Intelligent old-age care health management platform construction method and system based on Internet of Things

Abstract

The invention relates to the technical field of digital management platform construction, and discloses an Internet of Things-based smart pension health management platform construction method, which comprises the steps of collecting body detection and living environment data of pension users in a smart pension environment, and extracting key data to construct a structural similarity evaluation model, dividing the intelligent old-age care key data into structural data and non-structural data; analyzing an element association relationship of the structural data to organize the structural data to obtain organized data, and processing and mapping non-structural data by using an auto-encoder to obtain target data; after the target data is stored and sorted, permission setting and encryption are carried out on the stored data according to the differentiated identification attributes and the permission mapping rules; and evaluating the health of the pension user by using the encrypted data to obtain an initial evaluation report, then obtaining a continuous report through time sequence analysis, and constructing a user portrait and an intelligent pension health management platform. According to the invention, the analysis level of the intelligent old-age care health management platform based on the Internet of Things on the health data can be improved.

Description

Technical Field

[0001] The present invention relates to a method and system for constructing a smart elderly care health management platform based on the Internet of Things, and belongs to the technical field of digital management platform construction. Background Art

[0002] With the intensification of aging, smart elderly care and health management platforms have emerged. They are closely related to multiple fields such as the Internet of Things and medical health. They aim to use information technology to comprehensively manage the health of the elderly and improve the quality of elderly care services.

[0003] At present, the construction of smart elderly care health management platforms mostly relies on the Internet of Things technology. First, hardware such as wearable health monitoring devices and home smart sensors are equipped to collect data such as the elderly's heart rate, blood pressure, and activity trajectory, and then transmit the data to the server through wireless communications such as Bluetooth and Wi-Fi. In terms of software, a system with data storage, simple analysis, and basic reminder functions is developed, and user login, information query and other operation interfaces are set up to realize information interaction between different roles such as medical staff, the elderly and their families. However, when using this method, due to the inconsistent data standards of hardware equipment, the data formats output by sensors from different manufacturers are different, and compatibility problems such as data loss and errors are prone to occur during integration. In addition, the data analysis is at a shallow level, relying only on simple rules to judge, and lacks the use of advanced algorithms to build a deep analysis model. It is difficult to mine the health change trends and indicator associations hidden in the data. At the same time, in terms of information interaction, the needs of different user roles are not fully considered, the operation is cumbersome, and the interface is not friendly enough, resulting in a low level of analysis of elderly care health data. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a method for constructing a smart elderly care health management platform based on the Internet of Things, which can improve the analysis level of health data by the smart elderly care health management platform based on the Internet of Things.

[0005] To achieve the above-mentioned purpose, the present invention provides a method for constructing a smart elderly care health management platform based on the Internet of Things, comprising:

[0006] Using a pre-configured Internet of Things device to collect physical examination data and living environment data corresponding to elderly care users in a smart elderly care environment, extracting key data from the physical examination data and the living environment data to obtain smart elderly care key data, constructing a structural similarity evaluation model for the smart elderly care key data, and using the structural similarity evaluation model to perform data structure classification on the smart elderly care key data to obtain structured data and unstructured data;

[0007] Analyze the association relationship of data elements in the structured data, use the association relationship to construct a relational data table of the structured data, organize the structured data to obtain organized data, construct an autoencoder for the unstructured data, use the autoencoder to deconstruct the unstructured data to obtain initial processed data, map the initial processed data to a latent space, and obtain target data;

[0008] The sorted data and the target data are stored as data to obtain stored data, a differential identification attribute of the stored data is identified, permissions are set for the stored data based on the differential identification attribute and a preset permission mapping rule to obtain permission data, and the permission data is encrypted to obtain encrypted data;

[0009] The encrypted data is used to conduct a preliminary health assessment of the elderly care user to obtain a preliminary health assessment report; the preliminary health assessment report is used to conduct a time series analysis of the health of the elderly care user to obtain a continuous health analysis report; the continuous health analysis report is used to construct a user portrait of the elderly care user; the user portrait is used to conduct an intelligent health warning for the elderly care user; based on the warning results of the intelligent health warning, a smart elderly care health management platform for the elderly care user is constructed.

[0010] Optionally, extracting key data from the physical examination data and the living environment data to obtain key data for smart elderly care includes:

[0011] Querying the physiological data and neural data in the body detection data and the gas detection data and temperature and humidity data of the living environment data;

[0012] Based on the physiological data and neural data, construct a physiological and pathological model of the elderly user;

[0013] Constructing a quantitative model of environmental health risks of the elderly user according to the gas detection data and the temperature and humidity data;

[0014] Using the physiological and pathological model to analyze the trend of physical data of the elderly user, and obtain physical data analysis features;

[0015] Using the environmental health risk quantification model, analyzing the environmental data changes of the elderly care users to obtain environmental change analysis features;

[0016] Constructing a joint feature vector of the body data analysis feature and the environment change analysis feature;

[0017] Feature screening is performed on the joint feature vector to obtain key data for smart elderly care.

[0018] Optionally, the step of constructing a structural similarity evaluation model for the key data of smart elderly care includes:

[0019] Performing hash coding on the key data of smart elderly care to obtain a data code;

[0020] identifying a Hamming distance metric value encoding the data;

[0021] Constructing a coding matrix for encoding the data;

[0022] Based on the Hamming distance metric and the coding matrix, a structural similarity evaluation model of the key data of smart elderly care is constructed;

[0023] Wherein, the structural similarity evaluation model is in the following form:

[0024]

[0025] Among them, φ represents the structural similarity evaluation model, n represents the amount of key data of smart elderly care, α represents the model search coefficient, represents the Hamming distance metric, Represents the i-th structure data in the encoded data, represents the i-th unstructured data in the encoded data, M represents the number of rows in the encoding matrix, L represents the number of columns in the encoding matrix, β represents the weight coefficient of the encoding parameter corresponding to the data encoding, R(x j,k ) represents the hash loss function, x j,k Represents the data at the corresponding position in the data encoding (the data related to the j-th row and k-th column position in the encoding matrix).

[0026] Optionally, analyzing the association relationship of data elements in the structure data includes:

[0027] Classifying the data elements in the structure data to obtain classified elements;

[0028] The element frequency of the classification element is calculated using the following formula:

[0029]

[0030] Among them, δ represents the frequency of elements, T iE represents the number of times element i appears in the Eth type element, T eE Indicates the number of times element e appears in the Eth type of elements;

[0031] Based on the frequency of the elements, construct a frequent element combination;

[0032] Calculating the support and confidence of the frequent element combination;

[0033] Based on the support and the confidence, an association relationship between data elements in the structure data is determined.

