Service recommendation method and device, terminal, storage medium and product

Abstract

The embodiment of the invention discloses a service recommendation method and device, a terminal, a storage medium and a product, and belongs to the technical field of artificial intelligence. Comprising the steps that a first service list of recommendation demand information is determined, the first service list comprises a plurality of candidate services meeting recommendation demands, and the recommendation demand information is used for representing demands of current recommendation services of a terminal; determining a plurality of target evaluation nodes based on the recommendation demand information, the plurality of target evaluation nodes being used for performing quality evaluation on the plurality of candidate services from different evaluation dimensions; based on multiple evaluation dimensions corresponding to the multiple target evaluation nodes, multiple first quality parameter values of the candidate service are determined, and the first quality parameter values of the candidate service are used for representing the degree that the candidate service meets the recommendation demand corresponding to the recommendation demand information; and based on the first quality parameter values of the plurality of candidate services, determining at least one target service from the first service list to obtain a second service list.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of artificial intelligence technology, and in particular to a service recommendation method, device, terminal, storage medium and product. Background Art

[0002] Currently, mobile phones and other terminals have assistant applications, which have a service recommendation function, and the assistant application will recommend the recommended service in the service card on the desktop, so that users can view the service information of the service without entering the application where the service is located. For example, the assistant application will recommend a taxi service to the user based on the user's itinerary, and recommend the taxi service in the service card, so that the user can directly click on the taxi service on the service card to quickly get a taxi. Summary of the invention

[0003] The embodiments of the present application provide a service recommendation method, device, terminal, storage medium and product. The technical solution is as follows:

[0004] On the one hand, an embodiment of the present application provides a service recommendation method, the method comprising:

[0005] Determine a first service list of recommended demand information, the first service list including a plurality of candidate services that meet the recommended demand corresponding to the recommended demand information, the recommended demand information being used to indicate the current recommended service demand of the terminal;

[0006] Based on the recommendation requirement information, determining a plurality of target evaluation nodes, wherein the plurality of target evaluation nodes are used to perform quality evaluation on the plurality of candidate services from different evaluation dimensions;

[0007] For any candidate service, based on the multiple evaluation dimensions corresponding to the multiple target evaluation nodes, determine multiple first quality parameter values ​​of the candidate service, and determine different first quality parameter values ​​of the candidate service based on different evaluation dimensions, wherein the first quality parameter value of the candidate service is used to indicate the degree to which the candidate service satisfies the recommendation requirement corresponding to the recommendation requirement information;

[0008] Based on the first quality parameter values ​​of the multiple candidate services, at least one target service is determined from the first service list to obtain a second service list.

[0009] On the other hand, an embodiment of the present application provides a service recommendation device, the device comprising:

[0010] A first determining module, configured to determine a first service list of recommended demand information, wherein the first service list includes a plurality of candidate services that meet the recommended demand corresponding to the recommended demand information, wherein the recommended demand information is used to indicate the current recommended service demand of the terminal;

[0011] A second determination module is used to determine a plurality of target evaluation nodes based on the recommendation requirement information, wherein the plurality of target evaluation nodes are used to perform quality evaluation on the plurality of candidate services from different evaluation dimensions;

[0012] A third determination module is used to determine, for any candidate service, multiple first quality parameter values ​​of the candidate service based on multiple evaluation dimensions corresponding to the multiple target evaluation nodes, and determine different first quality parameter values ​​of the candidate service based on different evaluation dimensions, wherein the first quality parameter value of the candidate service is used to indicate the degree to which the candidate service satisfies the recommendation requirement corresponding to the recommendation requirement information;

[0013] The fourth determining module is configured to determine at least one target service from the first service list based on the first quality parameter values ​​of the plurality of candidate services to obtain a second service list.

[0014] On the other hand, an embodiment of the present application provides a terminal, which includes a processor and a memory, wherein the memory stores at least one computer instruction, and the at least one computer instruction is loaded and executed by the processor to implement the service recommendation method as described in the above aspects.

[0015] On the other hand, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores at least one computer instruction, and the at least one computer instruction is used to be executed by a processor to implement the service recommendation method as described in the above aspects.

[0016] On the other hand, an embodiment of the present application provides a computer program product, which includes computer instructions, and the computer instructions are stored in a computer-readable storage medium; a processor reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to implement the service recommendation method as described in the above aspects.

[0017] In the embodiment of the present application, a comprehensive evaluation of multiple candidate services is performed through multiple evaluation dimensions, which can effectively integrate the results of different evaluation dimensions. This comprehensive processing method can ensure the comprehensiveness of the evaluation, so that the target service finally determined is more in line with the actual needs of the user, thereby improving the accuracy of the target service recommendation and further improving the service recommendation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a schematic diagram of an implementation environment of a service recommendation method provided by an exemplary embodiment of the present application;

[0019] Figure 2 is a schematic diagram of a quality assessment system provided by an exemplary embodiment of the present application;

[0020] Figure 3 is a flow chart of a service recommendation method shown in an exemplary embodiment of the present application;

[0021] Figure 4 is a schematic diagram of a service recommendation method shown in an exemplary embodiment of the present application;

[0022] Figure 5 is a schematic diagram of determining a resource configuration file according to an exemplary embodiment of the present application;

[0023] Figure 6 is a schematic diagram of a service recommendation method shown in an exemplary embodiment of the present application;

[0024] Figure 7 is a schematic diagram of a service recommendation method shown in an exemplary embodiment of the present application;

[0025] Figure 8 is a flow chart of a service recommendation method shown in an exemplary embodiment of the present application;

[0026] Fig. 9 is a schematic diagram of a service card shown in an exemplary embodiment of the present application;

[0027] Fig.10 is a flow chart of a service recommendation method shown in an exemplary embodiment of the present application;

[0028] Fig.11 is a structural block diagram of a service recommendation device shown in an exemplary embodiment of the present application;

[0029] Fig.12 It is a structural block diagram of a terminal shown in an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0030] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0031] The term "multiple" as used herein refers to two or more than two. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.

