Intelligent old-age care system based on big data integration analysis

Abstract

The invention provides an intelligent old-age care system based on big data integration analysis, and relates to the technical field of computer information processing, multi-modal old-age care data is fused based on a distributed architecture, and an old-age care resource data set is obtained; performing adaptive partitioning on the pension resource data set to obtain a pension service demand vector and a service resource vector, and constructing an embedded model library of pension semantics through semantic matching and space-time semantic annotation operation; according to the embedded model library, carrying out data stream response based on the pension service information to obtain pension response data, and carrying out resource scheduling reasoning on the pension response data based on the pension service knowledge graph to obtain a scheduling optimization matrix; performing service efficiency modeling on the pension demand data to obtain a service evaluation factor set; and performing multi-dimensional constraint optimization on the service evaluation factor set based on the scheduling optimization matrix to obtain a matching optimization result, so that intelligent matching and optimal scheduling of the old-age service requirements can be realized, and the accuracy of service response is improved.

Description

Technical Field

[0001] The present application relates to the field of computer information processing technology, and more specifically, to a smart elderly care system based on big data integrated analysis. Background Art

[0002] With the continuous development of computer information processing technology, data-driven intelligent service systems have been widely used in various social service fields. Especially driven by technologies such as big data, artificial intelligence, and the Internet of Things, computer systems can collect, integrate, process, semantically understand, and automatically reason about structured and unstructured data in real time, providing a data foundation and intelligent engine for various decision support systems.

[0003] Smart elderly care, an emerging field integrating information technology with elderly care services, aims to enable dynamic health monitoring, service response, and resource scheduling for the elderly population through integrated analysis of multi-source data, including elderly health data, service needs, and medical resources. Some existing systems have demonstrated initial capabilities for big data-based elderly care services, primarily consisting of modules such as data collection terminals, service platforms, and rule engines. However, due to the complex and heterogeneous structure of multimodal elderly care data, traditional integration approaches lack hierarchical management and distributed processing capabilities, resulting in poor data consistency and low call efficiency. Furthermore, matching service needs with resources primarily relies on keywords, rules, or manual intervention, lacking intelligent matching mechanisms based on semantic understanding and contextual awareness. This results in inaccurate resource scheduling, delayed responses, and mismatched matching. Furthermore, existing smart elderly care systems lack knowledge reasoning and scheduling optimization capabilities for dynamically changing environments, making them difficult to support personalized, real-time elderly care service demands. Therefore, achieving intelligent matching and optimized scheduling of elderly care service needs to improve the accuracy of service responses remains a major challenge facing the industry. Summary of the Invention

[0004] This application provides a smart elderly care system based on big data integrated analysis, which can realize intelligent matching and optimized scheduling of elderly care service needs to improve the accuracy of service response.

[0005] This application provides a smart elderly care system based on big data integrated analysis, which includes: The data fusion module is used to fuse multimodal elderly care data based on a distributed architecture to obtain an elderly care resource dataset; A semantic matching module is used to adaptively partition the elderly care resource dataset to obtain elderly care service demand vectors and service resource vectors, perform semantic matching and spatiotemporal semantic annotation operations on both the elderly care service demand vectors and the service resource vectors, and obtain an embedded model library for elderly care semantics; A scheduling optimization module is used to respond to data streams based on the elderly care service information according to the embedded model library to obtain elderly care response data, and perform resource scheduling reasoning on the elderly care response data based on the elderly care service knowledge graph to obtain a scheduling optimization matrix; A service effectiveness module is used to obtain elderly care demand data, perform service effectiveness modeling on the elderly care demand data, and obtain a set of service evaluation factors; The matching optimization module is used to perform multi-dimensional constraint optimization on the service evaluation factor set based on the scheduling optimization matrix to obtain a matching optimization result for elderly care needs.

[0006] In this embodiment, the multimodal elderly care data includes: elderly care service demand data, service resource data and health record data.

[0007] In this embodiment, the distributed architecture includes a central node and an edge computing node. The edge computing node is used to collect and pre-process multimodal elderly care data, and the central node is used to fuse the multimodal elderly care data.

[0008] In this embodiment, adaptively partitioning the elderly care resource dataset to obtain the elderly care service demand vector and the service resource vector specifically includes: Clustering the pension resource dataset to obtain multiple partition resource sets; Based on semantic similarity, similarity analysis is performed on each partition resource set to obtain semantic resource partitions; Performing embedding vector encoding on the elderly care service demand data of the semantic resource partition to obtain an elderly care service demand vector; Semantic feature extraction and embedding conversion are performed on the service resource data of the semantic resource partition to obtain a service resource vector.