[0034] Optionally, mapping the initial processed data into a latent space to obtain target data includes:

[0035] Performing data normalization processing on the initial processed data to obtain normalized data;

[0036] extracting data points of the normalized data;

[0037] The high-dimensional similarity probability of different data points in the data points is calculated using the following formula:

[0038]

[0039] Among them, P represents the similarity probability, Y a represents data point i, Y b represents data point j, σ represents the similarity range control parameter, ∥Y a -Y b ∥ means Y a With Y b The Euclidean distance, H and F represent the indexes of the data points, and their values ​​range from 1 to the total number of data points n (assuming there are n data points in total), ∑ H≠F exp(-∥Y H -Y F ∥ 2 / 2σ 2 ) represents the normalization factor, Y H Represents the data point pair H, Y obtained by indexing the data point F Represents the data point pair F (H≠F) obtained by indexing the data point;

[0040] Mapping the normalized data into a latent space to obtain mapped data;

[0041] Constructing low-dimensional initialization data points of the mapping data;

[0042] Calculating the low-dimensional similarity probability of the low-dimensional initialization data points;

[0043] Based on the high-dimensional similarity probability and the low-dimensional similarity probability, the position of the low-dimensional initialization data point is updated to obtain target data.

[0044] Optionally, the setting of permissions for the stored data based on the differentiated identification attribute and a preset permission mapping rule to obtain permission data includes:

[0045] identifying data sensitivity of the stored data;

[0046] Based on the data sensitivity, user definition is performed on the access role corresponding to the stored data to obtain a defined user;

[0047] Associating the differentiated identification attribute with the defined user to obtain data-role information;

[0048] Based on the data-role information and the preset permission mapping rules, an attribute-permission mapping relationship table of the stored data is constructed to set permissions for the stored data, thereby obtaining permission data.

[0049] Optionally, encrypting the authority data to obtain encrypted data includes:

[0050] Identify a data node of the permission data;

[0051] Configuring a dynamic key for the permission data;

[0052] The dynamic key can be expressed by the following formula:

[0053] B=(J1+r1)×(J2+r2)×…×(J n +r n )

[0054] Where B represents the dynamic key, J1 represents the first node in the data node, r1 represents the random number of J1, J2 represents the second node in the data node, r2 represents the random number of J2, and J n Represents the nth node in the data node, r n Indicates J n A random number;

[0055] According to the data node, the dynamic key is dynamically updated using the following formula to obtain an updated key:

[0056] B ′ =G(r ′ ,B)

[0057] Among them, B ′ represents the update key, G represents the update function, r ′ represents the update of random number, and B represents the dynamic key;

[0058] The authority data is encrypted using the update key to obtain encrypted data.

[0059] Optionally, the preliminary health assessment report is used to perform a time series analysis of the health of the elderly user to obtain a continuous health analysis report, including:

[0060] Obtaining an initial time series of a target health indicator from the preliminary health assessment report;

[0061] Performing a stationarity test on the initial time series;

[0062] When the result of the stationarity test does not meet the preset requirements, performing difference processing on the initial time series to obtain a target time series;

[0063] Calculating the autocorrelation function and the partial autocorrelation function of the target time series to obtain an autocorrelation calculation result;

[0064] Based on the autocorrelation calculation result, a time series analysis of the health of the elderly care user is performed to obtain a continuous health analysis report.

[0065] Optionally, the step of constructing a smart elderly care health management platform for the elderly care user based on the warning result of the health intelligent warning includes:

[0066] Constructing the initial platform of the smart elderly care health management platform for the elderly care users;

[0067] Analyze the generation steps corresponding to the warning results of the health intelligent warning;

[0068] Based on the generating step, constructing a data processing port of the health management platform corresponding to the elderly care user;

[0069] Using the intelligent health warning data of the elderly care user, port debugging is performed on the data processing port;

[0070] When the port debugging result is normal, obtaining the target port;

[0071] The data processing port is configured in the initial platform, and a user management interface is configured for the initial platform to obtain a smart elderly care health management platform for the elderly care users.

[0072] In a second aspect, the present invention provides a system for constructing a smart elderly care health management platform based on the Internet of Things, the system comprising:

[0073] A data structure classification module is used to collect physical examination data and living environment data corresponding to elderly care users in a smart elderly care environment by using a pre-configured Internet of Things device, extract key data from the physical examination data and the living environment data, obtain key data for smart elderly care, construct a structural similarity evaluation model for the key data for smart elderly care, and use the structural similarity evaluation model to perform data structure classification on the key data for smart elderly care to obtain structured data and unstructured data;

[0074] A data processing module is used to analyze the association relationship of data elements in the structured data, construct a relational data table of the structured data by using the association relationship, organize the structured data to obtain organized data, construct an autoencoder for the unstructured data, deconstruct the unstructured data by using the autoencoder to obtain initial processed data, and map the initial processed data to a latent space to obtain target data;

[0075] A data encryption module, used for storing the sorted data and the target data to obtain stored data, identifying a differential identification attribute of the stored data, setting permissions for the stored data based on the differential identification attribute and a preset permission mapping rule to obtain permission data, and encrypting the permission data to obtain encrypted data;

[0076] The platform construction module is used to use the encrypted data to perform a preliminary health assessment on the elderly user to obtain a preliminary health assessment report, use the preliminary health assessment report to perform a time series analysis of the health of the elderly user to obtain a continuous health analysis report, use the continuous health analysis report to build a user portrait of the elderly user, use the user portrait to perform intelligent health warning for the elderly user, and based on the warning results of the intelligent health warning, build a smart elderly health management platform for the elderly user.

[0077] Compared with the existing technology, the present invention uses pre-configured Internet of Things devices to widely collect physical examination data and living environment data of elderly care users, covering multiple types of key information such as blood pressure, heart rate, temperature and humidity of the living environment, etc., comprehensively breaking through the limitations of traditional elderly care information acquisition, ensuring the timeliness and completeness of information, and providing a rich data source for subsequent precise services; by accurately extracting key data and building a structural similarity evaluation model to classify data, it can effectively distinguish between structured data and unstructured data, and improve the pertinence and efficiency of data processing. For structured data, we conduct in-depth analysis of element associations to construct relational data tables to ensure that data logic is clear and of high quality. We use autoencoders to deconstruct and process unstructured data to tap its potential value, promote synergy between different types of data, and support elderly care health management decisions in multiple dimensions. We store and organize data and target data to form a data resource library, identify differentiated identification attributes through data mining, set strict permissions and encrypt data based on permission mapping rules, and comprehensively ensure the confidentiality, integrity and availability of data, effectively prevent data leakage and illegal operation risks, and build a solid security line for the stable operation of the smart elderly care system. We conduct preliminary health assessments based on encrypted data and generate reports, and conduct time series analysis based on this to obtain continuous health analysis reports, integrate various types of information to build user portraits, and then set personalized early warning rules based on the portraits to achieve intelligent health early warnings, and feed back the early warning results to each link of the smart elderly care health management platform for optimization and improvement, thereby improving the analysis level of health data on the smart elderly care health management platform based on the Internet of Things. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 A flowchart of a method for constructing a smart elderly care health management platform based on the Internet of Things provided by one embodiment of the present invention;

[0079] Figure 2 A schematic diagram of modules of a system for building a smart elderly care and health management platform based on the Internet of Things provided by one embodiment of the present invention.