[0032] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions. For example, the recommended demand information and multiple candidate services involved in this application are all obtained with full authorization.

[0033] The service recommendation method provided in the embodiment of the present application can be executed by a terminal with computing, processing and display functions. The terminal can be a smart phone, a tablet computer, a smart wearable device, a personal computer, etc. The terminal includes an assistant application, which can obtain the current recommendation demand information of the terminal, and then determine a first service list based on the recommendation demand information, the first service list includes multiple candidate services that meet the recommendation demand corresponding to the recommendation demand information, select at least one target service from the first service list to obtain a second service list, and display the second service list in the service card of the terminal.

[0034] Recommended demand information is used to indicate the terminal's current demand for recommended services; and the recommended demand information can be intent information or event information; intent information is used to reflect the intention of the user using the terminal, and event information is used to indicate the event currently occurring in the terminal. Service refers to the ability to meet the recommended demand provided based on the recommended demand information; for example, if the recommended demand information is intent information, then service refers to the ability to meet the user's needs provided based on the user's intent. The types and quantities of services are diverse; for example, for a user's intention to take a taxi, multiple target services can be A taxi application, B taxi application, and C taxi application, etc.

[0035] Please refer to Figure 1 , which shows a schematic diagram of the implementation environment of the service recommendation method provided by an exemplary embodiment of the present application. The implementation environment includes a terminal 1, which can be any device with service capabilities such as a smart phone, a tablet computer, a smart watch or a smart bracelet. Terminal 1 includes a service discovery system 11 and a service evaluation system 12; the service discovery system 11 is used to perform service discovery on the recommendation demand information to obtain a first service list; the service evaluation system 12 is used to perform quality evaluation and screening on multiple candidate services in the first service list to obtain a second service list.

[0036] Please refer to Figure 2, the service evaluation system 12 includes a first management module, a service evaluation execution framework and multiple evaluation nodes. Among them, the first management module is used to obtain the first service list determined by the service discovery system 11, and send the first service list to the service evaluation execution framework; for example, the first management module can be a service decision management module. As the core carrier solution of the service evaluation system 12, the service evaluation execution framework is responsible for receiving the recommended demand information and its corresponding first service list from the first management module, and then calling multiple target evaluation nodes to perform multi-dimensional evaluation and aggregation on the multiple candidate services included in the first service list to obtain the first resource values ​​of the multiple candidate services; then based on the first quality parameter values ​​of the multiple candidate services, at least one target service is determined from the first service list to obtain the second service list; for example, the service evaluation execution framework can be a service decision execution framework. And, in the case of multiple recommended demand information, the evaluation execution framework classifies the second service list based on the multiple recommended demand information, and then returns it to the caller, which can be an assistant application.

[0037] In a possible implementation, the service evaluation execution framework includes a second management module, a third management module, a fourth management module and a fifth management module; the second management module is used to manage the process of determining the second service list from the first service list; for example, the second management module can be a decision process management module. The third management module is used to manage multiple target evaluation nodes; for example, the third management module can be a decision point management module. The fourth management module is used to manage resource configuration files where multiple target evaluation nodes are located; for example, the fourth management module can be a comprehensive decision logic management module. The fifth management module is used to manage the second service list; for example, the fifth management module can be a decision result management module.

[0038] Please refer to Figure 3 , which shows a flow chart of a service recommendation method provided by an exemplary embodiment of the present application. This embodiment is described by taking the method used in a terminal as an example, and the method may include the following steps:

[0039] Step 301: determine a first service list of recommended demand information, the first service list including a plurality of candidate services that meet the recommended demand corresponding to the recommended demand information, the recommended demand information being used to indicate the current recommended service demand of the terminal.

[0040] The recommendation demand information is used to indicate the terminal's current demand for recommended services; and the recommendation demand information can be intention information or event information; the intention information is used to reflect the intention of the user using the terminal, and the event information is used to indicate the event currently occurring in the terminal. In a possible implementation, this step can be implemented by the following steps (1) to (3), including:

[0041] (1) In response to a service recommendation request, determine recommendation requirement information.

[0042] An assistant application is installed in the terminal, and the assistant application is associated with at least one target application; accordingly, the recommended demand information is the recommended demand information corresponding to the recommended demand of at least one target application. And, when there is a recommended demand for at least one target application, it is determined that a service recommendation request has been received. For example, the assistant application is associated with the air ticket application. When there is a purchased air ticket in the air ticket application, and the time interval between the current time and the departure time of the purchased air ticket is less than the preset time length, it is determined that a service recommendation request has been received, and the recommended demand information is to take a taxi to the departure place of the purchased air ticket.