[0009] In this embodiment, semantic matching and spatiotemporal semantic annotation operations are performed on both the elderly care service demand vector and the service resource vector to obtain an embedded model library of elderly care semantics, specifically including: Mapping the elderly care service demand vector and the service resource vector into a semantic vector space, and then determining semantic similarity based on the semantic vector space; In the semantic vector space, adding spatiotemporal labels to the elderly care service demand vector and the service resource vector respectively, thereby obtaining spatiotemporal classification conditions; An embedding model library of elderly care semantics is constructed based on the semantic similarity and the spatiotemporal classification conditions.

[0010] In this embodiment, the data stream response is performed based on the elderly care service information according to the embedded model library, and the elderly care response data obtained specifically includes: Performing embedding vector encoding on the elderly care service information to obtain a service semantic vector library; Semantic matching is performed on the embedding model library and the service semantic vector library, and then the elderly care response data is obtained by screening according to the matching results.

[0011] In this embodiment, the elderly care service knowledge graph is a heterogeneous knowledge graph.

[0012] In this embodiment, resource scheduling reasoning is performed on the elderly care response data based on the elderly care service knowledge graph to obtain a scheduling optimization matrix, which specifically includes: Scheduling candidates for the elderly care response data based on the entity relationships in the elderly care knowledge graph to obtain a candidate set of service resources; Sorting and scoring the service resource candidate set to obtain a scoring result; The scoring results are inferred by the inference engine in the elderly care knowledge graph to obtain a resource scheduling set, which is then scored and optimized and converted into a scheduling optimization matrix.

[0013] In this embodiment, elderly care demand data is obtained through an Internet of Things terminal.

[0014] In this embodiment, the service effectiveness model is performed on the elderly care demand data to obtain a service evaluation factor set including: Obtain service indicator data for elderly care needs; Constructing a multi-factor weighted model based on the service index data, and then performing an efficiency analysis on the elderly care demand data based on the multi-factor weighted model to obtain service efficiency; A service evaluation is performed on the elderly care demand data based on the service effectiveness, thereby obtaining a service evaluation factor set.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: Based on a distributed architecture, multimodal elderly care data is integrated to obtain an elderly care resource dataset; the elderly care resource dataset is adaptively partitioned to obtain an elderly care service demand vector and a service resource vector, semantic matching and spatiotemporal semantic labeling operations are performed on the elderly care service demand vector and the service resource vector to obtain an embedded model library of elderly care semantics; data stream responses are performed based on elderly care service information according to the embedded model library to obtain elderly care response data, resource scheduling reasoning is performed on the elderly care response data based on the elderly care service knowledge graph to obtain a scheduling optimization matrix; elderly care demand data is acquired, service efficiency modeling is performed on the elderly care demand data to obtain a service evaluation factor set; multi-dimensional constraint optimization is performed on the service evaluation factor set based on the scheduling optimization matrix to obtain a matching optimization result for elderly care demand.

[0016] It can be seen that in this application, intelligent matching and optimized scheduling of elderly care service needs can be achieved. First, by fusing multimodal elderly care data based on a distributed architecture, efficient integration of heterogeneous elderly care data can be achieved, and by adopting a hierarchical data storage structure, the problems of complex data structure and low access efficiency are effectively solved, providing a unified and structured elderly care resource data set for subsequent semantic modeling and resource scheduling; secondly, by fusing semantic embedding and spatiotemporal semantic features through the semantic matching module, semantic understanding, spatiotemporal correlation modeling and feature alignment between service supply and demand can be achieved, effectively improving the semantic accuracy and context perception ability of service matching; based on the constructed elderly care service knowledge graph, resource scheduling reasoning is further performed on the response data to obtain multi-objective scheduling The degree optimization matrix can combine the embedding model and knowledge graph reasoning to realize service matching and resource recommendation from the semantic layer to the logical layer, which is conducive to enhancing the scheduling intelligence and real-time performance of the system; then, the service indicator data related to elderly care needs are obtained, and a service performance evaluation model is constructed to generate a service evaluation factor set. By introducing multi-factor modeling, the historical performance and resource adaptability of the service can be quantitatively evaluated from the response results, which is convenient for providing a quantitative basis for intelligent early warning and service optimization; finally, the matching optimization results are determined based on the scheduling optimization matrix and service evaluation factor set. By integrating multiple dimensions such as service capabilities, time delay response, and historical evaluation, the system's intelligent adaptation capabilities to different service needs are improved, and refined recommendations and high-accuracy scheduling of elderly care services are achieved.