[0080] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0081] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0082] The embodiment of the present application provides a method for constructing a smart elderly care health management platform based on the Internet of Things. The execution subject of the method for constructing a smart elderly care health management platform based on the Internet of Things includes but is not limited to at least one of the electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the method for constructing a smart elderly care health management platform based on the Internet of Things can be executed by software or hardware installed on a terminal device or a server device. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0083] Embodiment 1:

[0084] Reference Figure 1 FIG. 1 is a flow chart of a method for constructing a smart elderly care health management platform based on the Internet of Things provided by an embodiment of the present invention. In this embodiment, the method for constructing a smart elderly care health management platform based on the Internet of Things includes:

[0085] S1. Use a pre-configured Internet of Things device to collect physical examination data and living environment data corresponding to elderly care users in a smart elderly care environment, extract key data from the physical examination data and the living environment data to obtain smart elderly care key data, construct a structural similarity evaluation model for the smart elderly care key data, and use the structural similarity evaluation model to classify the data structure of the smart elderly care key data to obtain structured data and unstructured data.

[0086] The embodiment of the present invention can realize all-round monitoring of elderly care users by using the pre-configured Internet of Things device to collect the physical examination data and living environment data of elderly care users in the smart elderly care environment, changing the situation of untimely and incomplete information acquisition in the traditional elderly care model. For example, continuous heart rate data can help medical staff to timely detect whether the elderly have heart problems, and indoor harmful gas concentration data can prevent dangerous situations such as gas poisoning.

[0087] Among them, the preconfigured IoT device refers to a collection of smart devices that are pre-set and deployed in a smart elderly care environment and are used to collect various types of data, such as smart bracelets / watches, smart blood pressure monitors, temperature and humidity sensors, gas sensors, and smart cameras.

[0088] The embodiment of the present invention extracts the key data from the physical test data and the living environment data to obtain the key data of smart elderly care, which can screen out the most valuable parts for elderly health management. For example, in the physical test data, focus on extracting indicators related to common elderly diseases, such as blood pressure, blood sugar, blood oxygen saturation, etc.; in the living environment data, extract factors that may affect the health of the elderly, such as whether the temperature is too high or too low, whether the air quality meets the standard, etc.

[0089] As an embodiment of the present invention, the extraction of key data from the body detection data and the living environment data to obtain key data for smart elderly care includes: querying the physiological data and neural data in the body detection data and the gas detection data and temperature and humidity data of the living environment data, constructing a physiological and pathological model of the elderly user based on the physiological data and neural data, constructing an environmental health risk quantification model for the elderly user according to the gas detection data and the temperature and humidity data, performing a trend analysis of the body data of the elderly user using the physiological and pathological model to obtain body data analysis features, performing a change analysis of the environmental data of the elderly user using the environmental health risk quantification model to obtain environmental change analysis features, constructing a joint feature vector of the body data analysis features and the environmental change analysis features, performing feature screening on the joint feature vector to obtain key data for smart elderly care.

[0090] Among them, the physiological data refers to various quantifiable indicator data reflecting the physiological functions and physiological states of the elderly users, such as blood pressure (systolic pressure, diastolic pressure), blood sugar (fasting blood sugar, postprandial blood sugar), heart rate and blood oxygen saturation, etc. The neural data refers to data used to indirectly or directly reflect the functional state of the elderly users' nervous system, mainly focusing on the nervous system's control over body movement and sensation, such as the step length, step frequency, gait cycle, double-foot support time, plantar pressure distribution and other data obtained from gait analysis; the gas detection data refers to the data obtained by detecting the concentration of various gas components in the elderly living environment, such as Such as data such as carbon monoxide (CO) concentration, carbon dioxide (CO2) concentration and formaldehyde (HCHO) concentration; the temperature and humidity data refer to data obtained by real-time monitoring of the temperature and relative humidity in the elderly living environment; the physiological and pathological model refers to a model constructed based on medical knowledge, mathematical algorithms and data analysis technology, which is used to describe the complex relationship between the physiological data of elderly users and the occurrence and development process of the disease, and can be constructed through a machine learning model; the health risk quantification model is a model used to quantify the potential hazards of environmental factors (such as gas, temperature and humidity) to the health of elderly users, which can be constructed through a machine learning model.

[0091] Optionally, the physiological data and neural data can be retrieved or screened by database retrieval or screening tools to find out the physiological and neural data in the body test and the gas, temperature and humidity data of the living environment from the corresponding data sets by category. The physiological and pathological model can be constructed based on the association between the physiological and neural data and the corresponding diseases using statistical analysis or machine learning algorithms, such as using regression analysis of the relationship between cardiovascular-related indicators and risk factors. The environmental health risk quantification model can be constructed based on the diffusion principle of chemical substances in the indoor environment and the method for calculating human exposure dose. Taking formaldehyde as an example, the mass balance equation is used to consider the indoor space volume, ventilation rate, formaldehyde release rate, etc. Based on factors such as radiation source intensity (estimated based on furniture materials, decoration materials, etc.), a dynamic model of formaldehyde concentration changing over time is established, and combined with the dose-effect relationship of formaldehyde on human health hazards determined by toxicological studies, the health risks of the elderly in environments with different formaldehyde concentrations are quantified. The body data analysis features can be obtained by using the physiological and pathological model in combination with time series analysis techniques, such as the autoregressive moving average (ARIMA) model, exponential smoothing method, etc. to perform trend analysis on physiological data and neural data. The environmental change analysis features can be obtained by using the environmental health risk quantification model in combination with the change point detection algorithm to analyze the gas detection data and temperature and humidity data. For example, a change point detection method based on a cumulative sum (CUSUM) control chart is used to monitor carbon monoxide concentration data. When the carbon monoxide concentration exceeds a preset threshold and lasts for a certain period of time (set by algorithm parameters), it is determined to be a change point. The data before and after the change point and related environmental parameters (such as the operating status of the ventilation equipment, whether there is a combustion source nearby, etc.) are extracted as environmental change analysis features. The joint feature vector can be integrated with the body data analysis features and the environmental change analysis features. For example, the blood pressure trend features, gait abnormality features, carbon monoxide concentration change features, and temperature and humidity change features are combined into a vector. At the same time, metadata such as timestamp, location information of the room where the elderly is located, and data source device number are added. The feature screening of the joint feature vector to obtain key data for smart elderly care can be performed by adopting a machine learning feature selection algorithm (such as random forest feature importance evaluation), with health status as the output, and filtering out key data according to feature importance scores (specific settings need to be combined with actual applications).

[0092] The embodiment of the present invention can provide a quantitative standard and method for data classification by constructing the structural similarity evaluation model of the key data of smart elderly care, and reasonably divide the data according to the similarity of the data structure so as to adopt different processing strategies to adapt to different types of data processing requirements.

[0093] As an embodiment of the present invention, the construction of the structural similarity evaluation model of the key data of smart elderly care includes: performing hash coding on the key data of smart elderly care to obtain data coding, identifying the Hamming distance measurement value of the data coding, constructing a coding matrix of the data coding, and constructing a structural similarity evaluation model of the key data of smart elderly care based on the Hamming distance measurement value and the coding matrix, wherein the structural similarity evaluation model is in the following form:

[0094]

[0095] Among them, φ represents the structural similarity evaluation model, n represents the amount of key data of smart elderly care, α represents the model search coefficient, represents the Hamming distance metric, Represents the i-th structure data in the encoded data, represents the i-th unstructured data in the encoded data, M represents the number of rows in the encoding matrix, L represents the number of columns in the encoding matrix, β represents the weight coefficient of the encoding parameter corresponding to the data encoding, R(x j,k ) represents the hash loss function, x j,k Represents the data at the corresponding position in the data encoding (the data related to the j-th row and k-th column position in the encoding matrix).

