Service demand matching method and device, storage medium and program product

Abstract

The embodiment of the invention provides a service demand matching method and device, a storage medium and a program product. According to the method, user service demand information is collected through a dialogue interaction node in a natural language multi-round interaction mode; extracting structured demand data from the demand information through a code execution node; calling an external matching engine service through an HTTP request node; based on a pre-constructed service provider knowledge graph, performing hard constraint, skill and space-time multi-dimensional matching calculation on the structured demand data to obtain a matching result; and generating recommendation information through the template generation node, and returning the recommendation information to the user through the answer node. The process of natural language interaction, demand analysis and multi-dimensional matching is realized based on the dialogue workflow, so that the matching accuracy and efficiency are improved while the natural and convenient user interaction is ensured.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a service demand matching method, device, storage medium, and program product. Background Technology

[0002] With the widespread application of online service platforms, service demand matching, as a core link connecting users and service resources, directly impacts user experience and platform operational efficiency in terms of efficiency and quality. Current mainstream matching technologies primarily rely on two implementation methods: first, users submit their demands through pre-defined structured forms (such as service type, time, and location fields), and the system matches them based on a rule base; second, users input free text keywords, and the system returns a service list through keyword retrieval. However, the structured form method is mechanical and rigid, requiring users to pre-define various parameters, making it difficult to adapt to the ambiguity and diversity of natural language expression, easily leading to omissions of demand information or increased user workload. While the keyword retrieval method has a lower interaction threshold, it only achieves shallow literal matching, lacking the ability to deeply analyze the semantic connotations and implicit conditions of user demands. In real-world scenarios requiring comprehensive consideration of multiple factors, the reliability and applicability of the matching results are significantly insufficient. Furthermore, in existing technologies, demand collection and matching calculation are often handled separately, lacking a dynamic collaborative mechanism, further exacerbating the inherent contradiction between interactive experience and matching quality.

[0003] In summary, existing service demand matching methods struggle to ensure natural and convenient user interaction while effectively improving the depth of understanding of user intent and the accuracy of matching results. Summary of the Invention

[0004] This application provides a service demand matching method, device, storage medium, and program product, which improves the accuracy and efficiency of matching while ensuring natural and convenient user interaction.

[0005] This application provides a service demand matching method applicable to server-side devices. The server-side devices are pre-configured with a dialogue workflow system, which includes: dialogue interaction nodes, code execution nodes, HTTP request nodes, template generation nodes, and response nodes. The method includes: conducting multi-round interactions with a user via the dialogue interaction nodes using natural language dialogue to collect the user's service demand information; extracting structured demand data of at least one demand dimension from the service demand information via the code execution nodes; invoking an external matching engine service via the HTTP request nodes; performing multi-dimensional matching calculations on the structured demand data based on a pre-built service provider knowledge graph in the external matching engine service to obtain matching results; the multi-dimensional matching includes at least hard constraint matching, skill matching, and spatiotemporal matching; generating recommendation information based on the matching results via the template generation nodes; and returning the recommendation information to the user via the response nodes.

[0006] Optionally, the dialogue interaction node includes a large language model node and a conditional judgment node; through the dialogue interaction node, multiple rounds of interaction are conducted with the user in a natural language dialogue manner to collect the user's service demand information, including: in response to the user's initial input demand, generating and outputting a target question through the large language model node; determining the completeness of the user's answer to the target question through the conditional judgment node; if the information completeness of the answer is not higher than a preset completeness threshold, generating and outputting a new target question through the large language model node, until the conditional judgment node determines that the completeness of the user's answer to the new target question is higher than the preset completeness threshold, and generating service demand information based on the user's answer to the new target question.

[0007] Optionally, extracting structured requirement data for at least one requirement dimension from the service requirement information includes: identifying candidate information for at least one requirement dimension from the service requirement information, the candidate information including: device type, anomaly, location information, time requirement, and / or certificate information; standardizing the candidate information to obtain standardized data, and identifying security constraints present in the standardized data; and encapsulating the standardized data and the security constraints into structured requirement data in a target format.

[0008] Optionally, the candidate information is standardized to obtain standardized data, including: mapping the candidate information to a predefined standardized category to obtain candidate standardized data; in response to identifying that the location information contains floor information and the floor is higher than a preset threshold, adding safety constraints corresponding to high-altitude operations to the candidate standardized data; in response to identifying that the time requirement contains keywords indicating urgency, adding category labels corresponding to urgent needs to the candidate standardized data; in response to identifying that the certificate information contains a specific certificate name, adding attribute information associated with the certificate name to the candidate standardized data, wherein the attribute information includes common tools and / or service characteristics, to obtain the standardized data.

[0009] Optionally, based on a pre-built service provider knowledge graph, multi-dimensional matching calculations are performed on the structured requirement data to obtain matching results, including: selecting service providers from the service provider knowledge graph that meet the security constraints in the structured requirement data as first service providers matched through hard constraints; the first service provider may be one or more; performing skill matching on the first service provider and the structured requirement data to obtain a first degree of matching between the skills of the first service provider and the anomalies in the structured requirement data; performing spatiotemporal matching on the first service provider and the structured requirement data to obtain a second degree of matching between the spatial attribute information of the first service provider and the structured requirement data; calculating a target matching score based on the first and second matching scores; and selecting target service providers from the first service providers that match the structured requirement data in multiple dimensions based on the target matching score to obtain matching results.

[0010] Optionally, from the service provider knowledge graph, service providers that meet the safety constraints in the structured requirements data are selected as the first service providers through hard constraint matching. This includes: if the structured requirements data contains safety constraints corresponding to high-altitude operations, querying whether any service provider in the service provider knowledge graph is associated with valid high-altitude operation permit information; if not, then any service provider does not meet the safety constraints; if yes, then any service provider is the first service provider that meets the safety constraints.