[0017] To sum up, the technical solution adopted in this application can realize intelligent matching and optimized scheduling of elderly care service needs to improve the accuracy of service response. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0019] Figure 1 This is a module structure diagram of a smart elderly care system based on big data integrated analysis provided by this application; Figure 2 is a schematic diagram of a process for determining elderly care service demand vectors and service resource vectors in some embodiments of the present application; Figure 3 This is an exemplary flowchart for determining a scheduling optimization matrix in some embodiments of the present application. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0021] The embodiment of the present application provides a smart elderly care system based on big data integrated analysis, the core of which is to fuse multimodal elderly care data based on a distributed architecture to obtain an elderly care resource data set; adaptively partition the elderly care resource data set to obtain an elderly care service demand vector and a service resource vector, perform semantic matching and spatiotemporal semantic labeling operations on both the elderly care service demand vector and the service resource vector to obtain an embedded model library of elderly care semantics; perform data stream response based on elderly care service information according to the embedded model library to obtain elderly care response data, perform resource scheduling reasoning on the elderly care response data based on the elderly care service knowledge graph to obtain a scheduling optimization matrix; obtain elderly care demand data, perform service efficiency modeling on the elderly care demand data to obtain a service evaluation factor set; perform multi-dimensional constraint optimization on the service evaluation factor set based on the scheduling optimization matrix to obtain a matching optimization result for elderly care demand.

[0022] In order to better understand the above technical solution, the following will be described in detail with reference to the accompanying drawings and specific implementation methods. Figure 1 As shown in FIG, this figure is a module structure diagram of a smart elderly care system based on big data integrated analysis provided by the present application. The smart elderly care system includes: a data fusion module 100, a semantic matching module 200, a scheduling optimization module 300, a service efficiency module 400 and a matching optimization module 500, which are described as follows: The data fusion module 100 is used to fuse multimodal elderly care data based on a distributed architecture to obtain an elderly care resource data set.

[0023] In specific implementation, multimodal elderly care data can be collected through the edge computing nodes of the distributed architecture. It should be noted that the multimodal elderly care data in this application include: elderly care service demand data, service resource data and health record data, wherein the elderly care service demand data includes the service requests, service preferences and historical service records of elderly users; the service resource data includes the basic information of the service agency, service types, service availability and service capabilities; the health record data includes the vital sign monitoring information, historical medical records, drug use records and daily behavior patterns of elderly users; in addition, the edge computing nodes can perform preprocessing operations on the elderly care service demand data, service resource data and health record data respectively. The preprocessing specifically includes data format standardization, noise filtering, missing value filling and preliminary semantic labeling, and the preprocessed data is associated and matched based on dimensions such as timestamp, user ID and data category, thereby generating a structured elderly care resource data set. The elderly care resource data set can represent semantically clear and temporally continuous elderly care resource information.

[0024] It should be noted that the distributed architecture in this application includes a central node and an edge computing node. The edge computing node is used to collect and pre-process multimodal elderly care data, and the central node is used to fuse multimodal elderly care data. The central node is a data processing unit with powerful computing power, storage capacity and resource scheduling capabilities. The edge computing node refers to a lightweight computing unit, including: elderly care service terminals, wearable devices and health monitoring equipment; the central node is integrated with a streaming processing engine for processing continuously accessed data streams, and adopts an event-driven synchronization mechanism to dynamically update the content of the elderly care resource data set, thereby improving the system's response capability to sudden demands and the timeliness of data fusion; preferably, the edge computing node and the central node can communicate through a secure encrypted channel to ensure the integrity and privacy of the data during transmission.

[0025] The semantic matching module 200 is used to adaptively partition the elderly care resource data set to obtain an elderly care service demand vector and a service resource vector, perform semantic matching and spatiotemporal semantic labeling operations on both the elderly care service demand vector and the service resource vector, and obtain an embedded model library of elderly care semantics.