[0096] Optionally, the data code can be obtained by encoding the key data of smart elderly care using a suitable hash algorithm (such as a locally sensitive hash algorithm). The Hamming distance metric value can be obtained by organizing all the data codes into a data set form, and then for each pair of data codes, the Hamming distance metric value is determined by calculating the number of different corresponding bits. For example, if the two data codes are "10101" and "11100", the Hamming distance is 2 because there are two different bits. All data code pairs are traversed to obtain a corresponding set of Hamming distance metric values. The coding matrix can determine the number of rows and columns of the coding matrix based on factors such as the number of data codes and the coding length. For example, if there are m data codes and each code length is k, a coding matrix with m rows and k columns can be constructed.

[0097] The embodiment of the present invention utilizes the structural similarity evaluation model to perform data structure classification on the key data of smart elderly care, and obtains structured data and unstructured data, which can provide a clear path for the entire data processing flow. Different types of data can be deeply mined under their respective suitable processing frameworks, thereby improving the efficiency and accuracy of data processing.

[0098] Optionally, the structural similarity evaluation model is used to classify the key data of smart elderly care into data structures, and the process of obtaining structured data and unstructured data is as follows:

[0099] Substitute the Hamming distance metric, encoding matrix and related data into the model formula, use training data for training, and continuously adjust the model parameters through optimization algorithms (such as gradient descent method) so that the model can accurately evaluate the structural similarity of key data of smart elderly care according to data encoding. For example, during the training process, the loss value calculated by the loss function is used to adjust the weight coefficient and other parameters through back propagation. After multiple rounds of iterations, until the model converges, the target structural similarity evaluation model is obtained, and then the target structural similarity evaluation model is used to perform data structure classification on the key data of smart elderly care.

[0100] S2. Analyze the association relationship among the data elements in the structured data, and use the association relationship to construct a relational data table of the structured data to organize the structured data to obtain organized data, construct an autoencoder for the unstructured data, use the autoencoder to deconstruct the unstructured data to obtain initial processed data, and map the initial processed data to a latent space to obtain target data.

[0101] The embodiment of the present invention can analyze the association relationship of the data elements in the structured data to clearly identify the association relationship, which helps to integrate the scattered data elements in a meaningful way, avoid isolated storage and use of data, improve data management efficiency, and facilitate subsequent data query, analysis, and data-based decision making.

[0102] As an embodiment of the present invention, the analyzing the association relationship of the data elements in the structure data includes: classifying the data elements in the structure data to obtain classified elements, and calculating the element frequency of the classified elements using the following formula:

[0103]

[0104] Among them, δ represents the frequency of elements, T iE represents the number of times element i appears in the Eth type element, T eE Indicates the number of times element e appears in the Eth type of elements;

[0105] Based on the element frequency, a frequent element combination is constructed, the support and confidence of the frequent element combination are calculated, and based on the support and the confidence, the association relationship of the data elements in the structure data is determined.

[0106] Optionally, the data elements in the structured data are classified to obtain classified elements, which can be achieved by using a clustering algorithm in machine learning (such as a K-Means clustering algorithm) to assist in classification. The frequent element combination constructed based on the element frequency can first set a suitable element frequency threshold (such as 0.2, etc.) according to the characteristics and analysis requirements of the elderly health management data, and select elements with an element frequency greater than or equal to the threshold as frequent elements. For example, in cardiovascular elements, if the element frequency of "hypertension" is higher than the set threshold, then "hypertension" is a frequent element. From the screened frequent elements, element combinations of different lengths (such as binary combinations, ternary combinations, etc.) are generated by combination. Combinations can be generated by exhaustive enumeration or based on some optimization strategies (such as the prior properties in the Apriori algorithm to reduce unnecessary combination attempts). For example, among the selected cardiovascular frequent elements "hypertension", "high heart rate" and "hyperlipidemia", binary combinations "hypertension, high heart rate", "hypertension, high blood lipids", "high heart rate, high blood lipids", etc., and possible ternary combinations "hypertension, high heart rate, high blood lipids", etc. are generated; the calculation of the support and confidence of the frequent element combination can be realized by programming code (such as data processing and statistical functions in Python) or the query statistical function of the database (SQL statement), and the determination of the association relationship of the data elements in the structure data based on the support and the confidence can set the support threshold (such as 0.1) and the confidence threshold (such as 0.5) according to the business needs of elderly health management and data analysis experience. Only when the support of the frequent element combination is greater than or equal to the support threshold, and its confidence is greater than or equal to the confidence threshold, it is considered that there is a meaningful association relationship between the data elements in the combination. For example, the support of the combination of "hypertension, high heart rate" is 0.2, and the confidence is 0.5, which meets the set threshold requirements, then it can be determined that there is an association relationship between "hypertension" and "high heart rate".

[0107] The embodiment of the present invention utilizes the association relationship to construct a relational data table of the structured data to organize the structured data. The organized data can intuitively present the data logic, ensure the consistency and accuracy of the data, remove redundant and erroneous information in the structured data, and optimize the data format and order, thereby improving the data quality.

[0108] Optionally, the relationship data table can be constructed using a data visualization tool (such as Graphviz, etc.), and the sorted data can be obtained by using the visualization tool to display the screened meaningful association relationships in the form of a network diagram.

[0109] The relational data table refers to a visual chart for displaying data.

[0110] Furthermore, the embodiment of the present invention can extract a low-dimensional feature vector that best represents the key information of the original data by encoding and compressing the input data through the autoencoder for constructing the unstructured data, thereby realizing feature abstraction of the unstructured data.

[0111] The autoencoder refers to a neural network architecture that aims to learn effective feature representation of data by encoding input data into a low-dimensional representation and then decoding it back to the original data. It is often used for processing and feature extraction of unstructured data.

[0112] Optionally, the autoencoder can be obtained by training a neural network model.

[0113] The embodiment of the present invention uses the autoencoder to deconstruct the unstructured data to obtain initial processed data, which can make it easier to establish an association between the unstructured data and the structured data after processing. For example, the emotional features deconstructed from the voice are associated with the health test data of the same period, which helps to discover the potential impact of emotions on health and improve the accuracy of the overall platform's comprehensive assessment of elderly care users.

[0114] Optionally, the process of using the autoencoder to deconstruct the unstructured data to obtain initial processed data is: input the unstructured data into the encoder part of the trained autoencoder. For example, for a video data of daily activities of an elderly person, each frame of the video is input into the encoder of the autoencoder in a set order (such as time order), and then the encoder performs feature extraction operations on the input data, and converts the high-dimensional video data into a low-dimensional feature vector (initial processed data) according to the learned feature representation method.

[0115] Furthermore, the embodiment of the present invention can break the barriers between unstructured data of different modalities by mapping the initial processed data into the latent space to obtain target data, so that various data can complement each other and work together, providing a basis for elderly health management from multiple perspectives. For example, by combining the cross-modal target data corresponding to the elderly activity video and voice recording, it is possible to more accurately determine whether the elderly are in a state of depression, inconvenience in life, etc., and provide corresponding care and services in a timely manner.