[0011] Optionally, the service provider knowledge graph includes multiple skill nodes for each first service provider; skill matching is performed on the first service provider and the structured requirement data to obtain a first matching degree between the skills of the first service provider and the anomalies in the structured requirement data, including: generating multiple skill requirements based on the anomalies in the structured requirement data; determining multiple target skill nodes from the multiple skill nodes of each first service provider that match the semantic information of the multiple skill requirements; and performing a weighted calculation on the candidate matching degree of attribute information between each target skill node and its matched skill requirements to obtain a first matching degree between the skills of each first service provider and the anomalies in the structured requirement data.

[0012] Optionally, the service provider knowledge graph includes service range nodes for each first service provider; spatiotemporal matching is performed on the first service provider and the structured demand data to obtain a second matching degree of spatial attribute information between the first service provider and the structured demand data, including: calculating the geographical distance and service range conformity between the user and each first service provider based on the spatial attribute information of the structured demand data and the spatial attribute information of the service range nodes of each first service provider; calculating the response time of each first service provider based on the current service status and queuing status of each first service provider; and calculating the second matching degree of spatial attribute information between each first service provider and the structured demand data based on the geographical distance, the service range conformity, and the response time.

[0013] Optionally, generating recommendation information based on the matching results includes: selecting a target explanation template from predefined explanation templates based on the matching results; filling the target explanation template with the details of each target service provider according to the priority order of each target service provider to obtain candidate information; generating action suggestion information based on the details of each target service provider, and adding the action suggestion information to the candidate information to obtain recommendation information; the action suggestion information includes: selection method, contact information, and / or adjustment of needs options.

[0014] This application also provides a server-side device, including: a memory and a processor; wherein the memory is used to: store one or more computer instructions; and the processor is used to execute the one or more computer instructions to: perform the steps in the service demand matching method.

[0015] This application also provides a computer-readable storage medium that, when a computer program is executed by a processor, enables the processor to implement the steps in the service demand matching method.

[0016] This application also provides a computer program product, including a computer program / instructions, which, when executed by a processor, enable the processor to implement the steps in the service demand matching method.

[0017] In this embodiment, user service request information can be collected through multi-turn natural language interaction via dialogue interaction nodes; structured request data can be extracted from the request information via code execution nodes; external matching engine services can be invoked via HTTP request nodes; based on a pre-built service provider knowledge graph, hard constraints, skills, and spatiotemporal multi-dimensional matching calculations are performed on the structured request data to obtain matching results; recommendation information is generated via template generation nodes and returned to the user by answer nodes. This dialogue workflow enables natural language interaction, request parsing, and multi-dimensional matching, improving matching accuracy and efficiency while ensuring natural and convenient user interaction. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart illustrating a service demand matching method provided for an exemplary embodiment of this application; Figure 2 A schematic diagram of a server device provided for an exemplary embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, and displayed data) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use, and processing of such data must comply with the relevant laws, regulations, and standards of the relevant countries and regions, and corresponding access points are provided for users to choose to authorize or refuse. In addition, the various models involved in this application (including but not limited to language models or large models) comply with relevant laws and standards.

[0021] With the widespread application of online service platforms, service demand matching, as a core link connecting users and service resources, directly impacts user experience and platform operational efficiency in terms of efficiency and quality. Current mainstream matching technologies primarily rely on two implementation methods: first, users submit their needs through pre-set structured forms, and the system matches them based on a rule base; second, users input free text keywords, and the system returns a service list through keyword retrieval. However, the structured form method is mechanical and rigid, requiring users to pre-define various parameters, making it difficult to adapt to the ambiguity and diversity of natural language expression, easily leading to omissions of demand information or increased user workload. While the keyword retrieval method has a lower interaction threshold, it only achieves shallow literal matching, lacking the ability to deeply analyze the semantic connotations and implicit conditions of user needs. In real-world scenarios requiring comprehensive consideration of multiple factors, the reliability and applicability of the matching results are significantly insufficient. Furthermore, in existing technologies, demand collection and matching calculation are often handled separately, lacking a dynamic collaborative mechanism, further exacerbating the inherent contradiction between interactive experience and matching quality.

[0022] In summary, existing service demand matching methods struggle to ensure natural and convenient user interaction while effectively improving the depth of understanding of user intent and the accuracy of matching results.

[0023] To address the aforementioned technical issues, this application provides a technical solution that leverages a pre-configured dialogue workflow system to achieve fully automated matching from user natural language service requests to precise service recommendations. Specifically, multiple nodes can work together to transform unstructured natural language requests into structured data, combine this data with the service provider's knowledge graph to achieve multi-dimensional precise matching, and finally generate and provide recommendation information.

[0024] Figure 1 The service demand matching method provided in the exemplary embodiments of this application is applicable to server-side devices. The server-side device, serving as the execution carrier of the service demand matching method, is deployed on the service provider side and can be implemented as any type of server, such as a physical server or a cloud server based on a cloud computing architecture. The server-side device may be pre-configured with a dialogue workflow system, which can be implemented in any form, such as an application, a mini-program, or a WEB (World Wide Web) service page. The dialogue workflow system may include: dialogue interaction nodes, code execution nodes, HTTP (Hypertext Transfer Protocol) request nodes, template generation nodes, and response nodes. Within the dialogue workflow system, each node can be independently downloaded, deployed, installed, and uninstalled.

[0025] like Figure 1 As shown, the method may include the following steps: Step 11: Through dialogue interaction nodes, conduct multiple rounds of interaction with users in a natural language dialogue manner to collect information on users' service needs.

[0026] Step 12: Extract structured requirement data for at least one requirement dimension from the service requirement information through the code execution node.

[0027] Step 13: Call the external matching engine service via the HTTP request node.

[0028] Step 14: In the external matching engine service, based on the pre-built service provider knowledge graph, perform multi-dimensional matching calculations on the structured requirement data to obtain the matching results; multi-dimensional matching includes at least hard constraint matching, skill matching and spatiotemporal matching.

[0029] Step 15: Generate nodes using templates, generate recommended information based on the matching results, and return the recommended information to the user through the answer node.