[0026] Preferably, reference Figure 2 As shown in FIG, this figure is a schematic diagram of the process of determining the elderly care service demand vector and the service resource vector in some embodiments of the present application. In this embodiment, the elderly care resource dataset is adaptively partitioned to obtain the elderly care service demand vector and the service resource vector, which specifically includes the following steps: First, in step S21, the elderly care resource dataset is clustered to obtain multiple partition resource sets; Secondly, in step S22, similarity analysis is performed on each partition resource set based on the semantic embedding model to obtain semantic resource partitions; Then, in step S23, embedding vector encoding is performed on the elderly care service demand data of the semantic resource partition to obtain an elderly care service demand vector; Finally, in step S24, semantic feature extraction and embedding conversion are performed on the service resource data of the semantic resource partition to obtain a service resource vector.

[0027] In the specific implementation, first, the key feature fields in the elderly care resource dataset are standardized and preprocessed, for example: the numerical and categorical features such as user age, service time, and service level are uniformly encoded, and the processed data are clustered using the K-means algorithm, and then the multiple datasets obtained by clustering are combined into different partition resource sets, and the partition resource sets represent data clusters with high similarity in the elderly care resource dataset; secondly, the semantic embedding model based on deep learning is used to semantically vectorize the descriptive fields (such as service content description and user note information) of each partition resource set, and then the similarity between each partition resource set is calculated through the semantic embedding model, and the hierarchical merging algorithm is used to merge the partition sets with high similarity, and then the merged The data set obtained is used as a semantic resource partition; then, the core features of the elderly care service demand data in each semantic resource partition (such as the health status of the service recipient, the nursing level, the time period preference, etc.) are extracted, and the pre-trained contextual language model (ALBERT) is used to encode the core features of the elderly care service demand data into a fixed-length vector, so that the fixed-length vector is used as the elderly care service demand vector; finally, semantic features are extracted from the fields such as service type, service personnel capability, and service coverage area involved in the service resource data, and the extracted semantic features are mapped into structured semantic features through label embedding technology, and then the structured semantic features are mapped into a service resource vector through full-connection layer projection, wherein the dimensions of the service resource vector and the elderly care service demand vector are consistent.

[0028] It should be noted that the elderly care service demand vector represents the service demand characteristics of a specific user in a certain time and space background; the service resource vector describes the characteristics of the service resources in the semantic space; in addition, through the clustering and semantic analysis process, it is possible to achieve a dual division of elderly care data in terms of structural characteristics and semantic levels, thereby improving the accuracy of semantic partitioning; improving the contextual understanding ability of the embedded model can model the contextual semantic relationship of elderly care service needs, which is conducive to avoiding semantic deviations caused by data discreteness.

[0029] In this embodiment, semantic matching and spatiotemporal semantic annotation operations are performed on both the elderly care service demand vector and the service resource vector to obtain an embedded model library of elderly care semantics, specifically in the following manner: Mapping the elderly care service demand vector and the service resource vector into a semantic vector space, and then determining semantic similarity based on the semantic vector space; In the semantic vector space, adding spatiotemporal labels to the elderly care service demand vector and the service resource vector respectively, thereby obtaining spatiotemporal classification conditions; An embedding model library of elderly care semantics is constructed based on the semantic similarity and the spatiotemporal classification conditions.

[0030] In the specific implementation, first, a multi-layer deep semantic network is used to perform semantic alignment processing on the elderly care service demand vector and the service resource vector, and then mapped to the semantic vector space. The cosine similarity calculation formula is used to calculate the cosine similarity of all vectors in the semantic vector space, and then the cosine similarity is used as the semantic similarity between vectors in the semantic vector space, wherein the semantic similarity is an indicator to measure the similarity of the meanings of words or semantic representations, which facilitates the semantic alignment of elderly care resources; then, spatiotemporal labels are introduced into the semantic vector space to label the elderly care service demand vector and the service resource vector respectively, and a spatiotemporal classification condition with spatiotemporal context perception capability is obtained. The spatiotemporal labels include information such as service time period and service resources; finally, a similarity threshold is set through regression analysis, and vectors with semantic similarity greater than the similarity threshold are all taken as candidate vector pairs. Then, the time dimension and space dimension of the candidate vector pairs are screened based on the spatiotemporal classification condition, and the screened vector pairs are encoded into spatiotemporal vectors. The spatiotemporal vectors are embedded into the semantic vector space through the semantic embedding model, so that the semantic vector space after the spatiotemporal vector embedding is used as the embedding model library of elderly care semantics.