[0043] In one possible implementation, the recommendation demand information can be structured data or unstructured data; and the terminal can obtain the recommendation demand information with the help of a large language model; thereby achieving support for diversified data sources; and using big data analysis and machine learning to extract valuable recommendation demand information from large-scale data, thereby improving the intelligence level of service recommendations.

[0044] (2) Based on the recommendation requirement information, a plurality of candidate services that meet the recommendation requirement corresponding to the recommendation requirement information are determined through a service discovery system.

[0045] The multiple candidate services include multiple services in the terminal and the server; the terminal includes a service discovery system, which is used to perform service discovery based on the recommended demand information, that is, to determine multiple candidate services that meet the recommended requirements corresponding to the recommended demand information.

[0046] (3) Multiple candidate services are combined into a first service list.

[0047] After the service discovery system determines the first service list, it sends the first service list to the service evaluation system of the terminal, and the service evaluation system executes steps 302-304.

[0048] In another possible implementation, after the terminal determines the first service list, it determines whether there is an evaluation requirement for the first service list; if there is an evaluation requirement for the first service list, step 302 is performed; if there is no evaluation requirement for the first service list, the first service list is directly displayed in the service card. For example, if only one candidate service that meets the recommendation requirement corresponding to the recommendation requirement information is found based on the recommendation requirement information, service evaluation cannot be performed; therefore, it can be determined whether there is an evaluation requirement for the first service list based on the number of candidate services included in the first service list. The process can be: determining the number of candidate services included in the first service list; if the number of candidate services is multiple, determining that there is an evaluation requirement for the first service list; if the number of candidate services is one, determining that there is no evaluation requirement for the first service list.

[0049] For example, see Figure 4 , the input source sends an intent list to the first management module (specifically, it can be a service decision management module), the intent list includes at least one intent information; the first management module sends the intent list to the service discovery system, the service discovery system determines the first service list for each intent information in the intent list based on the intent list, the first service list of the intent information includes multiple candidate services that meet the intent corresponding to the intent information, and sends the first service list of each intent information in the intent list to the first management module. The first management module receives the first service list of each intent information in the intent list, and based on the number of candidate services in the first service list of each intent information, filters the first service list and intent information that do not have evaluation requirements, and sends the remaining intent lists and first service lists to the service evaluation execution framework (specifically, it can be a service decision execution framework).

[0050] Step 302: Based on the recommendation requirement information, a plurality of target evaluation nodes are determined, and the plurality of target evaluation nodes are used to perform quality evaluation on a plurality of candidate services from different evaluation dimensions.

[0051] Multiple target evaluation nodes can be decision points; and multiple target evaluation nodes can be connected to the service evaluation system in a configurable manner, thereby ensuring the flexibility of the service evaluation system. A target evaluation node corresponds to at least one evaluation dimension, and the evaluation dimension is a dimension obtained by abstracting the evaluation factors. In a possible implementation, different recommendation requirement information corresponds to different selection strategy information to select different resource configuration files; accordingly, this step can be implemented by the following steps (1) to (3), including:

[0052] (1) Determine selection strategy information for recommendation requirement information. The selection strategy information is used to represent a strategy for selecting a resource configuration file. Different resource configuration files include different candidate evaluation nodes.

[0053] In one possible implementation, the recommendation demand information carries selection strategy information; accordingly, in this step, the selection strategy information of the recommendation demand information is obtained from the recommendation demand information. In another possible implementation, the recommendation demand information does not carry selection strategy information, but is generated with the help of a large language model, that is, based on the real-time data analysis capability of the large language model, the recommendation demand information is dynamically adjusted to cope with changes in user needs; this adaptability improves the flexibility of the strategy for selecting resource configuration files, so that when the recommendation demand information does not carry selection strategy information, accurate resource configuration files can still be obtained, thereby providing high-quality services.

[0054] Based on business requirements, terminal-side hard constraints (constraints such as the power consumption of the terminal, the network to which the terminal is connected, and the storage space of the terminal), and scalability considerations at the design level, the embodiments of the present application support configuring resource configuration files based on the granularity of input source, configuring resource configuration files based on the granularity of recommended requirement information, configuring resource configuration files based on the granularity of the recommended requirement information set where the recommended requirement information is located, and configuring a default resource configuration file that does not distinguish between the recommended requirement information, thereby achieving the determination of the resource configuration file based on the input source, the recommended requirement information, and the recommended requirement information set where the recommended requirement information is located, or directly obtaining the default resource configuration file.

[0055] For example, the selection policy information is used to indicate the selection of a resource profile based on an input source; or, the selection policy information is used to indicate the selection of a resource profile based on recommended demand information; or, the selection policy information is used to indicate the selection of a resource profile based on a recommended demand information set in which the recommended demand information is located; or, the selection policy information is used to indicate the selection of a default resource profile. The default resource profile is a resource profile that does not distinguish between recommended demands and services, and the default resource profile is a catch-all resource profile.