[0030] In this embodiment, multiple rounds of communication with users can be conducted through preset dialogue interaction nodes using natural language dialogue, aiming to comprehensively and accurately collect users' service demand information. Service demand information includes, but is not limited to: service type (such as furniture repair, home installation, or pipe dredging), core demand indicators (such as skill level for furniture repair or time limit requirements for home installation), hard constraints (such as service area, service fee cap, or service qualification requirements), and preferred demands (such as the service provider's reputation or response speed), as well as implicit demands. For example, implicit demands can be extracted from the dialogue using an NLP intent inference model; if a user mentions "emergency handling," the implicit constraint of "service response time ≤ 2 hours" can be inferred.

[0031] The dialogue interaction nodes can include: large language model nodes and conditional judgment nodes. Large language model nodes integrate large language models of any type, possessing contextual understanding, intent recognition, and multi-turn dialogue memory capabilities. They employ a "guided questioning + intent completion" logic, guiding users to clarify their needs through follow-up questions and examples, while also capturing ambiguous expressions and completing implicit requirements. Furthermore, conditional judgment nodes can verify the completeness of user responses until the service request information meets the requirements, ultimately collecting unstructured natural language service request information. The collected service request information can be temporarily stored in unstructured natural language text format.

[0032] For example, if a user only states "I need to find a plumber to unclog the pipes," the dialogue interaction node can automatically ask follow-up questions such as "How long have the pipes been blocked? When would you like the repairman to come?" until the complete user request, i.e., service demand information, is obtained.

[0033] Subsequently, structured requirement data of at least one requirement dimension can be extracted from the service requirement information through code execution nodes. This involves invoking pre-defined code logic, such as natural language processing algorithms, entity recognition algorithms, regular expression matching rules, or data cleaning rules, to parse, extract, organize, and classify the unstructured service requirement information, thereby extracting structured requirement data of at least one requirement dimension. The requirement dimension can be flexibly defined according to the service scenario, such as service type, service capability requirements, service price thresholds, service geographical scope, service time requirements, and service qualification requirements. Each requirement dimension can correspond to a specific numerical type. The structured data format of the requirement data can be any form, and this embodiment does not impose any restrictions, such as key-value pairs or data tables.

[0034] In this embodiment, an external matching engine service can be invoked via an HTTP request node. The HTTP request node can initiate a service call request to the independently deployed external matching engine service using any network communication protocol, and the service call request carries structured requirement data. The external matching engine service is a professional computing service independent of the dialogue workflow system and can be deployed separately on a server, decoupled from the dialogue workflow system of the server-side device. Based on this, subsequent high-performance matching calculations can be separated from the dialogue workflow system, avoiding the consumption of server-side system resources by matching calculations and ensuring the smoothness of dialogue interaction. Furthermore, it facilitates independent upgrades, maintenance, and iterations of the external matching engine service without requiring additional modifications to the dialogue workflow system; leveraging the professional computing capabilities of the external matching engine service can improve the efficiency and accuracy of matching calculations.

[0035] In the external matching engine service, multi-dimensional matching calculations can be performed on structured requirement data based on a pre-built service provider knowledge graph to obtain matching results. This multi-dimensional matching includes at least hard constraint matching, skill matching, and spatiotemporal matching. The service provider knowledge graph is a semantically linked data model. Nodes in the knowledge graph can include service providers, service skills, service qualifications, service geographical scope, service time, service price, and service cases, while edges represent the relationships between nodes.

[0036] Hard constraint matching refers to a matching process that uses the basic service conditions (i.e., the hard conditions that need to be met and have a small negotiation space) in the structured demand data as the basis for matching, combined with the attribute information of service providers in the service provider knowledge graph, to achieve compliance matching. Hard constraint matching can be used to eliminate service providers that do not meet the user's hard conditions and output a preliminary set of qualified service provider candidates.

[0037] Skill matching refers to a matching method that uses skill requirements (such as professional competence, technical skills, service level, or professional qualification level) in structured demand data as the basis for matching, and combines the semantic association between service providers and various skills in the service provider knowledge graph to achieve a matchability determination.

[0038] Spatiotemporal matching refers to scenario adaptability matching that integrates spatial and temporal dimensions. It is a spatiotemporal adaptability determination made by combining the spatial service attributes (such as geographical service range and offline service coverage area) and temporal service attributes (such as service schedule, delivery timeliness, and service time period) of the service provider in the service provider's knowledge graph for scenario-based needs such as geographical services or time arrangements in structured demand data.

[0039] This multi-dimensional matching approach leverages the semantic association characteristics of knowledge graphs to achieve multi-dimensional, hierarchical, and precise matching calculations. Unlike traditional single-keyword matching, it solves the problems of "single matching dimension, low accuracy, and easy omission / mismatch" in traditional matching methods, improving computational efficiency while ensuring matching accuracy.

[0040] After obtaining the matching results, a template-based generation node can be used to generate recommended information based on the matching results, and then the recommended information can be returned to the user through a response node. The template-based generation node can have a built-in library of preset recommended information templates, which can automatically fill the corresponding dynamically populated fields in the recommended information templates with the matching results to generate the recommended information.

[0041] The response node can serve as the user-side output node of the dialogue workflow system, transforming the recommendation information generated by the templated generation node into a form that is easily accepted by the user (such as natural language text, visual lists, voice broadcasts, etc.), and returning it to the user through the interactive interface of the server-side device (such as APP, mini-program, web page, smart terminal).

[0042] In this embodiment, user service request information can be collected through multi-turn natural language interaction via dialogue interaction nodes; structured request data can be extracted from the request information via code execution nodes; external matching engine services can be invoked via HTTP request nodes; based on a pre-built service provider knowledge graph, hard constraints, skills, and spatiotemporal multi-dimensional matching calculations are performed on the structured request data to obtain matching results; recommendation information is generated via template generation nodes and returned to the user by answer nodes. This dialogue workflow enables natural language interaction, request parsing, and multi-dimensional matching, improving matching accuracy and efficiency while ensuring natural and convenient user interaction.