[0031] It should be noted that in this application, the embedded model library of elderly care semantics is a set of semantic structure models of service requirements and service resources represented in vector form, which is used to express, store and support efficient retrieval and matching based on semantics and spatiotemporal information, and has the ability to model semantic consistency and express multi-dimensional constraints; in this embodiment, the embedded model library calculates the semantic similarity of service demand vectors and resource vectors in a unified semantic vector space, and combines spatiotemporal semantic annotation with information such as service time and geographic location, which can realize cross-dimensional and cross-context service matching, effectively improving the intelligent response capability of the elderly care service system to complex demand scenarios; in addition, the semantic matching method based on vectorized modeling can reduce the ambiguity of semantic expression.

[0032] The scheduling optimization module 300 is used to respond to the data stream based on the elderly care service information according to the embedded model library to obtain elderly care response data, perform resource scheduling reasoning on the elderly care response data based on the elderly care service knowledge graph, and obtain a scheduling optimization matrix.

[0033] In this embodiment, the embedded model library is used to perform data stream response based on the elderly care service information to obtain elderly care response data in the following manner, namely: Performing embedding vector encoding on the elderly care service information to obtain a service semantic vector library; Semantic matching is performed on the embedding model library and the service semantic vector library, and then the elderly care response data is obtained by screening according to the matching results.

[0034] It should be noted that the elderly care service information in this application refers to a data set generated in real time or periodically by multiple sources in the elderly care service scenario, which is used to describe user needs and environmental status, including: physiological parameters collected by IoT terminals (such as wearable health monitoring devices, smart home sensors), service requests and preference settings submitted by mobile terminals, and health records pushed by public health systems. Among them, the elderly care service information can be uniformly accessed and pre-processed through the existing edge gateway combined with message bus technology, and then transmitted to the scheduling optimization module through the edge computing node.

[0035] In the specific implementation, first, the text fields and numerical fields of the elderly care service information are standardized based on regular expressions to obtain different text features. For example, structured fields (such as heart rate, temperature, and blood pressure) are directly normalized, and part-of-speech tagging technology is used to extract key phrases from text fields (service requests, preferences, and health records). Then, all text features are deeply semantically encoded through the pre-trained sentence embedding representation model (Sentence-BERT), and the multiple encoded results are used as service semantic vectors. The collection of service semantic vectors is used as a service semantic vector library. Then, through the vector The retrieval and similarity calculation technology performs semantic matching on the embedding model library and the service semantic vector library, and then filters the elderly care response data based on the matching results. Preferably, an efficient nearest neighbor search algorithm can be used to quickly retrieve the semantic vector in the embedding model library, wherein cosine similarity can be used as a matching metric, which is conducive to accurately reflecting the vector angle information in the high-dimensional vector space and maintaining calculation efficiency, and setting the similarity threshold through regression analysis, selecting the semantic vector above the similarity threshold, and then reversely inputting the selected semantic vector into the sentence embedding representation model, and obtaining the elderly care response data through the output of the sentence embedding representation model.

[0036] It should be noted that the elderly care response data in this application is service resource data matched from the embedded model library, which has high similarity and time-space coupling with user service needs in the semantic vector space. It can be used as input for subsequent resource scheduling reasoning to reduce response delay and improve resource utilization, which is conducive to improving the accuracy of scheduling decisions.

[0037] Preferably, reference Figure 3 As shown in FIG, this figure is a schematic diagram of the process of determining the scheduling optimization matrix in some embodiments of the present application. In this embodiment, resource scheduling reasoning is performed on the elderly care response data based on the elderly care service knowledge graph to obtain the scheduling optimization matrix, which specifically includes the following steps: First, in step S31, scheduling candidates are selected for the elderly care response data based on the entity relationships in the elderly care knowledge graph to obtain a candidate set of service resources; Then, in step S32, the service resource candidate set is sorted and scored to obtain a scoring result; Finally, in step S33, the scoring result is inferred by the inference engine in the elderly care knowledge graph to obtain a resource scheduling set, which is then scored and optimized and converted into a scheduling optimization matrix.

[0038] It should be noted that the elderly care service knowledge graph in this application is a heterogeneous knowledge graph, which is constructed by combining machine learning technology with multi-source heterogeneous elderly care service data fusion, including semantic relationships between different types of entities, including: service-provider, service-applicable object, institution-affiliated area, time-resource availability, etc.; the heterogeneous knowledge graph includes inference rules and inference engines, which can effectively represent the complex and multi-dimensional semantic associations in the elderly care service scenario. The heterogeneous knowledge graph performs resource scheduling reasoning, which is conducive to the scheduling optimization of elderly care service resources in terms of personalization, precision and dynamic availability, and then generates highly robust and highly matched elderly care service information.