[0056] The priority of resource profiles in the resource profile set is divided from high to low into resource profiles based on recommended requirements, resource profiles based on recommended requirement sets, and default resource profiles. Taking the recommended requirement information as the intent information and intentId as the intent identifier as an example, since intentId itself is an integer, if the policy information is selected as 0, it means that the default resource profile is selected; for example, metis_0_sdl.json, which identifies the intent input by Metis is used to select the default resource profile. If the policy information is selected as [10,10000], it means that the resource profile is selected based on the granularity of the intent set; if the policy information is selected as other values, it means that the resource profile is selected based on the granularity of the intent.

[0057] Taking the recommended demand information as the intent information and action as the intent identifier as an example, since action is of string type, if the policy information is selected as "default", it means selecting the default resource configuration file; if the policy information is selected as intentset, it means selecting the resource configuration file based on the intent set as the granularity; if the policy information is selected as other values, it means selecting the resource configuration file based on the intent as the granularity; for example, systemapp_pantanal.intent.travel.ESCORT_SERVICE_sdl.json means selecting the resource configuration file based on the intent intent.travel.ESCORT_SERVICE as the granularity.

[0058] One thing that needs to be explained is that the number of resource configuration files with intent as the granularity will increase as the number of preferred configurations increases; when the requirements for precision are not too high, the number of configuration files with intent sets as the granularity will increase slowly; and the default resource configuration files that do not distinguish between intents will be updated less frequently.

[0059] The target evaluation node is based on the resource configuration file. In compliance with the configuration specification, the business party can configure the resource configuration file by itself, and support online distribution of the resource configuration file, that is, the resource configuration file set supports configurability. Among them, the steps of configuring the resource configuration file set can be: obtaining the candidate resource file to be configured, adding the candidate resource configuration file to the resource configuration file set, so as to realize flexible configuration of the resource configuration file.

[0060] In the embodiment of the present application, the resource configuration file supports configurability, thereby allowing the business party to flexibly set the resource configuration file according to specific scenarios and requirements. This flexibility enables the solution to quickly adapt to market changes and the diversity of recommendation requirements, improving the scalability and application scope of the solution. In addition, the resource configuration file supports online distribution, so that the resource configuration file can respond to changes in recommendation requirements in real time. This instant update capability greatly enhances the agility of service recommendations and ensures that users can always obtain the most suitable target service.

[0061] Before adding a candidate resource configuration file to a resource configuration file collection, first determine whether the resource configuration file collection contains a resource configuration file with the same file name; if the resource configuration file collection includes a resource configuration file with the same file name, directly overwrite the resource configuration file with the same file name with the candidate resource configuration file, thereby achieving the goal that the resource configuration file with the same file name does not maintain historical versions but directly overwrites them, thereby reducing the number of resource configuration file collections and further improving the efficiency of selecting candidate resource configuration files.

[0062] In a possible implementation, the resource configuration file supports online distribution; accordingly, the step of obtaining the configured candidate resource configuration file can be: displaying a first configuration interface, the first configuration interface displays multiple resource configuration files obtained from the server; obtaining a candidate resource configuration file selected from the multiple resource configuration files.

[0063] In another possible implementation, the resource configuration file supports online configuration of evaluation dimensions; accordingly, the step of obtaining the configured candidate resource configuration file may be: displaying a second configuration interface, where the second configuration interface displays multiple evaluation dimensions; obtaining a selected candidate evaluation dimension from the multiple evaluation dimensions, configuring the evaluation node based on the selected candidate evaluation dimension, and obtaining the configured candidate resource configuration file.

[0064] (2) Based on the selection strategy information, determine a target resource profile from the resource profile set, where the target resource profile includes multiple candidate evaluation nodes.

[0065] The resource configuration file set includes at least one resource configuration file; this step can be implemented by the following steps (2-1) to (2-4), including:

[0066] (2-1) In the case where selection strategy information is used to indicate selection of a resource profile based on an input source, a target resource profile of the same input source is determined from a resource profile set based on an input source of the recommended requirement information, where the input source is used to indicate a source of the recommended requirement information.

[0067] The file name of the resource configuration file in the resource configuration file set includes the input source; for example, the naming rule of the resource configuration file is: input source_sdl.json; when the recommended requirement information is intent information, the input source can be the intent input source. In a possible implementation, based on the input source of the recommended requirement information, the step of determining the target resource configuration file of the same input source from the resource configuration file set can be: based on the input source of the recommended requirement information, determining from the resource configuration file set a resource configuration file whose file name includes the input source, and determining the determined resource configuration file as the target resource configuration file.

[0068] The file name of the resource configuration file is configurable; and, before this step, the terminal sets the file name of the resource configuration file, and the process may be: when the resource configuration file is set to adapt to the recommended requirement information of a certain input source, the terminal obtains the input source of the recommended requirement information and adds the input source to the resource configuration file. In addition, the file name of the resource configuration file is also extensible; when the resource configuration file is set to also adapt to the recommended requirement information of other input sources, the terminal obtains the other input sources and adds the other input sources to the file name of the resource configuration file. For example, when the resource configuration file is adapted to the Metis input source, the terminal adds Metis to the file name of the resource configuration file; when the resource configuration file is adapted to the SystemApp input source, the terminal adds SystemApp to the file name of the resource configuration file.

[0069] In a possible implementation, the file name of the resource configuration file may be set in the first management module when the input source inputs the recommendation requirement. At this time, the terminal obtains the input source and adds the input source to the file name of the resource configuration file.