[0043] In some optional embodiments, when collecting user service request information through multi-turn interactions with the user via dialogue interaction nodes using natural language dialogue, the system can generate and output target questions in response to the user's initial input. The training samples of the large language model set in the large language model node can include various service scenario request cases, request dimension labels, typical information gap questions, etc. After training, it possesses the following capabilities: generating targeted and structured follow-up questions (i.e., target questions) based on user needs, adapting to the request collection requirements of different service scenarios, and without fixed wording restrictions.

[0044] The conditional judgment node can determine the completeness of a user's answer to a target question. This node can be configured with information completeness judgment rules and preset completeness thresholds. The information completeness judgment rules are strongly tied to the required dimensions for subsequent structured extraction. These dimensions can include: whether the information covers the required dimensions of the corresponding service scenario, whether the information is unambiguous, and whether the information meets the basic requirements for structured extraction by subsequent code execution nodes.

[0045] If the completeness of the answer is not higher than the preset completeness threshold, a new target question is generated and output through the large language model node, until the condition judgment node determines that the completeness of the user's answer to the new target question is higher than the preset completeness threshold, and service requirement information is generated based on the user's answer to the new target question.

[0046] This approach enables intelligent and proactive follow-up questions, solving the problems of rigidity and easy omission of information in traditional manual pre-set follow-up questions, and can collect service demand information more accurately.

[0047] In some optional embodiments, when extracting structured demand data for at least one demand dimension from service demand information, candidate information for at least one demand dimension can be identified from the service demand information. Specifically, a pre-trained NLP (Natural Language Processing) entity recognition model can be used to perform entity recognition on the service demand information, identifying candidate information for at least one demand dimension. The candidate information for at least one demand dimension may include at least one of the following: device type, anomaly, location information, time requirement, and certificate information. Here, device type refers to the category of service-related device / service carrier; anomaly refers to an anomaly existing in the service object / service carrier; location information is the location where the service is provided; time requirement is the time interval for service provision; and certificate information is a service-related certificate or other qualification identifier.

[0048] Next, the candidate information can be standardized to obtain standardized data, and the security constraints present in the standardized data can be identified. Specifically, the candidate information can be transformed into standardized data with a uniform format and standardized information through unified standardization processing. At the same time, the constraints related to service security can be identified in the standardized data, and finally, the standardized data and security constraints are output.

[0049] Specifically, candidate information can be mapped to predefined standardized categories to obtain candidate standardized data. This mapping can be based on a pre-established standardized classification dictionary for each dimension of service requirements. For example, the standardized classification of location information can be standardized according to a hierarchical format of "province-city-district-street-community-building-floor," unifying the location description standard.

[0050] After obtaining the candidate standardized data, in response to the identification that the location information contains floor information and the floor is higher than a preset threshold, safety constraints corresponding to high-altitude operations can be added to the candidate standardized data. For example, the safety constraints for high-altitude operations could be: domestic workers must have safety awareness for high-altitude operations, or there are no records of domestic workers not wearing safety gloves when wiping windows or performing other high-altitude operations. In response to the identification that the time requirement contains keywords indicating urgency, category labels corresponding to urgent needs can be added to the candidate standardized data. For example, the category label for urgent needs could be "Priority: High". In response to the identification that the certificate information contains a specific certificate name, attribute information associated with the certificate name can be added to the candidate standardized data. The attribute information is: common tools and / or service characteristics, to obtain standardized data. The certificate information here corresponds to the service qualification information mentioned above. Among them, the common accessories associated with the elderly care worker professional qualification certificate are: "daily cleaning tools, basic tools for preparing complementary foods, and basic tools for elderly care"; the associated service characteristics are: "familiar with the common knowledge of caring for elderly people living alone, possessing emergency handling capabilities, able to prepare soft and easy-to-swallow complementary foods, and not leaving the post without authorization". In this way, candidate information can be standardized in a relatively efficient and accurate manner.

[0051] Then, standardized data and security constraints can be encapsulated into structured requirement data in the target format. The target format can be any format, such as XML (Extensible Markup Language) or database table structure format.

[0052] In this way, service demand information can be transformed from unstructured to structured, and through standardized processing, upfront security constraints, and data format adaptation, a foundation can be laid for subsequent multi-dimensional matching.

[0053] In some optional embodiments, step 14 in the foregoing embodiments, "based on a pre-built service provider knowledge graph, perform multi-dimensional matching calculations on structured requirement data to obtain matching results," can be implemented based on the following steps: Step 141: From the service provider knowledge graph, select the service providers that meet the security constraints in the structured requirement data as the first service providers through hard constraint matching; the first service provider can be one or more.

[0054] Specifically, if the structured requirements data includes safety constraints for high-altitude operations, the system can query the service provider knowledge graph to determine if any service provider is associated with a valid high-altitude operation permit. If not, the service provider does not meet the safety constraints; if so, it is the first service provider that meets the safety constraints. Optionally, after identifying the first service provider that passes the hard constraint matching, the issuing authority and validity period of the first service provider's work permit can be recorded as interpretable evidence. Interpretable evidence refers to structured and traceable supporting information. Based on this step, service providers that do not meet safety requirements can be quickly eliminated through pre-selective hard filtering based on safety constraints, ensuring the safety and compliance of service matching from the source. This also reduces the computational load of subsequent multi-dimensional matching, improves overall matching efficiency, and avoids resource consumption caused by invalid matching.

[0055] Step 142: Perform skill matching on the first service provider and the structured demand data to obtain the first matching degree between the skills of the first service provider and the anomalies in the structured demand data. The service provider knowledge graph may contain multiple skill nodes for each first service provider.

[0056] Specifically, multiple skill requirements can be generated based on anomalies in structured demand data. For example, if the anomaly is a broken water pipe, the generated skill requirements could include: the ability to locate the pipe break and troubleshoot leaks, as well as the ability to replace and install pipe fittings. Another example is if the anomaly is an elderly person living alone with paralysis; the generated skill requirements could include: basic daily care for an elderly person living alone, emergency response skills for the elderly, and the ability to process and prepare soft, mushy complementary foods.