[0039] In a specific implementation, first, key entity information is extracted from the elderly care response data, such as service type, demand time period, service urgency, user geographic location, etc., and entity nodes and their relationships in the elderly care service knowledge graph are used to construct entity association query paths. Semantic retrieval is performed on the knowledge graph based on graph query statements (such as SPARQL) to screen out resource entities that meet entity constraints and obtain a candidate set of service resources. Then, a multidimensional scoring index is constructed based on factors such as the matching degree between the candidate set of service resources and the elderly care response data, historical service records, current resource availability, and response timeliness. That is, the candidate resources are scored based on a machine learning model, and the scoring results are sorted by score, and the sorted results are used as the scoring results. Preferably, the machine learning model can use the extreme gradient boosting model (XGBoost), which is conducive to sorting in combination with the scoring results. Finally, the scoring results are logically reasoned using the inference rules and inference engine in the knowledge graph to screen out the optimal resource allocation plan, forming a resource scheduling set. The resource arrangement in the resource scheduling set is optimized for minimum delay through a reinforcement learning model, and the optimized resource scheduling results are used as a scheduling optimization matrix.

[0040] It should be noted that the scheduling optimization matrix in this application is the actual demand condition in the elderly care service scenario, and is the optimal resource configuration result calculated in combination with multi-dimensional factors such as resource availability, time and space constraints, and service priority. Each element in the scheduling optimization matrix represents the assignment relationship, assignment time, scheduling priority, etc. between service resources and demand entities, which can be used as the core basis for subsequent scheduling. The rows of the scheduling optimization matrix can represent service demand entities, and the columns represent optional resource units, which facilitates the realization of intelligent and efficient elderly care resource scheduling, is conducive to the realization of refined management of multi-source service resources, and improves the accuracy of scheduling decisions.

[0041] The service effectiveness module 400 is used to obtain elderly care demand data, perform service effectiveness modeling on the elderly care demand data, and obtain a service evaluation factor set.

[0042] In this embodiment, the service effectiveness modeling is performed on the elderly care demand data to obtain a service evaluation factor set in the following manner, namely: Obtain service indicator data for elderly care needs; Constructing a multi-factor weighted model based on the service index data, and then performing an efficiency analysis on the elderly care demand data based on the multi-factor weighted model to obtain service efficiency; A service evaluation is performed on the elderly care demand data based on the service effectiveness, thereby obtaining a service evaluation factor set.

[0043] It should be noted that in this application, elderly care demand data is obtained through the Internet of Things terminal. Elderly care demand data refers to a multi-dimensional dynamic data set that can reflect the health status, service needs and behavioral characteristics of the service recipients within a specific time period, including: static information and dynamic information. The static information includes: age, medical history, living habits, etc., and the dynamic information includes physiological parameters, behavioral trajectories, service feedback, etc. The elderly care demand data can provide an objective data basis for the optimization of elderly care service management; among them, the Internet of Things terminal includes: smart wearable devices, environmental sensors, health monitoring equipment, positioning trackers, etc., and the elderly care demand data can be collected in real time through the Internet of Things terminal.

[0044] In specific implementation, first, the key indicator information of the service process can be collected in real time through the Internet of Things terminals deployed on the user side or in the elderly care service facilities. The key indicator information includes: service response time, service completion time, service frequency, changes in user physiological state, user satisfaction feedback, etc., and then the collected key indicator information is used as service indicator data; then, weights can be set according to the business importance of each service indicator, and a multi-factor weighted model can be constructed based on the set weights using a linear weighted method. The service indicator data is aggregated through the multi-factor weighted model, and the aggregation results are used as service performance scores; finally, a classification evaluation is performed based on the service performance score through the multi-factor weighted model, and service evaluation factors are extracted, and then the set of service evaluation factors is used as a service evaluation factor set, wherein the service evaluation factors refer to elderly care service data used for service optimization and quality tracking.

[0045] It should be noted that the service evaluation factor set in this application is a set of structured indicators that reflect key dimensions such as service response quality, user satisfaction and resource utilization efficiency. The service evaluation factors in the service evaluation factor set can be used to quantitatively analyze service performance, providing data support and evaluation basis for subsequent service recommendation optimization and system iteration and upgrade.

[0046] The matching optimization module 500 is used to perform multi-dimensional constraint optimization on the service evaluation factor set based on the scheduling optimization matrix to obtain a matching optimization result for elderly care needs.

