple tasks (health assessment, chronic disease management, etc.) are achieved, and a dynamically weighted multi-objective reward function is adopted to achieve optimization of long-term nursing effects. In view of the fact that existing behavioral health guidance methods cannot fully take into account the personalized behavioral habits and emotions of the elderly, this solution creatively adopts a reinforcement learning method combined with a multi-task deep neural network to enhance nursing planning. Through innovative shared feature learning and time series modeling of emotional states, efficient information sharing and mutual optimization between tasks are achieved, and a dynamically weighted multi-objective reward function is adopted to achieve optimization of long-term nursing effects. In order to solve the technical problem of the impact of changes in the elderly's mood, this solution creatively adopts a bidirectional long short-term memory neural network combined with temporal convolution and an improved support vector classifier to provide behavioral health guidance. By dynamically capturing the behavioral patterns and emotional fluctuations of the elderly, the health guidance strategy is adjusted in real time to achieve personalized behavioral guidance, so that the health guidance strategy can more accurately respond to the specific needs of the elderly. In order to solve the technical problem that the existing mental health care methods lack technical means to monitor and evaluate the emotional state of the elderly in real time, and the guidance strategy lacks personalization and cannot be dynamically adjusted according to the emotional changes of the elderly, this solution creatively adopts the strategy of emotional dynamic modeling to generate an optimized network for mental health care. It automatically generates and adjusts the guidance strategy according to the emotional dynamics and the specific needs of the elderly, and realizes personalized and continuously effective mental health care.

[0005] The technical solution adopted by the present invention is as follows: the present invention provides an artificial intelligence-based elderly care management optimization system, which includes a nursing data management module, a nursing planning enhancement module, a behavioral health guidance module, a mental health care module and an elderly care optimization module;

[0006] The nursing data management module is used for nursing data management, obtains elderly care management optimization data through nursing data management, and sends the elderly care management optimization data to the nursing planning enhancement module, the behavioral health guidance module and the mental health care module;

[0007] The nursing plan enhancement module is used for nursing plan enhancement, obtains elderly nursing plan reference data through nursing plan enhancement, and sends the elderly nursing plan reference data to the behavioral health guidance module and the mental health nursing module;

[0008] The behavioral health guidance module is used for behavioral health guidance, obtains daily behavioral health guidance reference data for the elderly through behavioral health guidance, and sends the daily behavioral health guidance reference data for the elderly to the elderly care optimization module;

[0009] The mental health care module is used for mental health care, obtains reference data on mental health counseling for the elderly through mental health care, and sends the reference data on mental health counseling for the elderly to the elderly care optimization module;

[0010] The elderly care optimization module is used for elderly care optimization, and obtains a comprehensive elderly care reference data set through elderly care optimization.

[0011] Furthermore, the nursing data management is used to collect, store, organize and optimize the health data of the elderly. Specifically, the nursing data management is performed from the elderly medical care system and the elderly health survey records through nursing data collection, data optimization and storage management to obtain elderly nursing management optimization data, and the elderly nursing management optimization data is stored in the database for data encryption storage;

[0012] The nursing data collection specifically collects the health data, behavior data and nursing history data of the elderly to obtain the original data set of elderly care management;

[0013] The original data set of elderly care management specifically includes physiological health data, behavioral data, nursing record data, environmental auxiliary data and mental health data;

[0014] The physiological health data includes blood pressure, blood sugar, heart rate, body temperature and blood oxygen saturation data of the elderly, which are specifically recorded and collected through wearable devices;

[0015] The behavioral data includes the elderly's daily activity, exercise volume, gait records, sleep quality assessment records, and dietary records, which are specifically recorded and collected through motion sensors and elderly health surveys;

[0016] The nursing data includes elderly care record data, medical treatment record data, and medication record data, which are specifically collected through electronic health systems and manual records by nursing staff;

[0017] The environmental auxiliary data includes data on the humidity and air quality assessment of the elderly's living environment, which are collected through health surveys of the elderly and manual records by caregivers;

[0018] The mental health data includes records of the mental health status, emotional changes, and anxiety and depression level assessment data of the elderly, which are specifically collected through standardized elderly questionnaires and manual recording by caregivers;

[0019] The data optimization is specifically to preprocess and optimize the original data set of elderly care management to obtain elderly care management optimized data, including the following steps: missing value processing, outlier correction, data standardization and feature initialization.