[0057] From the multiple skill nodes of each first service provider, multiple target skill nodes that match the semantic information of multiple skill requirements are identified. Specifically, from the multiple skill nodes of each first service provider, multiple target skill nodes that match the semantic information of multiple skill requirements can be identified through natural language semantic similarity calculation and predefined skill vocabulary mapping. Each skill node pre-stored in the service provider's knowledge graph carries attribute information, which may include at least one of the following: skill proficiency, professional operational qualifications, and years of practical experience.

[0058] The candidate matching degree of attribute information between each target skill node and its matched skill requirements is weighted and calculated to obtain the first matching degree between the skills of each first service provider and the anomalies in the structured requirement data. Specifically, the attribute information of the target skill node can first be quantified into candidate matching degrees for the corresponding skill requirements. For example, skill proficiency can be quantified as 100 points for mastery, 80 points for proficiency, 60 points for basic mastery, and 0 points for no mastery. Then, preset weights are assigned based on the contribution of each skill requirement to resolving the anomalies. Finally, based on the preset weights, the candidate matching degrees of attribute information between each target skill node and its matched skill requirements are weighted and summed to obtain the first matching degree between the skills of each first service provider and the anomalies in the structured requirement data.

[0059] Optionally, after obtaining the first matching degree, interpretable evidence can be generated based on the attribute information of skill nodes whose candidate matching degrees meet preset conditions.

[0060] Step 143: Perform spatiotemporal matching on the first service provider and the structured demand data to obtain the second matching degree of the spatial attribute information of the first service provider and the structured demand data.

[0061] The service provider knowledge graph can include a service scope node for each primary service provider. This service scope node is a spatial attribute node constructed separately for each primary service provider within the knowledge graph, forming the core data foundation for spatiotemporal matching. The service scope node carries standardized spatial attribute information, rather than a vague description of the scope, specifically including the area covered by the service, the geographical radius of influence, service boundary restrictions, and spatial limitations of the service scenarios. For example, the service scope node for a plumber could be labeled as "the entire area of ​​XX District, XX City, within 8 kilometers of XX residential area, capable of undertaking plumbing repairs in residential / outdoor public areas."

[0062] The system can calculate the geographical distance and service range compliance between a user and each first service provider based on the spatial attribute information of the structured demand data and the spatial attribute information of the service range nodes of each first service provider. The spatial attribute information is the service location information extracted from the user's service demand information, such as "Room 101, Unit 2, Building 3, XX Community, XX City," which includes hierarchical spatial attribute information such as province / city / district / street / community / building, and can be directly compared in a structured manner with the spatial attributes of the service range nodes. When calculating the geographical distance between a user and each first service provider, the actual traversable distance between them can be calculated based on the spatial attribute information of the structured demand (user service location) and the permanent location / service center point of the service range node in the service provider's knowledge graph, and this distance is used as the geographical distance between the user and each first service provider. When calculating the service range compliance between a user and each first service provider, it can be determined whether the user's service location is within the coverage area of ​​the service provider's service range node, and the result is quantified as the service range compliance score, such as 100 points for being completely within the coverage area, 80 points for being partially in the edge coverage area, and 0 points for being outside the coverage area.

[0063] The response time for each first service provider is calculated based on its current service status and queuing situation. For example, in the emergency repair scenario of a broken water pipe, if the repairman is currently idle, the response time can be 2 minutes; if the repairman is in a service position and has 2 queued orders, the response time can be 60 minutes.

[0064] Based on geographical distance, service range compliance, and response time, a second degree of matching between the spatial attribute information of each first service provider and the structured demand data is calculated. Specifically, geographical distance, service range compliance, and response time can be weighted according to their respective preset weights to obtain the second degree of matching between the spatial attribute information of each first service provider and the structured demand data.

[0065] Based on steps 142-143 above, targeted matching is carried out for skills and spatiotemporal dimensions, realizing dual verification of the service provider's capabilities and scenario adaptability. Skill matching ensures that the service provider's professional capabilities are compatible with the needs for solving abnormal phenomena in the requirements, while spatiotemporal matching ensures that the service provider's spatial attributes are consistent with the spatiotemporal requirements of the requirement scenario. This solves the one-sidedness problem of single-dimensional matching and greatly improves the accuracy of matching.

[0066] Step 144: Calculate the target matching score based on the first matching degree and the second matching degree. This embodiment does not limit the specific calculation method; it can be either adding the first matching degree and the second matching degree to obtain the target matching score, or weighted summing the first matching degree and the second matching degree to obtain the target matching score.

[0067] Step 145: Based on the target matching score, select target service providers from the first service providers that match the structured demand data in multiple dimensions to obtain the matching results. Specifically, service providers with the highest target matching score or those meeting other preset conditions (such as being in the top 10%) can be selected from the first service providers as target service providers for multi-dimensional matching with the structured demand data. Optionally, the matching results may include not only detailed information about the target service providers but also interpretable evidence generated in the preceding steps.

[0068] In this way, multi-dimensional matching calculations can be performed based on pre-built service provider knowledge graphs. By making full use of the structured and relational attribute features of the knowledge graphs, it is possible to efficiently retrieve and mine relevant attribute information such as security, skills, and spatiotemporal information of service providers, thereby improving the efficiency and accuracy of matching calculations.

[0069] In some optional embodiments, when generating recommendation information based on the matching results, a target explanation template can be selected from predefined explanation templates based on the matching results. Specifically, based on features included in the matching results such as the matching degree between the service provider and the user's needs, the suitability of the need type, and the urgency of the need, a target explanation template that matches the current user's needs can be selected from predefined explanation templates pre-classified by service category, matching level, and need scenario.

[0070] Based on the priority order of each target service provider, the detailed information of each target service provider is populated into the target explanation template to obtain candidate information. The priority order can be determined based on preset rules, which include one or more of the following: matching degree from high to low, service reputation weight, service response efficiency, and service price cost-effectiveness. It can also be dynamically adjusted in combination with the service preference weights preset by the user. The detailed information of the target service provider may include name, service qualifications, service capability scope, typical service cases, service price range, and service delivery cycle.