[0047] In this embodiment, the following method is used to perform multi-dimensional constraint optimization on the service evaluation factor set based on the scheduling optimization matrix to obtain a matching optimization result for elderly care needs, namely: Fusing the service evaluation factor set and the service resource feature vector in the scheduling optimization matrix to obtain an evaluation vector set; Constructing a multi-objective optimization model, using the evaluation vector set as input to the multi-objective optimization model, and setting a service response delay threshold, a path distance weight, a service expertise weight, and a historical score threshold through the multi-objective optimization model; Based on a preset weighted scoring mechanism, the matching optimization result of the elderly care needs is determined based on the service response delay threshold, the path distance weight, the service professionalism weight and the historical scoring threshold.

[0048] In a specific implementation, first, the evaluation items in the service evaluation factor set are standardized, and the characteristic information of the service resources in the scheduling optimization matrix is ​​extracted. The two types of vectors are integrated through splicing, mapping, or weighted fusion to form an evaluation vector set. Then, multiple optimization goals are set, such as shortening service response time, reducing resource path distance, improving service professional matching, and improving historical service satisfaction. According to the actual needs of the elderly care service scenario, corresponding optimization weights and constraints are set, such as maximum acceptable response delay, minimum path restriction, service priority standard, etc., and a particle swarm algorithm is used to perform a global search on the evaluation vector set, so that the particle swarm algorithm outputs a service response delay threshold, path distance weight, service professional weight, and historical score threshold. Finally, the preset weighted scoring mechanism is to perform a multi-dimensional scoring on each candidate service resource based on the service response delay threshold, path distance weight, service professional weight, and historical score threshold output by the particle swarm algorithm. Through this weighted scoring mechanism, resources that meet the preset response delay, professional adaptability, and historical score requirements are assigned higher weight scores through a weighted accumulation method. Then, all resource candidates are ranked, thereby outputting a matching optimization result for elderly care needs.

[0049] It should be noted that the matching optimization result in this application refers to the result obtained by optimally matching service resources and elderly care needs while satisfying multiple service scheduling conditions, which can improve the intelligence level of the smart elderly care system; in addition, in this embodiment, the flexibility and scalability of the multi-objective optimization model can be used to customize strategies according to the characteristics of elderly care services in different regions and for different populations.

[0050] It can be seen that in the present application, the intelligent matching and optimized scheduling of pension service demand can be realized. First, based on the distributed architecture, the multi-modal pension data is fused, the efficient integration of heterogeneous pension data can be realized, and through the adoption of a hierarchical data storage structure, the problems of complex data structure and low access efficiency are effectively solved, providing a unified and structured pension resource dataset for subsequent semantic modeling and resource scheduling. Second, through the semantic matching module, the semantic embedding and spatio-temporal semantic features are fused, the semantic understanding between service supply and demand, spatio-temporal correlation modeling and feature alignment can be realized, and the semantic accuracy and context awareness of service matching are effectively improved. Third, based on the constructed pension service knowledge graph, resource scheduling reasoning is further performed on the response data, and a multi-objective scheduling optimization matrix is obtained. The embedding model and knowledge graph reasoning are combined to realize service matching and resource recommendation from the semantic layer to the logic layer, which is conducive to enhancing the scheduling intelligence and real-time performance of the system. Fourth, the service index data related to the pension demand is obtained, and a service efficiency evaluation model is constructed to generate a service evaluation factor set. Through the introduction of multi-factor modeling, the historical efficiency and resource adaptation degree of the service can be quantitatively evaluated from the response results, which provides a quantitative basis for intelligent early warning and service optimization. Finally, based on the scheduling optimization matrix and the service evaluation factor set, the matching optimization result is determined, and through the fusion of service capability, time delay response, historical evaluation and other multiple dimensions, the intelligent adaptation ability of the system to different service demands is improved, realizing the fine recommendation and high accuracy scheduling of pension services.

[0051] In summary, the technical scheme adopted in the present application can realize intelligent matching and optimized scheduling of pension service demand to improve the accuracy of service response.

[0052] The present application is described with reference to flowcharts and / or block diagrams according to the methods, devices (systems) and computer program products of the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of the flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The device that implements the functions specified in one flow or multiple flows and / or blocks. Figure 1 The device that implements the functions specified in one flow or multiple flows and / or blocks.