[0116] The latent space refers to an abstract, low-dimensional space used to represent the hidden features or potential structures of data.

[0117] As an embodiment of the present invention, mapping the initial processed data to a latent space to obtain target data includes: performing data normalization processing on the initial processed data to obtain normalized data, extracting data points of the normalized data, and calculating the high-dimensional similarity probability of different data points in the data points using the following formula:

[0118]

[0119] Among them, P represents the similarity probability, Y a represents data point i, Y b represents data point j, σ represents the similarity range control parameter, ∥Y a -Y b ∥ means Y a With Y b The Euclidean distance, H and F represent the indexes of the data points, and their values ​​range from 1 to the total number of data points n (assuming there are n data points in total), ∑ H≠F exp(-∥Y H -Y F ∥ 2 / 2σ 2 ) represents the normalization factor, Y H Represents the data point pair H, Y obtained by indexing the data point F Represents the data point pair F (H≠F) obtained by indexing the data point;

[0120] The normalized data is mapped into a latent space to obtain mapping data, low-dimensional initialization data points of the mapping data are constructed, low-dimensional similarity probabilities of the low-dimensional initialization data points are calculated, and based on the high-dimensional similarity probability and the low-dimensional similarity probability, the positions of the low-dimensional initialization data points are updated to obtain target data.

[0121] It should be further explained that in the above formula, Y a and Y b It is mainly used to calculate the similarity probability P between these two specific data points. They are a pair of data points that the formula focuses on, which is used to measure the similarity between these two points in high-dimensional space, Y H and Y F (H≠F) is a normalization factor used to traverse all possible pairs of data points to calculate the similarity probability distribution of the entire data set. Their combination is used to comprehensively consider the similarity relationship between all data points to ensure that the final calculated P is a legal probability distribution.

[0122] Optionally, the data normalization processing is performed on the initial processed data to obtain normalized data, which can be achieved through Z-Score standardization. The mapping of the normalized data to the latent space to obtain the mapped data can be achieved through the t-distributed stochastic neighbor embedding (t-SNE) algorithm. The principle and steps of calculating the low-dimensional similarity probability of the low-dimensional initialization data points are the same as the high-dimensional similarity probability of different data points in the data points calculated using the following formula, so they are not repeated here. The position of the low-dimensional initialization data points is updated based on the high-dimensional similarity probability and the low-dimensional similarity probability to obtain the target data, which can be obtained by adopting a method such as gradient descent, with the difference between the high-dimensional similarity probability and the low-dimensional similarity probability (commonly measured by KL divergence) as the target, calculating the gradient and updating the coordinates of the low-dimensional initialization data points. After multiple iterations, the difference is reduced, and the final stable low-dimensional data point is the target data.

[0123] S3. Storing the sorted data and the target data to obtain stored data, identifying the differential identification attributes of the stored data, setting permissions for the stored data based on the differential identification attributes and a preset permission mapping rule to obtain permission data, and encrypting the permission data to obtain encrypted data.

[0124] The embodiment of the present invention obtains the stored data by storing the sorted data and the target data. By storing these data, a data resource library in the field of smart elderly care can be gradually accumulated to provide a rich data basis for subsequent data analysis, health assessment, service decision-making and other work at different stages. For example, in the subsequent study of the health change trend of elderly care users, data of the corresponding time period can be directly extracted from the stored data for analysis, while also protecting the security of the data.

[0125] Optionally, the sorted data and the target data may be stored via a preconfigured data repository.

[0126] The embodiment of the present invention can help to classify and organize the stored data according to different identification attributes by identifying the differentiated identification attributes of the stored data, thereby improving the efficiency and accuracy of data management. For example, the data of elderly people with cardiovascular diseases can be classified into one category for special management and analysis, so that medical staff can focus on the data changes of this group. Among them, the differentiated identification attribute refers to a feature that can distinguish different individuals, different categories or different states in the stored data.

[0127] Optionally, the differential identification attributes may be identified by utilizing data mining and machine learning techniques.

[0128] Furthermore, the embodiment of the present invention sets permissions for the stored data based on the differentiated identification attributes and pre-set permission mapping rules, and obtains permission data to prevent the data from being arbitrarily accessed, tampered with or deleted by unauthorized personnel, thereby ensuring the integrity, confidentiality and availability of data in the smart elderly care system, and avoiding the risk of data leakage and system operation failures caused by confusion in permissions.

[0129] Among them, the pre-set permission mapping rules refer to a set of pre-defined rule systems used to clarify the correspondence between different differentiated identification attributes (such as user's age, health status, service level, etc.) and different user roles (such as medical staff, elderly care service personnel, managers, etc.) for data access and operation permissions.

[0130] As an embodiment of the present invention, the method of setting permissions for the stored data based on the differentiated identification attribute and the preset permission mapping rules to obtain permission data includes: identifying the data sensitivity of the stored data, based on the data sensitivity, user-defining the access role corresponding to the stored data to obtain a defined user, associating the differentiated identification attribute with the defined user to obtain data-role information, and constructing an attribute-permission mapping relationship table of the stored data based on the data-role information and the preset permission mapping rules to set permissions for the stored data to obtain permission data.

[0131] The data sensitivity refers to the importance of data in terms of confidentiality, integrity and availability, as well as the potential risks and impact that may be caused by data leakage, tampering or loss.

[0132] Optionally, the data sensitivity can be divided according to the purpose and scope of use of the data. If the data is only used for internal medical diagnosis and is strictly confidential, such as details of mental illness diagnosis, it is marked as high sensitivity; if the data is used for daily elderly care service communication, such as hobbies, it can be identified as low sensitivity;

[0133] Based on the data sensitivity, the access roles corresponding to the stored data are user defined, and the defined users can determine the role range allowed to access according to the data sensitivity level. For example, high-sensitivity data is only accessible to specific attending physicians, data security administrators and other roles, and these roles are defined as the access roles of the corresponding data; medium-sensitivity data can be extended to relevant nursing staff, etc.; low-sensitivity data can be opened to more elderly care service personnel, etc. A corresponding access role list is sorted out for each sensitivity data to form a defined user set. The differentiated identification attribute is associated with the defined user role to obtain data-role information. The access requirements corresponding to each attribute can be analyzed by viewing the determined differentiated identification attributes (such as the age of the elderly, medical history, service level, etc.). For example, for the attribute of "suffering from cardiovascular disease", it is associated with defined users such as doctors and cardiovascular specialist nurses who need to access such patient data; for the attribute of "special nursing service level", it is associated with defined users such as members of the nursing team responsible for the service of this level. The attribute-permission mapping relationship table can be used using spreadsheet software (such as Excel) or a database based on data-role information, combined with previously established pre-set permission mapping rules (clarifying the permission scope of each role corresponding to different operations such as viewing, modifying, etc.), to correspond specific permission content to each data-role association.

[0134] Furthermore, the embodiment of the present invention can construct a secure data environment by encrypting the permission data to obtain encrypted data, thereby enhancing data confidentiality.