[0071] Action suggestions are generated based on the detailed information of each target service provider and added to the candidate information to obtain recommended information. Action suggestions include one or more of the following: selection method, contact information, and adjustment of requirements. Specifically, action suggestions are generated by combining the detailed information of each target service provider with the characteristics of the user's service needs and general industry knowledge. These suggestions are then added to the preset action suggestion field area of ​​the target explanation template and integrated with the already filled candidate information to obtain recommended information. Selection methods include direct selection, multi-solution comparison and filtering, and trial application. Contact information includes channels such as customer service hotline, dedicated online communication portal, offline contact address, and contact person contact information.

[0072] In this way, nodes can be generated based on templates, and recommendation information can be generated more accurately based on the matching results.

[0073] In some optional embodiments, user feedback on recommendation information can also be recorded, including accepting recommendations, rejecting recommendations, or requesting adjustments. User feedback is associated with and stored in relation to corresponding matching query requests and matching results. Matching weights in the external matching engine service and the question generation logic in the dialogue workflow system are periodically optimized based on user feedback.

[0074] In some optional embodiments, real-time status nodes of service providers can also be maintained in the service provider knowledge graph, and the location, service status and availability of service providers can be updated in real time through service provider clients or IoT devices; in the matching calculation, service providers with an idle status and geographical proximity are given priority as target service providers.

[0075] In some optional embodiments, a caching layer can be set between the dialogue workflow system and the external matching engine service. A cache key can be generated based on structured requirement data. The system checks whether there is a matching result in the cache. If there is a matching result in the cache and it has not expired, the cached result is used directly. If there is no matching result in the cache or it has expired, the external matching engine service is called. The result returned by the external matching engine service is stored in the cache and a reasonable expiration time is set.

[0076] In some optional embodiments, different matching strategies can be selected based on the complexity of the requirements and the user's historical behavior; the matching strategies include at least a fast matching strategy and a deep matching strategy. The fast matching strategy can filter and sort based on keywords and simple rules; the deep matching strategy can perform multi-dimensional matching calculations based on knowledge graphs and machine learning models.

[0077] In some optional embodiments, an exception capture node can be set up in the dialogue workflow system to monitor the execution status of each node. In response to a node execution failure, error information can be logged and backup logic can be executed; the backup logic may include returning a default recommendation list, prompting the user to try again later, or transferring the user to a human agent.

[0078] In some alternative embodiments, the above embodiments may be applicable to the field of home repair services, including at least one of air conditioner repair, pipe dredging, appliance repair, and home installation services.

[0079] It should be noted that the execution subject of each step of the method provided in the above embodiments can be the same device, or the method can be executed by different devices. For example, the execution subject of steps 11 to 15 can be device A; or the execution subject of steps 11 to 12 can be device A, and the execution subject of steps 13 to 15 can be device B; and so on.

[0080] Furthermore, some processes described in the above embodiments and accompanying drawings include multiple operations that appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or they may be executed in parallel. The operation numbers, such as 12, 13, etc., are merely used to distinguish different operations and do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel.

[0081] It should be noted that the terms "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0082] Figure 2 This is a schematic diagram of the structure of a server-side device provided in an exemplary embodiment of this application. This server-side device is applicable to the service demand matching method provided in the foregoing embodiments, such as... Figure 2 As shown, the server-side device may include: a memory 201, a processor 202, and a communication component 203. The server-side device is pre-configured with a dialogue workflow system, which includes: dialogue interaction nodes, code execution nodes, HTTP request nodes, template generation nodes, and response nodes.

[0083] Memory 201 is used to store computer programs and can be configured to store various other data to support operation on the server device. Examples of this data include instructions for any application or method operating on the server device, contact data, phonebook data, messages, pictures, videos, etc.

[0084] In some exemplary embodiments, processor 202, coupled to memory 201, is configured to execute computer programs in memory 201 for: engaging in multi-turn interactions with a user via the dialogue interaction node in a natural language dialogue manner to collect the user's service request information; extracting structured demand data of at least one demand dimension from the service request information via the code execution node; invoking an external matching engine service via the HTTP request node; performing multi-dimensional matching calculations on the structured demand data based on a pre-built service provider knowledge graph in the external matching engine service to obtain matching results; the multi-dimensional matching includes at least hard constraint matching, skill matching, and spatiotemporal matching; generating recommendation information based on the matching results via the template generation node, and returning the recommendation information to the user via the answer node.

[0085] Optionally, the dialogue interaction node includes a large language model node and a conditional judgment node. When the processor 202 interacts with the user in a natural language dialogue manner through the dialogue interaction node to collect the user's service request information, it is specifically used to: respond to the user's initial input request, generate and output a target question through the large language model node; determine the completeness of the user's answer to the target question through the conditional judgment node; if the information completeness of the answer is not higher than a preset completeness threshold, generate and output a new target question through the large language model node, until the conditional judgment node determines that the completeness of the user's answer to the new target question is higher than the preset completeness threshold, and generate service request information based on the user's answer to the new target question.

[0086] Optionally, when the processor 202 extracts structured requirement data for at least one requirement dimension from the service requirement information, it is specifically used to: identify candidate information for at least one requirement dimension from the service requirement information, the candidate information including: device type, anomaly, location information, time requirement and / or certificate information; perform standardization processing on the candidate information to obtain standardized data, and identify the security constraints existing in the standardized data; and encapsulate the standardized data and the security constraints into structured requirement data in a target format.

[0087] Optionally, when the processor 202 performs standardization processing on the candidate information to obtain standardized data, it specifically performs the following: mapping the candidate information to a predefined standardized category to obtain candidate standardized data; in response to recognizing that the location information contains floor information and the floor is higher than a preset threshold, adding safety constraints corresponding to high-altitude operations to the candidate standardized data; in response to recognizing that the time requirement contains keywords indicating urgency, adding category labels corresponding to urgent needs to the candidate standardized data; and in response to recognizing that the certificate information contains a specific certificate name, adding attribute information associated with the certificate name to the candidate standardized data, wherein the attribute information includes common tools and / or service characteristics, to obtain the standardized data.