[0020] Furthermore, the nursing planning enhancement is used to automatically generate and continuously optimize personalized nursing plans. Specifically, based on the elderly care management optimization data, a reinforcement learning method combined with a multi-task deep neural network is used to perform nursing planning enhancement to obtain elderly care planning reference data, including the following steps: constructing a multi-task learning model, improving multi-task weight sharing, constructing a reinforcement learning model, constructing an improved reward function, and enhancing nursing planning;

[0021] The multi-task learning model is constructed by constructing a prediction task combination and performing a standard deep neural network construction; the prediction task combination specifically includes a health assessment task, a chronic disease management task, a lifestyle optimization task, a diet management task, and an exercise volume prediction task;

[0022] The standard deep neural network specifically refers to a standard deep neural network including an input layer, a hidden layer, and an output layer;

[0023] The multi-task weight sharing improvement is specifically performed by designing a multi-task shared hidden layer, setting a task-specific routing, and setting a task weight dynamic sharing mechanism to obtain a multi-task weight sharing improved deep neural network;

[0024] The construction of the reinforcement learning model specifically involves constructing a basic reinforcement learning model by constructing state space and action space parameters; the state space specifically includes the health status, behavioral habits, and historical care data of the elderly; the action space specifically includes adjustment options for the care plan, and the adjustment options are specifically predicted by the multi-task weight sharing improved deep neural network;

[0025] The constructing of the improved reward function is specifically to construct the improved reward function by integrating the health improvement reward and the personalized care matching reward;

[0026] The nursing plan enhancement is specifically performed by constructing a multi-task learning model, improving the multi-task weight sharing, constructing a reinforcement learning model, and constructing an improved reward function to obtain elderly care planning reference data;

[0027] The reference data for elderly care planning specifically include health assessment results, chronic disease management plans, lifestyle optimization plans, dietary recommendation plans and exercise planning plans.

[0028] Furthermore, the behavioral health guidance is used to assess whether the elderly have bad living habits and provide them with personalized health guidance. Specifically, based on the elderly care management optimization data and the elderly care planning reference data, a bidirectional long short-term memory neural network combining temporal convolution and an improved support vector classifier is used to conduct behavioral health guidance to obtain daily behavioral health guidance reference data for the elderly, including the following steps: constructing an attention-enhanced bidirectional long short-term subnet, constructing a multi-scale optimized temporal convolution subnet, fusing temporal features, constructing a dynamic multi-objective optimized support vector machine, and behavioral health guidance;

[0029] The construction of the attention-enhanced bidirectional long-short-term subnetwork specifically involves constructing a standard long-short-term memory neural network with a bidirectional structure, and optimizing the dynamic focusing ability of time-step data by constructing an attention-enhancing mechanism to obtain trend behavioral health characteristics;

[0030] The multi-scale optimized temporal convolutional subnetwork is constructed, specifically by constructing a standard temporal convolutional network and optimizing the extraction and capture of local temporal pattern features by introducing a multi-scale feature extraction mechanism to obtain local sexual behavior health features;

[0031] The time series feature fusion is specifically to perform weighted fusion on the trend behavior health feature and the local behavior health feature through weighted fusion to obtain time series fusion feature data;

[0032] The said constructing a dynamic multi-objective optimization support vector machine specifically comprises constructing a support vector machine model with dynamic weight control optimization as a classifier of the model, and performing dynamic multi-objective optimization classification prediction based on the said time series fusion feature data to obtain dynamic prediction output data;

[0033] The behavioral health guidance is specifically carried out by constructing the attention-enhanced bidirectional long-term and short-term subnet, constructing the multi-scale optimized temporal convolutional subnet, fusing temporal features, and constructing the dynamic multi-objective optimized support vector machine to obtain reference data for daily behavioral health guidance for the elderly;

[0034] The reference data for daily behavioral health guidance for the elderly specifically includes behavioral assessment data and behavioral guidance program data;

[0035] The behavioral assessment data specifically refers to category-labeled classification data;

[0036] The behavior guidance program data specifically includes dietary intake recommendation data, daily exercise recommendation data, and work and rest schedule adjustment recommendation data.

[0037] Furthermore, the mental health care is used to detect potential emotional problems, specifically based on the elderly care management optimization data, the elderly care planning reference data and the elderly daily behavioral health guidance reference data, using the emotional dynamic modeling strategy to generate an optimization network, perform mental health care, and obtain elderly mental health counseling reference data, including the following steps: constructing an emotional dynamic modeling classification subnet, constructing a mental health counseling strategy generation subnet, constructing a counseling strategy optimization reinforcement learning model and mental health care;

[0038] The emotional dynamic modeling classification subnet is constructed, specifically by extracting and fusing multimodal features to obtain mental health fusion feature data, and by constructing a multi-task learning network combining a temporal convolutional network and a transformer model to construct the emotional dynamic modeling classification subnet, and by introducing an emotional dynamic modeling module to optimize the emotional state sequence modeling, thereby obtaining dynamic emotional modeling classification data;