[0135] As an embodiment of the present invention, encrypting the permission data to obtain encrypted data includes: identifying a data node of the permission data, and configuring a dynamic key of the permission data, wherein the dynamic key can be expressed by the following formula:

[0136] B=(J1+r1)×(J2+r2)×…×(J n +r n )

[0137] Where B represents the dynamic key, J1 represents the first node in the data node, r1 represents the random number of J1, J2 represents the second node in the data node, r2 represents the random number of J2, and J n Represents the nth node in the data node, r n Indicates J n A random number;

[0138] According to the data node, the dynamic key is dynamically updated using the following formula to obtain an updated key:

[0139] B ′=G(r ′ ,B)

[0140] Among them, B ′ represents the update key, G represents the update function, r ′ represents the update of random number, and B represents the dynamic key;

[0141] The authority data is encrypted using the update key to obtain encrypted data.

[0142] Optionally, the data node can be identified through a data management tool (such as a database management client for viewing the database table structure) or a data parsing tool (for files in a specific format), and the dynamic key can be edited through a programming library in a Python tool.

[0143] S4. Using the encrypted data, conduct a preliminary health assessment on the elderly care user to obtain a preliminary health assessment report; using the preliminary health assessment report, conduct a time series analysis of the health of the elderly care user to obtain a continuous health analysis report; using the continuous health analysis report, construct a user portrait of the elderly care user; using the user portrait, conduct an intelligent health warning for the elderly care user; based on the warning results of the intelligent health warning, construct a smart elderly care health management platform for the elderly care user.

[0144] The embodiment of the present invention utilizes the encrypted data to conduct a preliminary health assessment of the elderly care user, and the preliminary health assessment report obtained can help medical staff, health managers or elderly care service personnel quickly understand the user's health profile, so as to formulate a preliminary health management plan in a targeted manner, such as arranging specific physical examination items or recommending suitable health activities.

[0145] Optionally, the encrypted data is used to conduct a preliminary health assessment of the elderly care user, and the preliminary health assessment report can be obtained by using data processing technology, such as Python's Pandas library for data sorting. Various health assessment indicators are calculated based on the extracted data, such as calculating the BMI value based on height and weight data, and calculating the cardiovascular disease risk score through blood pressure, blood sugar and other indicators (relevant medical statistical formulas can be used). Based on the calculation results, a comprehensive health assessment is performed in combination with the medical knowledge base or expert system rules. For example, compare the indicator with the normal range to determine the health status. Finally, a preliminary health assessment report containing basic information, health indicator details, risk warnings and preliminary management suggestions is generated using document generation technology or database storage procedures.

[0146] The embodiment of the present invention uses the preliminary health assessment report to perform a time series analysis of the health of the elderly user, and obtains a continuous health analysis report, which can mine the inherent laws of health data changes over time and predict the possible trend of future health conditions. For example, by analyzing the fluctuation trend of indicators such as blood pressure and blood sugar over a period of time, it is predicted whether there will be risks such as hypertension and diabetic complications, providing a basis for formulating response strategies in advance.

[0147] As an embodiment of the present invention, the use of the preliminary health assessment report to perform a health time series analysis on the elderly user to obtain a health continuous analysis report includes: obtaining an initial time series of a target health indicator from the preliminary health assessment report, performing a stationarity test on the initial time series, and when a result of the stationarity test does not meet a preset requirement, performing a differential process on the initial time series to obtain a target time series, calculating an autocorrelation function and a partial autocorrelation function of the target time series to obtain an autocorrelation calculation result, and based on the autocorrelation calculation result, performing a health time series analysis on the elderly user to obtain a health continuous analysis report.

[0148] The autocorrelation function refers to a tool for measuring the correlation between a time series and its own lagged value, and the partial autocorrelation function refers to an indicator for measuring the correlation between a time series and its lagged value after controlling the influence of the intermediate lagged value of the time series.

[0149] Optionally, the initial time series can extract data of target health indicators such as blood pressure and blood sugar at different time points from the preliminary health assessment report, and organize them into a sequence arranged in chronological order, which can be achieved with the help of database query statements or data processing tools (such as Python's Pandas library); the stationarity test of the initial time series can adopt a unit root test (such as an ADF test) method, and use statistical analysis software (such as R language or Python's statsmodels library) to determine whether the sequence is stationary, and determine the result based on the comparison of the test statistic with the critical value, the differential processing of the initial time series can use a first-order difference method, and the calculation of the autocorrelation function and partial autocorrelation function of the target time series can be calculated using statistical software such as R language. Based on the autocorrelation calculation results, the health time series analysis of the elderly care user is performed to obtain a continuous health analysis report. According to the calculation results of the autocorrelation function and the partial autocorrelation function, a suitable time series model (such as ARIMA) can be selected, and the model fitting and parameter estimation are performed using software to analyze the changing trend and seasonality of the health indicators, and the analysis results are organized into a continuous health analysis report.

[0150] It should be explained that in the stationarity test, the commonly used unit root test (such as the ADF test) will generate a statistic. If the value of this statistic is greater than the corresponding critical value, it means that "the result of the stationarity test does not meet the preset requirements." For example, in the ADF test, for a given significance level (such as 5%), there is a corresponding critical value. When the calculated ADF statistic is greater than this critical value, we cannot reject the null hypothesis, that is, the null hypothesis believes that the sequence has a unit root, that is, the sequence is non-stationary.

[0151] First, obtain the time series of the target health indicator (such as heart rate data) from the health continuous analysis report. Use a unit root test (such as an ADF test) to determine the stationarity of the data. If the data is non-stationary, it needs to be differentiated. For example, perform a first-order difference on the original heart rate data, calculate the difference between the heart rate values ​​at two adjacent time points, obtain a new time series, and test the stationarity of the differentiated series again until a stationary series is obtained. This is because the ARIMA model requires the data to be stationary, otherwise the model's estimation and prediction may be biased.

[0152] The embodiment of the present invention utilizes the health continuous analysis report to construct a user portrait of the elderly care user. Various health information, living habit characteristics, medical history and other data in the health continuous analysis report can be integrated to form a comprehensive and multi-dimensional elderly care user portrait, which can more comprehensively reflect the individual characteristics of the elderly care user.

[0153] Among them, the user portrait refers to a comprehensive description of the characteristics of a specific user group or individual. It collects, integrates and analyzes multi-dimensional data information, labels the user with various "labels", thereby outlining the user's overall picture, so as to better understand the data on user needs, behavioral habits, preferences and other characteristics.

[0154] Optionally, the process of using the health continuous analysis report to construct the user portrait of the elderly care user is as follows: extract key features from the health continuous analysis report manually or with the help of natural language processing technology (such as NLTK library), including health indicators (such as types of chronic diseases, indicator change trends), living habits (such as exercise frequency, whether dietary preferences are mentioned) and other information, and define labels based on the extracted features, such as "hypertension tendency" and "active sports type". According to the degree of impact of the features on health or the focus of business attention, weights are assigned to labels through expert experience or data analysis algorithms (such as hierarchical analysis method), and a label-based user portrait model is adopted to integrate labels and their weights into database records or JSON format data structures, and use database management systems (such as MySQL) or NoSQL databases (such as MongoDB) to store user portrait information.