[0070] (2-2) When the selection strategy information is used to indicate the selection of a resource configuration file based on the recommended requirement information, a target resource configuration file corresponding to the recommended requirement information is determined from a resource configuration file set based on the recommended requirement corresponding to the recommended requirement information, and different recommended requirements correspond to different resource configuration files.

[0071] The file name of the resource configuration file in the resource configuration file set includes the first identification information of the recommended requirement corresponding to the recommended requirement information, and the first identification information is used to uniquely identify the recommended requirement; for example, the naming rule of the resource configuration file is: input source_first identification information of the recommended requirement_sdl.json. In a possible implementation, based on the recommended requirement corresponding to the recommended requirement information, the step of determining the target resource configuration file corresponding to the recommended requirement information from the resource configuration file set may be: based on the first identification information of the recommended requirement corresponding to the recommended requirement information, determining the resource configuration file including the first identification information from the resource configuration file set, and determining the determined resource configuration file as the target resource configuration file.

[0072] The file name of the resource configuration file is configurable; and, before this step, the terminal sets the file name of the resource configuration file. The process may be: when the resource configuration file is set to adapt to a certain recommended requirement information, the terminal obtains the first identification information of the recommended requirement corresponding to the recommended requirement information, and adds the first identification information of the recommended requirement corresponding to the recommended requirement information in the resource configuration file.

[0073] When the recommendation demand information is intent information, the first identification information of the recommendation demand can be the intent identifier. Multiple applications in the terminal interact and communicate with each other through Intent, that is, the recommendation demand information is described through Intent. Intent describes the action, data involved, and additional information of an operation; Intent includes two attributes, namely intentId or action, intentId is a unique identifier used to represent Intent; action is used to describe the action that Intent will perform; therefore, the intent identifier can be intentId or action. And, in the case where Intent includes the intentId attribute, the intent identifier is intentId. When the transfer Intent does not include intentID but includes action, the intent identifier is action.

[0074] (2-3) When the selection strategy information is used to indicate the selection of a resource configuration file based on a recommended requirement information set in which the recommended requirement information is located, the recommended requirement information set in which the recommended requirement information is located is determined, and based on the recommended requirement information set, a target resource configuration file corresponding to the recommended requirement information set is determined from the resource configuration file set, different recommended requirement sets correspond to different resource configuration files, and the resource configuration file corresponding to the recommended requirement set is adapted to at least one recommended requirement included in the recommended requirement set.

[0075] The file name of the resource configuration file in the resource configuration file set also includes the second identification information of the recommended requirement set, and the second identification information is used to uniquely identify the recommended requirement set. In the case where the recommended requirement information is intent information, the recommended requirement set can be an intent set.

[0076] In one possible implementation, based on the recommended requirement information set, the step of determining a target resource configuration file corresponding to the recommended requirement information set from a resource configuration file set may be: based on the second identification information of the recommended requirement information set, determining a resource configuration file including the second identification information from the resource configuration file set, and determining the determined resource configuration file as the target resource configuration file.

[0077] (2-4) When the selection policy information is used to indicate the selection of a default resource profile, the default resource profile is determined from the resource profile set, and the default resource profile is determined as the target resource profile.

[0078] The default resource profile is a catch-all resource profile that does not distinguish between recommended requirements; and the default resource profile includes at least one of a first default resource profile corresponding to the terminal side of the terminal and a second default resource profile corresponding to the cloud side of the server. Moreover, in the case where the selection policy information is used to indicate the selection of a default resource profile, the selection policy information may also indicate an input source of the default resource profile, which is used to indicate the source of the default resource profile. For example, in the case where the selection policy information indicates that the input source is the terminal side, a first default resource profile is determined from a resource profile set, and the first default resource profile is determined as a target resource profile; for another example, in the case where the selection policy information indicates that the input source is the cloud side, a second default resource profile is determined from a resource profile set, and the second default resource profile is determined as a target resource profile; for another example, in the case where the selection policy information indicates that the input source is the terminal side and the cloud side, a first default resource profile and a second default resource profile are determined from a resource profile set, and the first default resource profile and the second default resource profile are determined as target resource profiles.

[0079] For example, when a system application (recording application or call application, etc.) is used as an input source, the recommendation requirements involved only require the participation of the terminal-side evaluation node, so only a resource configuration file that does not distinguish between recommendation requirements and is purely terminal-side is required. When Metis1 is used as an input source, the recommendation requirements involved may require the participation of terminal-side evaluation nodes and cloud-side evaluation nodes; then, based on business needs, it is determined that a resource configuration file that does not distinguish between recommendation requirements and services is required that integrates terminal and cloud. When Metis2 is used as an input source, the recommendation requirements involved are distinguished based on the recommendation requirement set, and the resource configuration file is selected based on the recommendation requirement information set where the recommendation requirement information is located. Please refer to Figure 5 The fourth management module (specifically, the comprehensive decision-making logic management module) determines the selection policy information of Metis1; when the selection policy information of Metis1 is used to indicate the selection of the default resource configuration file, the target resource configuration file of Metis1 is determined to be the default resource configuration file metis1_default_sdl.json; the selection policy information of the system application is determined, and when the selection policy information of the system application is used to indicate the selection of the default resource configuration file, the target resource configuration file of the system application is determined to be the default resource configuration file systemapp_default_sdl.json; the selection policy information of Metis2 is determined, and when the selection policy information of Metis2 is used to indicate the selection of a resource configuration file based on the recommended demand information set in which the recommended demand information is located, the target resource configuration file of Metis2 is determined to be the resource configuration file metis2_intentset_sdl.json that supports dynamic distribution.