[0039] The constructing of the mental health counseling strategy generation subnetwork specifically involves generating an adaptive mental health counseling strategy by combining a generative adversarial network and an attention mechanism to obtain the mental health counseling strategy generation subnetwork;

[0040] The constructing of the grooming strategy optimization reinforcement learning model is specifically to construct a standard reinforcement learning model and improve the grooming strategy optimization reward function to obtain the grooming strategy optimization reinforcement learning model;

[0041] The improved guidance strategy optimizes the reward function, specifically by dynamically adjusting and calculating the change range of emotion intensity, the matching degree of guidance and nursing behavior, and the duration of emotion improvement;

[0042] The mental health care is specifically carried out by constructing the emotional dynamic modeling classification subnet, constructing the mental health counseling strategy generation subnet and constructing the counseling strategy optimization reinforcement learning model to obtain mental health care reference data for the elderly;

[0043] The elderly psychological health counseling reference data specifically includes emotion detection result data and psychological counseling suggestion data;

[0044] The emotion detection result data specifically includes an emotion state classification label and an emotion intensity quantitative evaluation value;

[0045] The psychological counseling suggestions specifically include counseling strategies and behavior optimization strategies.

[0046] Furthermore, the elderly care optimization is used to comprehensively analyze data from nursing data management, nursing planning enhancement, behavioral health guidance and mental health care modules and implement elderly care optimization. Specifically, the elderly care planning reference data, the elderly daily behavioral health guidance reference data and the elderly mental health counseling reference data are combined to perform comprehensive elderly care optimization and obtain a comprehensive elderly care reference data set.

[0047] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0048] (1) In view of the technical problems in existing nursing planning enhancement methods, multi-task learning cannot efficiently share the features of each task, resulting in insufficient information transfer between multiple tasks. At the same time, the nursing plan lacks the ability to adjust in real time for individuals. This solution creatively adopts a reinforcement learning method combined with a multi-task deep neural network to enhance nursing planning. Through innovative shared feature learning and time series modeling of emotional states, it realizes efficient information sharing and mutual optimization between multiple tasks (health assessment, chronic disease management, etc.), and adopts a dynamically weighted multi-objective reward function to achieve the optimization of long-term nursing effects.

[0049] (2) In view of the technical problem that the existing behavioral health guidance methods cannot fully consider the impact of the personalized behavioral habits and emotional changes of the elderly, this solution creatively adopts a bidirectional long short-term memory neural network that combines temporal convolution and an improved support vector classifier to provide behavioral health guidance. By dynamically capturing the behavioral patterns and emotional fluctuations of the elderly, the health guidance strategy is adjusted in real time to achieve personalized behavioral guidance, so that the health guidance strategy can more accurately respond to the specific needs of the elderly.

[0050] (3) In view of the technical problems that the existing mental health care methods lack the technical means to monitor and evaluate the emotional state of the elderly in real time, and the counseling strategies lack personalization and cannot be dynamically adjusted according to the emotional changes of the elderly, this solution creatively adopts the strategy of emotional dynamic modeling to generate an optimized network for mental health care, and automatically generates and adjusts counseling strategies according to the emotional dynamics and the specific needs of the elderly, so as to achieve personalized and continuously effective mental health care. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a schematic diagram of the structure of an artificial intelligence-based elderly care management optimization system provided by the present invention;

[0052] Figure 2 A flowchart of the steps performed for data optimization in the nursing data management module;

[0053] Figure 3Flowchart of the steps performed for the care planning enhancement module;

[0054] Figure 4 A flowchart of the steps performed for the behavioral health coaching module;

[0055] Figure 5 Flowchart of the steps performed for the Elder Care Optimization module.

[0056] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0058] Example 1, see Figure 1 The present invention provides an artificial intelligence-based elderly care management optimization system, which includes a nursing data management module, a nursing planning enhancement module, a behavioral health guidance module, a mental health care module and an elderly care optimization module;

[0059] The nursing data management module is used for nursing data management, obtains elderly care management optimization data through nursing data management, and sends the elderly care management optimization data to the nursing planning enhancement module, the behavioral health guidance module and the mental health care module;

[0060] The nursing plan enhancement module is used for nursing plan enhancement, obtains elderly nursing plan reference data through nursing plan enhancement, and sends the elderly nursing plan reference data to the behavioral health guidance module and the mental health nursing module;

[0061] The behavioral health guidance module is used for behavioral health guidance, obtains daily behavioral health guidance reference data for the elderly through behavioral health guidance, and sends the daily behavioral health guidance reference data for the elderly to the elderly care optimization module;

[0062] The mental health care module is used for mental health care, obtains reference data on mental health counseling for the elderly through mental health care, and sends the reference data on mental health counseling for the elderly to the elderly care optimization module;

[0063] The elderly care optimization module is used for elderly care optimization, and obtains a comprehensive elderly care reference data set through elderly care optimization.