[0053] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program. The program can be stored in a computer-readable storage medium, including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disc storage, magnetic disk storage, or magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0054] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

Claims

1. A smart elderly care system based on big data integrated analysis, characterized by: The smart elderly care system includes: The data fusion module is used to fuse multimodal elderly care data based on a distributed architecture to obtain an elderly care resource dataset; A semantic matching module is used to adaptively partition the elderly care resource dataset to obtain elderly care service demand vectors and service resource vectors, perform semantic matching and spatiotemporal semantic annotation operations on both the elderly care service demand vectors and the service resource vectors, and obtain an embedded model library for elderly care semantics; A scheduling optimization module is used to respond to data streams based on the elderly care service information according to the embedded model library to obtain elderly care response data, and perform resource scheduling reasoning on the elderly care response data based on the elderly care service knowledge graph to obtain a scheduling optimization matrix; A service effectiveness module is used to obtain elderly care demand data, perform service effectiveness modeling on the elderly care demand data, and obtain a set of service evaluation factors; The matching optimization module is used to perform multi-dimensional constraint optimization on the service evaluation factor set based on the scheduling optimization matrix to obtain a matching optimization result for elderly care needs.

2. The smart elderly care system based on big data integrated analysis as claimed in claim 1, characterized in that: The multimodal elderly care data includes: elderly care service demand data, service resource data and health record data.

3. The smart elderly care system based on big data integrated analysis as claimed in claim 1, characterized in that: The distributed architecture includes a central node and an edge computing node. The edge computing node is used to collect and pre-process multimodal elderly care data, and the central node is used to fuse the multimodal elderly care data.

4. The smart elderly care system based on big data integrated analysis as claimed in claim 1, characterized in that: Adaptively partitioning the elderly care resource dataset to obtain elderly care service demand vectors and service resource vectors specifically includes: Clustering the pension resource dataset to obtain multiple partition resource sets; Based on semantic similarity, similarity analysis is performed on each partition resource set to obtain semantic resource partitions; Performing embedding vector encoding on the elderly care service demand data of the semantic resource partition to obtain an elderly care service demand vector; Semantic feature extraction and embedding conversion are performed on the service resource data of the semantic resource partition to obtain a service resource vector.

5. The smart elderly care system based on big data integrated analysis as claimed in claim 1, characterized in that: The semantic matching and spatiotemporal semantic labeling operations are performed on both the elderly care service demand vector and the service resource vector to obtain an embedded model library of elderly care semantics, specifically including: Mapping the elderly care service demand vector and the service resource vector into a semantic vector space, and then determining semantic similarity based on the semantic vector space; In the semantic vector space, adding spatiotemporal labels to the elderly care service demand vector and the service resource vector respectively, thereby obtaining spatiotemporal classification conditions; An embedding model library of elderly care semantics is constructed based on the semantic similarity and the spatiotemporal classification conditions.

6. The smart elderly care system based on big data integrated analysis as claimed in claim 1, characterized in that: According to the embedded model library, a data stream response is performed based on the elderly care service information, and the elderly care response data obtained specifically includes: Performing embedding vector encoding on the elderly care service information to obtain a service semantic vector library; Semantic matching is performed on the embedding model library and the service semantic vector library, and then the elderly care response data is obtained by screening according to the matching results.

7. The smart elderly care system based on big data integrated analysis as claimed in claim 1, characterized in that: The elderly care service knowledge graph is a heterogeneous knowledge graph.

8. The smart elderly care system based on big data integrated analysis as claimed in claim 1, characterized in that: Based on the elderly care service knowledge graph, the resource scheduling reasoning of the elderly care response data is performed to obtain the scheduling optimization matrix, which specifically includes: Scheduling candidates for the elderly care response data based on the entity relationships in the elderly care knowledge graph to obtain a candidate set of service resources; Sorting and scoring the service resource candidate set to obtain a scoring result; The scoring results are inferred by the inference engine in the elderly care knowledge graph to obtain a resource scheduling set, which is then scored and optimized and converted into a scheduling optimization matrix.

9. The smart elderly care system based on big data integrated analysis as claimed in claim 1, characterized in that: Obtain pension demand data through IoT terminals.

10. The smart elderly care system based on big data integrated analysis as claimed in claim 1, characterized in that: The service effectiveness model is performed on the elderly care demand data to obtain a service evaluation factor set including: Obtain service indicator data for elderly care needs; Constructing a multi-factor weighted model based on the service index data, and then performing an efficiency analysis on the elderly care demand data based on the multi-factor weighted model to obtain service efficiency; A service evaluation is performed on the elderly care demand data based on the service effectiveness, thereby obtaining a service evaluation factor set.

Images