[0080] (3) Determine multiple target evaluation nodes from the target resource configuration file.

[0081] In a possible implementation, multiple target evaluation nodes are determined based on the available status and priority levels of multiple candidate evaluation nodes included in the target resource configuration file; accordingly, step (3) may be: determining the available status and priority levels of multiple candidate evaluation nodes, the available status of the candidate evaluation node is used to indicate whether the candidate evaluation node is available, and the available status of the candidate evaluation node may be available or unavailable; the priority level of the candidate evaluation node is used to indicate the importance of the candidate evaluation node; based on the available status and priority levels of the multiple candidate evaluation nodes, determining a first preset number of candidate evaluation nodes with an available status and the highest priority level from the multiple candidate evaluation nodes, to obtain multiple target evaluation nodes.

[0082] In another possible implementation, multiple target evaluation nodes are determined based on the available status of multiple candidate evaluation nodes included in the target resource configuration file; accordingly, step (3) may be: determining the available status of multiple candidate evaluation nodes, where the available status of the candidate evaluation nodes may be available or unavailable; based on the available status of the multiple candidate evaluation nodes, determining multiple candidate evaluation nodes with an available status from the multiple candidate evaluation nodes to obtain multiple target evaluation nodes.

[0083] Please refer to Figure 5 , the fourth management module (specifically the comprehensive decision logic management module) maintains the policy selection information; the service evaluation framework entry sends a service evaluation request to the fourth management module, and the service evaluation request carries the intent information; the fourth management module selects the target resource configuration file based on the selection policy information of the intent information; for example, when the intent information is configured with the selection policy information, the target resource configuration file is selected based on the selection policy information; when the intent information is not configured with the selection policy information, the default resource configuration file is used as the target resource configuration file. After obtaining the target resource configuration file, multiple candidate evaluation nodes are obtained from the target resource configuration file, and the available status of the multiple candidate evaluation nodes is queried; multiple target evaluation nodes with available status are selected according to the available status of the multiple candidate evaluation nodes.

[0084] Among them, for any candidate evaluation node, the step of determining the available status of the candidate evaluation node can be: based on the SPI interface of the candidate evaluation node, sending a business test request to the candidate evaluation node; when a business test response returned by the evaluation candidate node is received within a preset time after sending the business test request, determining that the available status of the candidate evaluation node is available; when no business test response is received within a preset time after sending the business test request, determining that the available status of the candidate evaluation node is unavailable.

[0085] After the terminal determines that the candidate evaluation node is unavailable, it stores the node identifier of the candidate evaluation node in the state storage file and sets the expiration timestamp of the candidate evaluation node. The state storage file can be a shared preference; when the expiration timestamp arrives, the terminal deletes the node identifier of the candidate evaluation node from the state storage file.

[0086] Different resource configuration files correspond to different comprehensive evaluation logics; if the recommendation requirements need to define multiple comprehensive evaluation logics due to complex changes in the business, and the evaluation dimensions corresponding to the evaluation nodes involved in different comprehensive evaluation logics are also quite different, then integrating all the evaluation dimensions involved in different comprehensive evaluation logics into one resource configuration file will cause the resource configuration file to be difficult to maintain and the scalability to add new evaluation dimensions to be poor due to the large number of combinations of different evaluation dimensions. In an embodiment of the present application, different resource configuration files correspond to different comprehensive evaluation logics, so that it is easy to maintain the resource configuration file and to expand the evaluation dimensions. Among them, the expansion of the evaluation dimensions can be carried out from two aspects; the first aspect: the comprehensive evaluation logic can be expanded, and the resource configuration file corresponding to the comprehensive evaluation logic can be added; the second aspect: the comprehensive evaluation logic can be expanded to select a smaller granularity, that is, the resource configuration file is selected based on a smaller granularity.

[0087] The resource configuration file (SDL) is also configured with multiple fields, and the multiple field lists are shown in Table 1 below:

[0088] Table 1

[0089]

[0090]

[0091] For example, please refer to Figure 4 The service evaluation execution framework includes a service evaluation core logic, and multiple target evaluation nodes are selected based on the service evaluation core logic. For each intent information in the intent list, a target resource configuration file is selected based on the selection strategy information of the intent information; and a target evaluation node corresponding to the intent information is selected from the target resource configuration file. Since there are multiple intent information in the intent list, the above steps are performed for each intent information to determine the target evaluation node corresponding to each intent information; since the target evaluation nodes corresponding to each intent information may have the same target evaluation node, for each target evaluation node, the intent information of the target evaluation node and the first service list of the intent information are integrated.

[0092] Step 303: For any candidate service, multiple first quality parameter values ​​of the multiple candidate services are determined based on multiple evaluation dimensions corresponding to multiple target evaluation nodes, and different first quality parameter values ​​of the candidate services are determined based on different evaluation dimensions. The first quality parameter value of the candidate service is used to indicate the degree to which the candidate service satisfies the recommendation requirement corresponding to the recommendation requirement information.