[0064] Example 2: This example is based on the above example. Figure 1 and Figure 2 The nursing data management is used to collect, store, organize and optimize the health data of the elderly. Specifically, it collects, optimizes and manages nursing data from the elderly medical care system and the elderly health survey records, obtains elderly nursing management optimization data, and stores the elderly nursing management optimization data in the database for data encryption storage;

[0065] The nursing data collection specifically collects the health data, behavior data and nursing history data of the elderly to obtain the original data set of elderly care management;

[0066] The original data set of elderly care management specifically includes physiological health data, behavioral data, nursing record data, environmental auxiliary data and mental health data;

[0067] The physiological health data includes blood pressure, blood sugar, heart rate, body temperature and blood oxygen saturation data of the elderly, which are specifically recorded and collected through wearable devices;

[0068] The behavioral data includes the elderly's daily activity, exercise volume, gait records, sleep quality assessment records, and dietary records, which are specifically recorded and collected through motion sensors and elderly health surveys;

[0069] The nursing data includes elderly care record data, medical treatment record data, and medication record data, which are specifically collected through electronic health systems and manual records by nursing staff;

[0070] The environmental auxiliary data includes data on the humidity and air quality assessment of the elderly's living environment, which are collected through health surveys of the elderly and manual records by caregivers;

[0071] The mental health data includes records of the mental health status, emotional changes, and anxiety and depression level assessment data of the elderly, which are specifically collected through standardized elderly questionnaires and manual recording by caregivers;

[0072] The data optimization is specifically to preprocess and optimize the original data set of elderly care management to obtain elderly care management optimized data, including the following steps: missing value processing, outlier correction, data standardization and feature initialization;

[0073] The missing value processing is specifically to process the missing values in the original data set of the elderly care management using the mean interpolation method to obtain missing value optimized data;

[0074] The outlier correction is specifically to use statistical analysis methods to identify and process data outliers through Z-score analysis, and to remove outliers and fill in the mean to obtain abnormal optimized data;

[0075] The data standardization specifically involves standardizing numerical data using maximum and minimum normalization, and classifying and labeling non-numerical data through manual labeling to obtain standard optimized data;

[0076] The feature initialization specifically involves performing initialization feature engineering through manual feature extraction to obtain feature optimization data, and performing data integration and division based on the obtained feature optimization data to obtain the elderly care management optimization data;

[0077] The storage management specifically involves data encryption storage and data management of the elderly care management optimization data.

[0078] Example 3: This example is based on the above example. Figure 3 The nursing planning enhancement is used to automatically generate and continuously optimize personalized nursing plans. Specifically, based on the elderly care management optimization data, a reinforcement learning method combined with a multi-task deep neural network is used to perform nursing planning enhancement to obtain elderly care planning reference data, including the following steps: constructing a multi-task learning model, improving multi-task weight sharing, constructing a reinforcement learning model, constructing an improved reward function, and enhancing nursing planning;

[0079] The multi-task learning model is constructed by constructing a prediction task combination and performing a standard deep neural network construction; the prediction task combination specifically includes a health assessment task, a chronic disease management task, a lifestyle optimization task, a diet management task, and an exercise volume prediction task;

[0080] The standard deep neural network specifically refers to a standard deep neural network including an input layer, a hidden layer, and an output layer;

[0081] The multi-task weight sharing improvement is specifically carried out by multi-task weight sharing through multi-task shared hidden layer design, task-specific routing setting and task weight dynamic sharing mechanism setting, thereby obtaining a multi-task weight sharing improved deep neural network. The calculation formula of the task weight dynamic sharing mechanism setting is:

[0082] ;

[0083] Where h kis the output hidden state of the multi-task weight sharing improved deep neural network, where k is the prediction task index, which is used to represent the specific prediction task in the prediction task combination, sig(·) is the S-type activation function, and a k is a dynamic shared weight, which is used to indicate the degree of independence of the k-th prediction task. is the independent static weight of the k-th prediction task, x is the original data input, is the shared weight updated over time t;

[0084] The construction of the reinforcement learning model specifically involves constructing a basic reinforcement learning model by constructing state space and action space parameters; the state space specifically includes the health status, behavioral habits, and historical care data of the elderly; the action space specifically includes adjustment options for the care plan, and the adjustment options are specifically predicted by the multi-task weight sharing improved deep neural network;