[0155] Furthermore, the embodiments of the present invention can provide intelligent health warnings for the elderly care users by utilizing the user portraits. The health status of the elderly care users can be monitored in real time and risks can be identified by combining real-time health data collection and intelligent algorithms based on the user portraits of the elderly care users.

[0156] Optionally, the process of using the user portrait to provide intelligent health warning for the elderly user is as follows: based on the features in the user portrait, such as disease risk labels, health indicator ranges, etc., personalized warning rules are set in combination with medical expertise. For example, for users with "high risk of cardiovascular disease" and "large blood pressure fluctuations in recent times", a warning is triggered if the blood pressure exceeds a specific threshold for a certain period of time.

[0157] The embodiment of the present invention constructs the smart elderly care health management platform for the elderly care users through the warning results based on the health intelligent warning, and can feed back the warning results of the health intelligent warning to each link of the smart elderly care health management platform, forming a complete health management closed loop. From data collection, preliminary health assessment, time series analysis, user portrait construction, health intelligent warning to warning result processing, each link is interrelated and interacts with each other, realizing data-driven continuous health management and platform optimization.

[0158] Among them, the smart elderly care health management platform refers to an intelligent platform for storing and analyzing data related to smart elderly care users.

[0159] As an embodiment of the present invention, the construction of the smart elderly care health management platform for the elderly users based on the warning results of the health intelligent warning includes: constructing an initial platform of the smart elderly care health management platform for the elderly users, analyzing the generation steps corresponding to the warning results of the health intelligent warning, and based on the generation steps, constructing the data processing port of the health management platform corresponding to the elderly users, using the health intelligent warning data of the elderly users to perform port debugging on the data processing port, and when the port debugging result is normal, obtaining the target port, configuring the data processing port to the initial platform, and configuring a user management interface for the initial platform to obtain the smart elderly care health management platform for the elderly users.

[0160] Optionally, the initial platform can be constructed using a mainstream software development architecture (such as a microservice architecture), and the generation steps can be obtained by reviewing the various environments of data processing in the health intelligent early warning system. The data processing port can be analyzed according to the early warning result generation steps to determine the type of data that the port needs to receive, such as a data acquisition port, a data analysis port, a data storage port, and an early warning port. The data processing port is configured in the initial platform, and a user management interface is configured for the initial platform. The smart elderly care health management platform for the elderly users can be obtained by utilizing front-end development technologies (such as HTML, CSS, JavaScript) to achieve interface layout and interactive effects, ultimately forming a complete smart elderly care health management platform.

[0161] Embodiment 2:

[0162] like Figure 2 As shown, it is a system function module diagram of a method for constructing a smart elderly care health management platform based on the Internet of Things provided by one embodiment of the present invention.

[0163] The system 200 for constructing a smart elderly care health management platform based on the Internet of Things of the present invention can be installed in an electronic device. According to the functions to be implemented, the system for constructing a smart elderly care health management platform based on the Internet of Things can include a data structure classification module 201, a data processing module 202, a data encryption module 203, and a platform construction module 204. The module of the present invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0164] In the embodiment of the present invention, the functions of each module / unit are as follows:

[0165] The data structure classification module 201 is used to collect the physical examination data and living environment data of the elderly care user in the smart elderly care environment by using the pre-configured Internet of Things device, extract the key data in the physical examination data and the living environment data, obtain the key data of the smart elderly care, construct a structural similarity evaluation model of the key data of the smart elderly care, and use the structural similarity evaluation model to perform data structure classification on the key data of the smart elderly care to obtain structured data and unstructured data;

[0166] The data processing module 202 is used to analyze the association relationship of the data elements in the structured data, use the association relationship to construct a relational data table of the structured data, organize the structured data to obtain organized data, construct an autoencoder for the unstructured data, use the autoencoder to deconstruct the unstructured data to obtain initial processed data, and map the initial processed data to a latent space to obtain target data;

[0167] The data encryption module 203 is used to store the sorted data and the target data to obtain stored data, identify the differential identification attribute of the stored data, set permissions for the stored data based on the differential identification attribute and a preset permission mapping rule to obtain permission data, and encrypt the permission data to obtain encrypted data;

[0168] The platform construction module 204 is used to use the encrypted data to perform a preliminary health assessment on the elderly user to obtain a preliminary health assessment report, use the preliminary health assessment report to perform a time series analysis of the health of the elderly user to obtain a continuous health analysis report, use the continuous health analysis report to build a user portrait of the elderly user, use the user portrait to perform intelligent health warning for the elderly user, and based on the warning results of the intelligent health warning, build a smart elderly health management platform for the elderly user.

[0169] In detail, each module in the system 200 for constructing a smart elderly care health management platform based on the Internet of Things in the embodiment of the present invention is used in the same manner as described above. Figure 1 The same technical means are used as the method for building a smart elderly care health management platform based on the Internet of Things, and can produce the same technical effects, so I will not go into details here.

[0170] In the several embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only illustrative, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0171] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0172] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0173] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0174] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0175] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for constructing a smart elderly care health management platform based on the Internet of Things, characterized in that: The method comprises: Using a pre-configured Internet of Things device to collect physical examination data and living environment data corresponding to elderly care users in a smart elderly care environment, extracting key data from the physical examination data and the living environment data to obtain smart elderly care key data, constructing a structural similarity evaluation model for the smart elderly care key data, and using the structural similarity evaluation model to perform data structure classification on the smart elderly care key data to obtain structured data and unstructured data; Analyze the association relationship of data elements in the structured data, use the association relationship to construct a relational data table of the structured data, organize the structured data to obtain organized data, construct an autoencoder for the unstructured data, use the autoencoder to deconstruct the unstructured data to obtain initial processed data, map the initial processed data to a latent space, and obtain target data; The sorted data and the target data are stored as data to obtain stored data, a differential identification attribute of the stored data is identified, permissions are set for the stored data based on the differential identification attribute and a preset permission mapping rule to obtain permission data, and the permission data is encrypted to obtain encrypted data; The encrypted data is used to conduct a preliminary health assessment of the elderly care user to obtain a preliminary health assessment report; the preliminary health assessment report is used to conduct a time series analysis of the health of the elderly care user to obtain a continuous health analysis report; the continuous health analysis report is used to construct a user portrait of the elderly care user; the user portrait is used to conduct an intelligent health warning for the elderly care user; based on the warning results of the intelligent health warning, a smart elderly care health management platform for the elderly care user is constructed.

2. The method for constructing a smart elderly care health management platform based on the Internet of Things according to claim 1, characterized in that: The step of extracting key data from the physical examination data and the living environment data to obtain key data for smart elderly care includes: Querying the physiological data and neural data in the body detection data and the gas detection data and temperature and humidity data of the living environment data; Based on the physiological data and neural data, construct a physiological and pathological model of the elderly user; Constructing a quantitative model of environmental health risks of the elderly user according to the gas detection data and the temperature and humidity data; Using the physiological and pathological model to analyze the trend of physical data of the elderly user, and obtain physical data analysis features; Using the environmental health risk quantification model, analyzing the environmental data changes of the elderly care users to obtain environmental change analysis features; Constructing a joint feature vector of the body data analysis feature and the environment change analysis feature; Feature screening is performed on the joint feature vector to obtain key data for smart elderly care.