[0093] In a possible implementation, this step can be implemented by following the steps (1) to (3), including:

[0094] (1) Based on multiple evaluation dimensions corresponding to multiple target evaluation nodes, an evaluation dimension set is determined, wherein the target evaluation node corresponds to at least one evaluation dimension, and the evaluation dimension set includes the multiple evaluation dimensions corresponding to the multiple target evaluation nodes.

[0095] For example, multiple target evaluation nodes include evaluation node 1 and evaluation node 2, evaluation node 1 corresponds to evaluation dimension 1 and evaluation dimension 2, and evaluation node 2 corresponds to evaluation dimension 3, then the determined evaluation dimension set includes evaluation dimension 1, evaluation dimension 2, and evaluation dimension 3.

[0096] After determining the evaluation dimension set, determine the available status of each evaluation dimension in the evaluation dimension set, and the available status of the evaluation dimension is used to indicate whether the evaluation dimension is available; when there are evaluation dimensions in the evaluation dimension set whose available status is unavailable, delete the evaluation dimensions in the evaluation dimension set, and then adjust the weights of the remaining evaluation dimensions in the evaluation dimension set.

[0097] (2) For any evaluation dimension in the evaluation dimension set, call the target evaluation node, and perform quality evaluation on the candidate service based on the evaluation dimension to obtain a second quality parameter value.

[0098] The terminal sends a first service list to the target evaluation node based on the SPI interface of the target evaluation node; based on the SPI interface of the target evaluation node, receives a second quality parameter value obtained by the target evaluation node performing a quality evaluation on multiple candidate services in the first service list based on at least one evaluation dimension corresponding to the target evaluation node. It should be noted that, in the event that the target evaluation node performs an abnormal quality evaluation on any candidate service based on at least one evaluation dimension, the target evaluation node returns an error code to the terminal based on the SPI interface, and the error code is used to indicate the abnormal cause of the quality evaluation.

[0099] For example, please refer to Figure 4For each target evaluation node, the SPI interface of the target evaluation node is called, and the first service list is passed to the target evaluation node, and the target evaluation node performs quality evaluation on the candidate services in the first service list; when the target evaluation node is in a normal state, the second quality parameter value of the candidate service is returned to the service evaluation execution framework; when the target evaluation node is in an abnormal state, an error code is returned to the service execution framework.

[0100] (3) Based on the weights of multiple evaluation dimensions in the evaluation dimension set, multiple second quality parameter values ​​of the candidate service are weighted and summed to obtain the first quality parameter value of the candidate service. The multiple second service quality parameter values ​​of the candidate service are obtained by performing quality evaluation on the candidate service based on multiple evaluation dimensions.

[0101] This step can be achieved by the following formula:

[0102] Formula 1: finalcScore=∑W i *DD i

[0103] Among them, finalcScore represents any candidate service, W i represents the weight of the i-th evaluation dimension in the evaluation dimension set, DD i represents the second quality parameter value of the candidate service determined by the target evaluation node based on the i-th evaluation dimension; ∑(W i )=1, that is, the sum of the weights of multiple evaluation dimensions in the evaluation dimension set is 1.

[0104] The resource configuration file is strongly related to the business, and business personnel are required to select the target evaluation node and the evaluation dimension corresponding to the target evaluation node included in the resource configuration file, and set the weight of the evaluation dimension. This application supports the configurability and dynamic update of the weight of the evaluation dimension; and the weight of the evaluation dimension corresponding to the target evaluation node is a certain weight in the weight range given by the SPI specification.

[0105] For example, evaluation node 1 performs quality evaluation on the candidate service based on evaluation dimension 1 and evaluation dimension 2 to obtain second quality parameter values ​​of 10 and 20 respectively, and evaluation node 2 performs quality evaluation on the candidate service based on evaluation dimension 3 to obtain a second quality parameter value of 30. The weights of evaluation dimension 1, evaluation dimension 2, and evaluation dimension 3 are 0.5, 0.2, and 0.3 respectively, then based on formula one, the second quality parameter values ​​10, 20, and 30 obtained by quality evaluation based on evaluation dimension 1, evaluation dimension 2, and evaluation dimension 3 are weighted and summed to obtain the first quality parameter value of the candidate service of 18.

[0106] For example, please refer to Figure 6The fourth management module calls the SPI interface of the target evaluation node, obtains the second quality parameter value obtained by evaluating the candidate service according to each evaluation dimension of the target evaluation node, and obtains the first quality parameter value of the candidate service by weighted summing up the second quality parameter value of the candidate service according to the weight of each evaluation dimension. For each intent information, the first service list of the intent information is sorted in descending order based on the first quality parameter value, and the first second preset number of candidate services in the first service list (which can be called target services at this time) are composed of a second service list. The second service list is the final decision result of the intent information, and then the final decision result is returned to the service evaluation framework entrance.

[0107] In the embodiment of the present application, a multi-dimensional evaluation and weighting mechanism are used to ensure that the quality evaluation of multiple candidate services is more comprehensive and objective. This method reduces the deviation that may be caused by a single evaluation dimension and improves the accuracy of service recommendations. This method can also optimize service selection in different scenarios to meet diverse user needs.