[0085] The improved reward function is constructed by integrating the health improvement reward and the personalized care matching reward to construct the improved reward function. The calculation formula is:

[0086] ;

[0087] Where R t is the improved reward function, is the health goal weight, The whole is the health improvement reward part, The whole is the personalized care matching reward part, where tanh(·) is the hyperbolic tangent function, is the output health prediction score of the multi-task weight sharing improved deep neural network, exp(·) is the natural base function, T is the total time, t is the time index, and p t is the characteristic vector of the nursing plan corresponding to the current time t, q hist,t is the feature vector of the historical successful nursing plan corresponding to the current time t;

[0088] Preferably, the health goal weight The specific value of is 0.65;

[0089] The nursing plan enhancement is specifically performed by constructing a multi-task learning model, improving the multi-task weight sharing, constructing a reinforcement learning model, and constructing an improved reward function to obtain elderly care planning reference data;

[0090] The reference data for elderly care planning specifically include health assessment results, chronic disease management plans, lifestyle optimization plans, dietary recommendation plans and exercise planning plans.

[0091] By performing the above operations, we can address the technical issues in existing nursing planning enhancement methods, such as the inability of multi-task learning to efficiently share the features of each task, resulting in insufficient information transfer between multiple tasks and the lack of real-time adjustment capabilities for individuals. This solution creatively adopts a reinforcement learning method combined with a multi-task deep neural network to enhance nursing planning. Through innovative shared feature learning and time series modeling of emotional states, we can achieve efficient information sharing and mutual optimization between multiple tasks (health assessment, chronic disease management, etc.), and adopt a dynamically weighted multi-objective reward function to optimize long-term nursing effects.

[0092] Example 4: This example is based on the above example. Figure 4 The behavioral health guidance is used to assess whether the elderly have bad living habits and provide them with personalized health guidance. Specifically, based on the elderly care management optimization data and the elderly care planning reference data, a bidirectional long short-term memory neural network combining temporal convolution and an improved support vector classifier is used to conduct behavioral health guidance and obtain reference data for daily behavioral health guidance for the elderly, including the following steps: constructing an attention-enhanced bidirectional long short-term subnet, constructing a multi-scale optimized temporal convolution subnet, fusing temporal features, constructing a dynamic multi-objective optimized support vector machine and behavioral health guidance;

[0093] The construction of the attention-enhanced bidirectional long-short-term subnetwork specifically involves constructing a standard long-short-term memory neural network with a bidirectional structure, and optimizing the dynamic focusing ability of time-step data by constructing an attention-enhancing mechanism to obtain trend behavioral health characteristics;

[0094] The multi-scale optimized temporal convolutional subnetwork is constructed, specifically by constructing a standard temporal convolutional network and optimizing the extraction and capture of local temporal pattern features by introducing a multi-scale feature extraction mechanism to obtain local sexual behavior health features;

[0095] The time series feature fusion is specifically to perform weighted fusion on the trend behavior health feature and the local behavior health feature through weighted fusion to obtain time series fusion feature data;

[0096] The said constructing a dynamic multi-objective optimization support vector machine specifically comprises constructing a support vector machine model with dynamic weight control optimization as a classifier of the model, and performing dynamic multi-objective optimization classification prediction based on the said time series fusion feature data to obtain dynamic prediction output data;

[0097] The calculation formula of the dynamic multi-objective optimization support vector machine is:

[0098] ;

[0099] Where L is the objective function of the dynamic multi-objective optimization support vector machine, w is the standard weight vector of the support vector machine, C is the penalty parameter, N is the total number of features in the time series fusion feature data, n is the feature index, and w is the standard weight vector of the support vector machine. n is the dynamic weight corresponding to the nth feature, is the slack variable corresponding to the nth feature, is the balance weight, The overall function is a personalized matching error function, where z n is the output data predicted by the standard support vector machine, g n It is ideal behavior guidance data;

[0100] The behavioral health guidance is specifically carried out by constructing the attention-enhanced bidirectional long-term and short-term subnet, constructing the multi-scale optimized temporal convolutional subnet, fusing temporal features, and constructing the dynamic multi-objective optimized support vector machine to obtain reference data for daily behavioral health guidance for the elderly;

[0101] The reference data for daily behavioral health guidance for the elderly specifically includes behavioral assessment data and behavioral guidance program data;

[0102] The behavioral assessment data specifically refers to category-labeled classification data;

[0103] The behavior guidance program data specifically includes dietary intake recommendation data, daily exercise recommendation data, and work and rest schedule adjustment recommendation data.