3. The method for constructing a smart elderly care health management platform based on the Internet of Things according to claim 1, characterized in that: The constructing of the structural similarity evaluation model of the key data of smart elderly care includes: Performing hash coding on the key data of smart elderly care to obtain a data code; identifying a Hamming distance metric value encoding the data; Constructing a coding matrix for encoding the data; Based on the Hamming distance metric and the coding matrix, a structural similarity evaluation model of the key data of smart elderly care is constructed; Wherein, the structural similarity evaluation model is in the following form: Among them, φ represents the structural similarity evaluation model, n represents the amount of key data of smart elderly care, α represents the model search coefficient, represents the Hamming distance metric, Represents the i-th structure data in the encoded data, represents the i-th unstructured data in the encoded data, M represents the number of rows in the encoding matrix, L represents the number of columns in the encoding matrix, β represents the weight coefficient of the encoding parameter corresponding to the data encoding, R(x j,k ) represents the hash loss function, x j,k Represents the data at the corresponding position in the data encoding (the data related to the j-th row and k-th column position in the encoding matrix).

4. The method for constructing a smart elderly care health management platform based on the Internet of Things according to claim 1, characterized in that: The analyzing the association relationship of the data elements in the structure data includes: Classifying the data elements in the structure data to obtain classified elements; The element frequency of the classification element is calculated using the following formula: Among them, δ represents the frequency of elements, T iE represents the number of times element i appears in the Eth type element, T eE Indicates the number of times element e appears in the Eth type of elements; Based on the frequency of the elements, construct a frequent element combination; Calculating the support and confidence of the frequent element combination; Based on the support and the confidence, an association relationship between data elements in the structure data is determined.

5. The method for constructing a smart elderly care health management platform based on the Internet of Things according to claim 1, characterized in that: Mapping the initial processed data into a latent space to obtain target data includes: Performing data normalization processing on the initial processed data to obtain normalized data; extracting data points of the normalized data; The high-dimensional similarity probability of different data points in the data points is calculated using the following formula: Among them, P represents the similarity probability, Y a represents data point i, Y b represents data point j, σ represents the similarity range control parameter, ∥Y a -Y b ∥ means Y a With Y b The Euclidean distance, H and F represent the indexes of the data points, and their values ​​range from 1 to the total number of data points n (assuming there are n data points in total), ∑ H≠F exp(-∥Y H -Y F ∥ 2 / 2σ 2 ) represents the normalization factor, Y H Represents the data point pair H, Y obtained by indexing the data point F Represents the data point pair F (H≠F) obtained by indexing the data point; Mapping the normalized data into a latent space to obtain mapped data; Constructing low-dimensional initialization data points of the mapping data; Calculating the low-dimensional similarity probability of the low-dimensional initialization data points; Based on the high-dimensional similarity probability and the low-dimensional similarity probability, the position of the low-dimensional initialization data point is updated to obtain target data.

6. The method for constructing a smart elderly care health management platform based on the Internet of Things according to claim 1, characterized in that: The step of setting permissions on the stored data based on the differentiated identification attribute and a preset permission mapping rule to obtain permission data includes: identifying data sensitivity of the stored data; Based on the data sensitivity, user definition is performed on the access role corresponding to the stored data to obtain a defined user; Associating the differentiated identification attribute with the defined user to obtain data-role information; Based on the data-role information and the preset permission mapping rules, an attribute-permission mapping relationship table of the stored data is constructed to set permissions for the stored data, thereby obtaining permission data.

7. The method for constructing a smart elderly care health management platform based on the Internet of Things according to claim 1, characterized in that: The step of encrypting the authority data to obtain encrypted data includes: Identify a data node of the permission data; Configuring a dynamic key for the permission data; The dynamic key can be expressed by the following formula: B=(J1+r1)×(J2+r2)×…×(J n +r n ) Where B represents the dynamic key, J1 represents the first node in the data node, r1 represents the random number of J1, J2 represents the second node in the data node, r2 represents the random number of J2, and J n Represents the nth node in the data node, r n Indicates J n A random number; According to the data node, the dynamic key is dynamically updated using the following formula to obtain an updated key: B ′ =G(r ′ ,B) Among them, B ′ represents the update key, G represents the update function, r ′ represents the update of random number, and B represents the dynamic key; The authority data is encrypted using the update key to obtain encrypted data.

8. The method for constructing a smart elderly care health management platform based on the Internet of Things according to claim 1, characterized in that: The preliminary health assessment report is used to perform a time series analysis of the health of the elderly user to obtain a continuous health analysis report, including: Obtaining an initial time series of a target health indicator from the preliminary health assessment report; Performing a stationarity test on the initial time series; When the result of the stationarity test does not meet the preset requirements, performing difference processing on the initial time series to obtain a target time series; Calculating the autocorrelation function and the partial autocorrelation function of the target time series to obtain an autocorrelation calculation result; Based on the autocorrelation calculation result, a time series analysis of the health of the elderly user is performed to obtain a continuous health analysis report.

9. The method for constructing a smart elderly care health management platform based on the Internet of Things according to claim 1, characterized in that: The method of constructing a smart elderly care health management platform for elderly care users based on the warning result of the health intelligent warning includes: Constructing the initial platform of the smart elderly care health management platform for the elderly care users; Analyze the generation steps corresponding to the warning results of the health intelligent warning; Based on the generating step, constructing a data processing port of the health management platform corresponding to the elderly care user; Using the intelligent health warning data of the elderly care user, port debugging is performed on the data processing port; When the port debugging result is normal, obtaining the target port; The data processing port is configured in the initial platform, and a user management interface is configured for the initial platform to obtain a smart elderly care health management platform for the elderly care users.

10. A system for building a smart elderly care health management platform based on the Internet of Things, characterized in that: The system comprises: A data structure classification module is used to collect physical examination data and living environment data corresponding to elderly care users in a smart elderly care environment by using a pre-configured Internet of Things device, extract key data from the physical examination data and the living environment data, obtain key data for smart elderly care, construct a structural similarity evaluation model for the key data for smart elderly care, and use the structural similarity evaluation model to perform data structure classification on the key data for smart elderly care to obtain structured data and unstructured data; A data processing module is used to analyze the association relationship of data elements in the structured data, construct a relational data table of the structured data by using the association relationship, organize the structured data to obtain organized data, construct an autoencoder for the unstructured data, deconstruct the unstructured data by using the autoencoder to obtain initial processed data, and map the initial processed data to a latent space to obtain target data; A data encryption module, used for storing the sorted data and the target data to obtain stored data, identifying a differential identification attribute of the stored data, setting permissions for the stored data based on the differential identification attribute and a preset permission mapping rule to obtain permission data, and encrypting the permission data to obtain encrypted data; The platform construction module is used to use the encrypted data to perform a preliminary health assessment on the elderly user to obtain a preliminary health assessment report, use the preliminary health assessment report to perform a time series analysis of the health of the elderly user to obtain a continuous health analysis report, use the continuous health analysis report to build a user portrait of the elderly user, use the user portrait to perform intelligent health warning for the elderly user, and based on the warning results of the intelligent health warning, build a smart elderly health management platform for the elderly user.

Images