[0088] Optionally, the processor 202 performs multi-dimensional matching calculations on the structured requirement data based on a pre-built service provider knowledge graph. When obtaining the matching results, it specifically performs the following: From the service provider knowledge graph, it identifies service providers that meet the security constraints in the structured requirement data as first service providers matched through hard constraints; the first service provider can be one or more; it performs skill matching on the first service provider and the structured requirement data to obtain a first degree of matching between the skills of the first service provider and the anomalies in the structured requirement data; it performs spatiotemporal matching on the first service provider and the structured requirement data to obtain a second degree of matching between the spatial attribute information of the first service provider and the structured requirement data; it calculates a target matching score based on the first and second matching scores; and it selects target service providers from the first service providers that are multi-dimensionally matched with the structured requirement data based on the target matching score, thus obtaining the matching results.

[0089] Optionally, when the processor 202 selects a service provider that meets the safety constraints in the structured requirements data from the service provider knowledge graph as the first service provider through hard constraint matching, it specifically performs the following: when the structured requirements data contains safety constraints corresponding to high-altitude operations, it queries whether any service provider in the service provider knowledge graph is associated with valid high-altitude operation permit information; if not, then any service provider does not meet the safety constraints; if yes, then any service provider is the first service provider that meets the safety constraints.

[0090] Optionally, the service provider knowledge graph includes multiple skill nodes for each first service provider; when the processor 202 performs skill matching on the first service provider and the structured requirement data to obtain a first matching degree between the skills of the first service provider and the anomalies in the structured requirement data, it is specifically used to: generate multiple skill requirements based on the anomalies in the structured requirement data; determine multiple target skill nodes from the multiple skill nodes of each first service provider that match the semantic information of the multiple skill requirements; and perform a weighted calculation on the candidate matching degree of attribute information between each target skill node and its matched skill requirements to obtain a first matching degree between the skills of each first service provider and the anomalies in the structured requirement data.

[0091] Optionally, the service provider knowledge graph includes service range nodes for each first service provider; when the processor 202 performs spatiotemporal matching on the first service provider and the structured demand data to obtain a second matching degree of spatial attribute information between the first service provider and the structured demand data, it is specifically used to: calculate the geographical distance and service range conformity between the user and each first service provider based on the spatial attribute information of the structured demand data and the spatial attribute information of the service range nodes of each first service provider; calculate the response time of each first service provider based on the current service status and queuing status of each first service provider; and calculate the second matching degree of spatial attribute information between each first service provider and the structured demand data based on the geographical distance, the service range conformity, and the response time.

[0092] Optionally, when the processor 202 generates recommendation information based on the matching result, it is specifically used to: select a target explanation template from predefined explanation templates based on the matching result; fill the detailed information of each target service provider into the target explanation template according to the priority order of each target service provider to obtain candidate information; generate action suggestion information based on the detailed information of each target service provider, and add the action suggestion information to the candidate information to obtain recommendation information; the action suggestion information includes: selection method, contact information and / or adjustment of needs options.

[0093] This application also provides a computer-readable storage medium that, when executed by a processor, enables the processor to implement the steps in the service demand matching method.

[0094] This application also provides a computer program product, including a computer program / instructions, which, when executed by a processor, enable the processor to implement the steps in the service demand matching method.

[0095] In this embodiment, user service request information is collected through multi-turn natural language interaction via dialogue interaction nodes; structured request data is extracted from the request information via code execution nodes; an external matching engine service is invoked via HTTP request nodes; based on a pre-built service provider knowledge graph, hard constraints, skills, and spatiotemporal multi-dimensional matching calculations are performed on the structured request data to obtain matching results; recommendation information is generated via template generation nodes and returned to the user by answer nodes. This dialogue workflow enables natural language interaction, request parsing, and multi-dimensional matching, improving matching accuracy and efficiency while ensuring natural and convenient user interaction.

[0096] Furthermore, such as Figure 2 As shown, the server device also includes other components such as a display 204, a power supply component 205, and an audio component 206. Figure 2 The diagram only shows a portion of the components and does not imply that the server-side device only includes... Figure 2 The components shown.

[0097] The aforementioned memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0098] The aforementioned communication components are configured to facilitate wired or wireless communication between the device containing the communication components and other devices. The device containing the communication components can access wireless networks based on communication standards, such as WiFi, 2G, 3G, 4G / LTE, 5G, or combinations thereof. In one exemplary embodiment, the communication components receive broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication components also include a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID), Infrared Data Association (IrDA), Ultra Wide Band (UWB), Bluetooth (BT), and other technologies.

[0099] The aforementioned display includes a screen, which may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a Touch Panel, the screen can be implemented as a touchscreen to receive input signals from the user. The Touch Panel includes one or more touch sensors to sense touches, swipes, and gestures on the Touch Panel. The touch sensors can sense not only the boundaries of touch or swipe actions but also the duration and pressure associated with the touch or swipe operation.

[0100] The aforementioned power supply components provide power to various components within the device in which they reside. These power supply components may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power to the device in which they reside.

[0101] The aforementioned audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC) configured to receive external audio signals when the device containing the audio component is in an operating mode, such as call mode, recording mode, or voice recognition mode. The received audio signals can be further stored in memory or transmitted via a communication component. In some embodiments, the audio component also includes a speaker for outputting audio signals.

[0102] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including, but not limited to, disk storage, compact disc read-only memory (CD-ROM), optical storage, etc.) containing computer-usable program code.

[0103] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0104] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0105] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0106] In a typical configuration, a computing device includes one or more processors (Central Processing Unit, CPU), input / output interfaces, network interfaces, and memory.