[0104] By performing the above operations, in order to address the technical problem that existing behavioral health guidance methods cannot fully consider the impact of the personalized behavioral habits and emotional changes of the elderly, this solution creatively adopts a bidirectional long short-term memory neural network combined with temporal convolution and an improved support vector classifier to provide behavioral health guidance. By dynamically capturing the behavioral patterns and emotional fluctuations of the elderly, the health guidance strategy is adjusted in real time to achieve personalized behavioral guidance, so that the health guidance strategy can more accurately respond to the specific needs of the elderly.

[0105] Example 5: This example is based on the above example. Figure 5 The mental health care is used to detect potential emotional problems, specifically based on the elderly care management optimization data, the elderly care planning reference data and the elderly daily behavioral health guidance reference data, using the emotional dynamic modeling strategy to generate an optimization network to perform mental health care and obtain the elderly mental health counseling reference data, including the following steps: constructing an emotional dynamic modeling classification subnet, constructing a mental health counseling strategy generation subnet, constructing a counseling strategy optimization reinforcement learning model and mental health care;

[0106] The emotional dynamic modeling classification subnet is constructed, specifically by extracting and fusing multimodal features to obtain mental health fusion feature data, and by constructing a multi-task learning network combining a temporal convolutional network and a transformer model to construct the emotional dynamic modeling classification subnet, and by introducing an emotional dynamic modeling module to optimize the emotional state sequence modeling, thereby obtaining dynamic emotional modeling classification data;

[0107] The constructing of the mental health counseling strategy generation subnetwork specifically involves generating an adaptive mental health counseling strategy by combining a generative adversarial network and an attention mechanism to obtain the mental health counseling strategy generation subnetwork;

[0108] The constructing of the grooming strategy optimization reinforcement learning model is specifically to construct a standard reinforcement learning model and improve the grooming strategy optimization reward function to obtain the grooming strategy optimization reinforcement learning model;

[0109] The improved guidance strategy optimizes the reward function, which is specifically composed of dynamic adjustment calculations based on the change range of emotion intensity, the matching degree of guidance and nursing behavior, and the duration of emotion improvement. The calculation formula is:

[0110] ;

[0111] Where, is the reward function for improving the guidance strategy, r1 is the weight of the change in emotion intensity, is the emotional intensity change amplitude term, which is specifically obtained from the dynamic emotional modeling classification data, r2 is the guidance and nursing behavior matching weight, M t is the matching degree of counseling nursing behavior, which is calculated by the cosine similarity between the nursing strategy generated by the mental health counseling strategy generation subnet and the expected nursing goal. r3 is the weight of the degree of emotional improvement. S t is the duration of mood improvement, which is obtained by extracting and calculating the reduction of the standard deviation of the mood state sequence from the dynamic mood modeling classification data;

[0112] Preferably, the specific value of the emotion intensity change amplitude weight r1 is 0.4, the specific value of the counseling and nursing behavior matching weight r2 is 0.3, and the specific value of the emotion improvement duration weight r3 is 0.3;

[0113] The mental health care is specifically carried out by constructing the emotional dynamic modeling classification subnet, constructing the mental health counseling strategy generation subnet and constructing the counseling strategy optimization reinforcement learning model to obtain mental health care reference data for the elderly;

[0114] The elderly psychological health counseling reference data specifically includes emotion detection result data and psychological counseling suggestion data;

[0115] The emotion detection result data specifically includes an emotion state classification label and an emotion intensity quantitative evaluation value;

[0116] The psychological counseling suggestions specifically include counseling strategies and behavior optimization strategies.

[0117] By performing the above operations, in order to address the technical problems in existing mental health care methods, such as the lack of technical means to monitor and evaluate the emotional state of the elderly in real time, and the lack of personalization of counseling strategies, which cannot be dynamically adjusted according to the emotional changes of the elderly, this solution creatively adopts the strategy of emotional dynamic modeling to generate an optimized network for mental health care, automatically generates and adjusts counseling strategies according to emotional dynamics and the specific needs of the elderly, and realizes personalized and continuously effective mental health care.

[0118] Example 6: This example is based on the above example. Figure 1 The elderly care optimization is used to comprehensively analyze the data from the nursing data management, nursing planning enhancement, behavioral health guidance and mental health care modules and implement elderly care optimization. Specifically, the elderly care planning reference data, the elderly daily behavioral health guidance reference data and the elderly mental health counseling reference data are combined to perform comprehensive elderly care optimization and obtain a comprehensive elderly care reference data set.

[0119] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include," "comprise," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process or method comprising a set of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process or method.

[0120] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

[0121] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.

Claims

1. An artificial intelligence-based elderly care management optimization system, characterized by: Including nursing data management module, nursing planning enhancement module, behavioral health guidance module, mental health nursing module and elderly care optimization module; The nursing data management module is used for nursing data management, obtains elderly care management optimization data through nursing data management, and sends the elderly care management optimization data to the nursing planning enhancement module, the behavioral health guidance module and the mental health care module; The nursing plan enhancement module is used for nursing plan enhancement, obtains elderly nursing plan reference data through nursing plan enhancement, and sends the elderly nursing plan reference data to the behavioral health guidance module and the mental health nursing module; The behavioral health guidance module is used for behavioral health guidance, obtains daily behavioral health guidance reference data for the elderly through behavioral health guidance, and sends the daily behavioral health guidance reference data for the elderly to the elderly care optimization module; The mental health care module is used for mental health care, obtains reference data on mental health counseling for the elderly through mental health care, and sends the reference data on mental health counseling for the elderly to the elderly care optimization module; The elderly care optimization module is used for elderly care optimization, and obtains an elderly care comprehensive reference data set through elderly care optimization.

2. The artificial intelligence-based elderly care management optimization system according to claim 1 is characterized by: The nursing data management is used to collect, store, organize and optimize the health data of the elderly. Specifically, it is to collect nursing data from the elderly medical care system and the elderly health survey records, optimize data and store them in the database to obtain the elderly nursing management optimization data. The nursing data collection specifically collects the health data, behavior data and nursing history data of the elderly to obtain the original data set of elderly care management; The original data set of elderly care management specifically includes physiological health data, behavioral data, nursing record data, environmental auxiliary data and mental health data.

3. The artificial intelligence-based elderly care management optimization system according to claim 2 is characterized by: The data optimization is specifically to preprocess and optimize the original data set of elderly care management to obtain elderly care management optimized data, including the following steps: missing value processing, outlier correction, data standardization and feature initialization.

4. The artificial intelligence-based elderly care management optimization system according to claim 3 is characterized by: The nursing planning enhancement is used to automatically generate and continuously optimize personalized nursing plans. Specifically, based on the elderly care management optimization data, a reinforcement learning method combined with a multi-task deep neural network is used to perform nursing planning enhancement to obtain elderly care planning reference data, including the following steps: constructing a multi-task learning model, improving multi-task weight sharing, constructing a reinforcement learning model, constructing an improved reward function, and enhancing nursing planning; The multi-task learning model is constructed by constructing a prediction task combination and performing a standard deep neural network construction; the prediction task combination specifically includes a health assessment task, a chronic disease management task, a lifestyle optimization task, a diet management task, and an exercise volume prediction task; The standard deep neural network specifically refers to a standard deep neural network including an input layer, a hidden layer, and an output layer; The multi-task weight sharing improvement is specifically performed by designing a multi-task shared hidden layer, setting a task-specific routing, and setting a task weight dynamic sharing mechanism to obtain a multi-task weight sharing improved deep neural network; The construction of the reinforcement learning model specifically involves constructing a basic reinforcement learning model by constructing state space and action space parameters; the state space specifically includes the health status, behavioral habits, and historical care data of the elderly; the action space specifically includes adjustment options for the care plan, and the adjustment options are specifically predicted by the multi-task weight sharing improved deep neural network; The constructing of the improved reward function is specifically to construct the improved reward function by integrating the health improvement reward and the personalized care matching reward; The nursing plan enhancement is specifically performed by constructing a multi-task learning model, improving the multi-task weight sharing, constructing a reinforcement learning model, and constructing an improved reward function to obtain elderly care planning reference data; The reference data for elderly care planning specifically include health assessment results, chronic disease management plans, lifestyle optimization plans, dietary recommendation plans and exercise planning plans.

5. The artificial intelligence-based elderly care management optimization system according to claim 4 is characterized by: The improved reward function is constructed by integrating the health improvement reward and the personalized care matching reward to construct the improved reward function. The calculation formula is: ； Where R t is the improved reward function, is the health goal weight, The whole is the health improvement reward part, The whole is the personalized care matching reward part, where tanh(·) is the hyperbolic tangent function, is the output health prediction score of the multi-task weight sharing improved deep neural network, exp(·) is the natural base function, T is the total time, t is the time index, and p t is the characteristic vector of the nursing plan corresponding to the current time t, q hist,t is the feature vector of the historical successful nursing plan corresponding to the current time t.