[0107] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0108] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, Digital Video Disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0109] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0110] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A service demand matching method, characterized in that, Applicable to server-side devices, the server-side devices are pre-configured with a dialogue workflow system, the dialogue workflow system including: dialogue interaction nodes, code execution nodes, HTTP request nodes, template generation nodes, and response nodes; the method includes: Through the dialogue interaction node, multiple rounds of interaction are conducted with the user in a natural language dialogue manner to collect the user's service request information; Through the code execution node, structured requirement data of at least one requirement dimension is extracted from the service requirement information; The external matching engine service is invoked through the HTTP request node; In the external matching engine service, based on a pre-built service provider knowledge graph, multi-dimensional matching calculations are performed on the structured requirement data to obtain matching results; the multi-dimensional matching includes at least hard constraint matching, skill matching, and spatiotemporal matching. The template generation node generates recommendation information based on the matching results, and the answer node returns the recommendation information to the user.

2. The method according to claim 1, characterized in that, The dialogue interaction nodes include: large language model nodes and condition judgment nodes; Through the aforementioned dialogue interaction node, multiple rounds of interaction are conducted with the user in a natural language dialogue manner to collect the user's service request information, including: In response to the initial request from the user input, the target question is generated and output through the large language model node; The completeness of the user's answer to the target question is determined by the condition judgment node. If the completeness of the answer is not higher than a preset completeness threshold, a new target question is generated and output through the large language model node, until the condition judgment node determines that the completeness of the user's answer to the new target question is higher than the preset completeness threshold, and service demand information is generated based on the user's answer to the new target question.

3. The method according to claim 1, characterized in that, Extracting structured demand data from at least one demand dimension from the service demand information includes: Identify candidate information for at least one demand dimension from the service demand information, the candidate information including: device type, anomaly, location information, time requirement and / or certificate information; The candidate information is standardized to obtain standardized data, and the security constraints present in the standardized data are identified. The standardized data and the security constraints are encapsulated into structured requirement data in the target format.

4. The method according to claim 3, characterized in that, The candidate information is standardized to obtain standardized data, including: The candidate information is mapped to a predefined standardized category to obtain candidate standardized data; In response to the identification that the location information contains floor information and the floor is higher than a preset threshold, safety constraints corresponding to high-altitude operations are added to the candidate standardized data. In response to the identification of keywords indicating urgency in the time requirement, a category label corresponding to the urgency requirement is added to the candidate standardized data; In response to the identification of a specific certificate name in the certificate information, attribute information associated with the certificate name is added to the candidate standardized data, the attribute information being common tool and / or service characteristics, to obtain the standardized data.

5. The method according to claim 1, characterized in that, Based on a pre-built service provider knowledge graph, multi-dimensional matching calculations are performed on the structured requirement data to obtain matching results, including: From the service provider knowledge graph, service providers that meet the security constraints in the structured requirement data are selected as the first service providers through hard constraint matching; the first service provider can be one or more. Skill matching is performed on the first service provider and the structured demand data to obtain a first degree of matching between the skills of the first service provider and the anomalies in the structured demand data; Spatiotemporal matching is performed on the first service provider and the structured demand data to obtain a second matching degree of spatial attribute information between the first service provider and the structured demand data; Calculate the target matching score based on the first matching degree and the second matching degree; Based on the target matching score, target service providers that match the structured demand data in multiple dimensions are selected from the first service providers to obtain the matching results.

6. The method according to claim 5, characterized in that, From the service provider knowledge graph, service providers that meet the security constraints in the structured requirement data are selected as the first service providers through hard constraint matching, including: If the structured requirements data includes safety constraints for high-altitude operations, query whether any service provider in the service provider knowledge graph is associated with valid high-altitude operation permit information; If no, then any of the service providers does not meet the security constraints; if yes, then any of the service providers is the first service provider that meets the security constraints.

7. The method according to claim 5, characterized in that, The service provider knowledge graph contains multiple skill nodes for each first service provider; Skill matching is performed on the first service provider and the structured demand data to obtain a first degree of matching between the skills of the first service provider and the anomalies in the structured demand data, including: Based on the anomalies in the structured requirements data, multiple skill requirements are generated; From the multiple skill nodes of each first service provider, determine multiple target skill nodes that match the semantic information of the multiple skill requirements; The candidate matching degree of attribute information between each target skill node and its matching skill requirements is weighted and calculated to obtain the first matching degree between the skills of each first service provider and the anomalies in the structured requirement data.

8. The method according to claim 5, characterized in that, The service provider knowledge graph contains service scope nodes for each first service provider; Perform spatiotemporal matching on the first service provider and the structured demand data to obtain a second matching degree of spatial attribute information between the first service provider and the structured demand data, including: Based on the spatial attribute information of the structured demand data and the spatial attribute information of the service range nodes of each first service provider, the geographical distance and service range conformity between the user and each first service provider are calculated. Calculate the response time for each first service provider based on its current service status and queuing situation; Based on the geographical distance, the service range compliance, and the response time, a second matching degree is calculated between each first service provider and the spatial attribute information of the structured demand data.

9. The method according to claim 5, characterized in that, Recommendation information is generated based on the matching results, including: Based on the matching results, a target interpretation template is selected from the predefined interpretation templates; According to the priority order of each target service provider, the detailed information of each target service provider is filled into the target explanation template to obtain candidate information; Action suggestion information is generated based on the details of each target service provider, and the action suggestion information is added to the candidate information to obtain recommended information; the action suggestion information includes: selection method, contact information and / or adjustment of needs options.

10. A server-side device, characterized in that, include: A memory and a processor; wherein the memory is used to: store one or more computer instructions; The processor is configured to execute the one or more computer instructions for performing the steps of the method according to any one of claims 1-9.

11. A computer-readable storage medium, characterized in that, When the computer program is executed by the processor, it causes the processor to perform the steps of the method according to any one of claims 1-9.

12. A computer program product, characterized in that, Includes a computer program / instruction that, when executed by a processor, causes the processor to perform the steps of the method according to any one of claims 1-9.

Images