6. The artificial intelligence-based elderly care management optimization system according to claim 5 is characterized by: The behavioral health guidance is used to assess whether the elderly have bad living habits and provide them with personalized health guidance. Specifically, based on the elderly care management optimization data and the elderly care planning reference data, a bidirectional long short-term memory neural network combining temporal convolution and an improved support vector classifier is used to provide behavioral health guidance, thereby obtaining reference data for daily behavioral health guidance for the elderly. The method includes the following steps: constructing an attention-enhanced bidirectional long short-term subnet, constructing a multi-scale optimized temporal convolution subnet, fusing temporal features, constructing a dynamic multi-objective optimized support vector machine, and providing behavioral health guidance. The construction of the attention-enhanced bidirectional long-short-term subnetwork specifically involves constructing a standard long-short-term memory neural network with a bidirectional structure, and optimizing the dynamic focusing ability of time-step data by constructing an attention-enhancing mechanism to obtain trend behavioral health characteristics; The multi-scale optimized temporal convolutional subnetwork is constructed, specifically by constructing a standard temporal convolutional network and optimizing the extraction and capture of local temporal pattern features by introducing a multi-scale feature extraction mechanism to obtain local sexual behavior health features; The time series feature fusion is specifically to perform weighted fusion on the trend behavior health feature and the local behavior health feature through weighted fusion to obtain time series fusion feature data; The said constructing a dynamic multi-objective optimization support vector machine specifically comprises constructing a support vector machine model with dynamic weight control optimization as a classifier of the model, and performing dynamic multi-objective optimization classification prediction based on the said time series fusion feature data to obtain dynamic prediction output data; The behavioral health guidance is specifically carried out by constructing the attention-enhanced bidirectional long-term and short-term subnet, constructing the multi-scale optimized temporal convolutional subnet, fusing temporal features, and constructing the dynamic multi-objective optimized support vector machine to obtain reference data for daily behavioral health guidance for the elderly; The reference data for daily behavioral health guidance for the elderly specifically includes behavioral assessment data and behavioral guidance program data; The behavioral assessment data specifically refers to category-labeled classification data; The behavior guidance program data specifically includes dietary intake recommendation data, daily exercise recommendation data, and work and rest schedule adjustment recommendation data.

7. The artificial intelligence-based elderly care management optimization system according to claim 6 is characterized by: The mental health care is used to detect potential emotional problems. Specifically, based on the elderly care management optimization data, the elderly care planning reference data and the elderly daily behavioral health guidance reference data, an emotional dynamic modeling strategy is used to generate an optimization network to perform mental health care and obtain elderly mental health counseling reference data, including the following steps: constructing an emotional dynamic modeling classification subnet, constructing a mental health counseling strategy generation subnet, constructing a counseling strategy optimization reinforcement learning model and mental health care; The emotional dynamic modeling classification subnet is constructed, specifically by extracting and fusing multimodal features to obtain mental health fusion feature data, and by constructing a multi-task learning network combining a temporal convolutional network and a transformer model to construct the emotional dynamic modeling classification subnet, and by introducing an emotional dynamic modeling module to optimize the emotional state sequence modeling, thereby obtaining dynamic emotional modeling classification data; The constructing of the mental health counseling strategy generation subnetwork specifically involves generating an adaptive mental health counseling strategy by combining a generative adversarial network and an attention mechanism to obtain the mental health counseling strategy generation subnetwork; The constructing of the grooming strategy optimization reinforcement learning model is specifically to construct a standard reinforcement learning model and improve the grooming strategy optimization reward function to obtain the grooming strategy optimization reinforcement learning model; The improved reward loss function is specifically constructed by dynamically adjusting and calculating the change range of emotion intensity, the matching degree of guidance and nursing behavior, and the duration of emotion improvement; The mental health care is specifically carried out by constructing the emotional dynamic modeling classification subnet, constructing the mental health counseling strategy generation subnet and constructing the counseling strategy optimization reinforcement learning model to obtain mental health care reference data for the elderly; The elderly psychological health counseling reference data specifically includes emotion detection result data and psychological counseling suggestion data; The emotion detection result data specifically includes an emotion state classification label and an emotion intensity quantitative evaluation value; The psychological counseling suggestions specifically include counseling strategies and behavior optimization strategies.

8. The artificial intelligence-based elderly care management optimization system according to claim 7 is characterized by: The elderly care optimization is used to comprehensively analyze data from nursing data management, nursing planning enhancement, behavioral health guidance and mental health care modules and implement elderly care optimization. Specifically, it combines the elderly care planning reference data, the elderly daily behavioral health guidance reference data and the elderly mental health counseling reference data to perform comprehensive elderly care optimization and obtain a comprehensive elderly care reference data set.

Images