[0108] In one possible implementation, after multiple target evaluation nodes are determined based on the recommended demand information, step 304 can be directly executed to determine the first quality parameter values ​​of multiple candidate services based on the multiple evaluation dimensions corresponding to the multiple target evaluation nodes, thereby improving the efficiency of determining the first quality parameter values ​​of multiple candidate services. In another possible implementation, after multiple target evaluation nodes are determined based on the recommended demand information, the service level agreements (Service Level Agreements, SLAs) of the multiple target evaluation nodes are first verified. If the SLA verification passes, step 304 is executed to determine the first quality parameter values ​​of multiple candidate services based on the multiple evaluation dimensions corresponding to the multiple target evaluation nodes. The process can be:

[0109] The service level agreements (SLAs) of multiple target evaluation nodes are detected; when it is detected that the SLAs of multiple target evaluation nodes meet the protocol requirements, first quality parameter values ​​of multiple candidate services are determined based on multiple evaluation dimensions corresponding to the multiple target evaluation nodes. When it is detected that the SLAs of multiple target evaluation nodes do not meet the protocol requirements, the target evaluation nodes are discarded, and first quality parameter values ​​of multiple candidate services are determined based on multiple evaluation dimensions corresponding to the remaining target evaluation nodes.

[0110] For example, see Figure 7, multiple target evaluation nodes include target evaluation node 1, target evaluation node 2 and target evaluation node 3, target evaluation node 1 corresponds to evaluation dimension A, target evaluation node 2 corresponds to evaluation dimension B and evaluation dimension C, and target evaluation node 3 corresponds to evaluation dimension D; when the SLA of target evaluation node 1, the SLA of target evaluation node 2 and the SLA of target evaluation node 3 all meet the protocol requirements, the first quality parameter values ​​of multiple candidate services are determined based on the evaluation dimensions A, B, C and D corresponding to the target evaluation nodes 1, 2 and 3.

[0111] In an embodiment of the present application, the service evaluation execution framework defines a unified specification for access of target evaluation nodes; therefore, when multiple target evaluation nodes meet the specification (that is, meet the protocol requirements), multiple target evaluation nodes are called to evaluate the first quality parameter values ​​of multiple candidate services, and when multiple target evaluation nodes do not meet the specification, multiple evaluation nodes are not called, thereby improving the reliability of the first quality parameter value.

[0112] Step 304: Based on the first quality parameter values ​​of the plurality of candidate services, determine at least one target service from the first service list to obtain a second service list.

[0113] Based on the first quality parameter values ​​of multiple candidate services, a second preset number of candidate services with the highest first quality parameter values ​​are determined from the first service list, the determined second preset number of candidate services are determined as target services, and the second preset number of target services are combined into a second service list, which is a service list corresponding to the recommended demand information.

[0114] In one possible implementation, the step of forming a second preset number of target services into a second service list may be: sorting the second preset number of target services in descending order based on the first quality parameter values ​​of the second preset number of target services, and forming the sorted second preset number of target services into a second service list.

[0115] In another possible implementation, based on the first quality parameter values ​​of multiple candidate services, the step of determining a second preset number of candidate services with the highest first quality parameter values ​​from the first service list may be: sorting the multiple candidate services in order from high to low based on the first quality parameter values ​​of the multiple candidate services, and selecting a second preset number of target services with the highest sorting among the sorted multiple candidate services.

[0116] For example, please refer to Figure 4The service evaluation execution framework performs weighted summation of the second quality parameter values ​​of the candidate services according to the weights of multiple evaluation dimensions to obtain the first quality parameter value of the candidate service, and the first service list is sorted in descending order based on the first quality parameter value, and the first second preset number of candidate services in the first service list (which can be called target services at this time) form the second service list.

[0117] In an embodiment of the present application, an efficient sorting algorithm is used to sort multiple candidate services in descending order, so as to quickly select a target service. This high-performance processing capability ensures that users can quickly obtain feedback after submitting a request, significantly reducing response time.

[0118] In a possible implementation, the recommended demand information also carries business data; the business data is the business data of the service used by the recommended demand information; for example, if the recommended demand information is to take a taxi, the business data may be the taxi time, the starting point and the destination, etc. For another example, if the recommended demand information is to order takeout, the business data may be the delivery address and contact information of the takeout, etc. Correspondingly, after the terminal determines the target service, it generates service information of the target service based on the business data and the target service, and displays the service information on the service card of the terminal. For example, if the recommended demand information is to take a taxi, the business data may be the taxi time, the starting point and the destination, and the target service is A taxi application, then the service information is a taxi order generated based on A taxi application, the time of the taxi order is the taxi time carried by the business data, and the starting point and the destination are the starting point and the destination carried by the business data, respectively.

[0119] In another possible implementation, the recommendation demand information also carries style information of the service card, and the style information is used to indicate the display style of the service card; accordingly, the terminal sets the style of the service card based on the style information, thereby personalizing the style of the service card based on user needs.

[0120] In a possible implementation, when there are multiple recommended demand information, the second service list of each recommended demand information can be determined based on steps 301-304, and then the target services in the second service list of each recommended demand information are deduplicated. In the second service list, multiple target services are arranged in order from high to low based on the first quality parameter value; accordingly, the step of deduplicating the target services in the second service list of each recommended demand information can be: for the same target service among the multiple target services in the second service list of each recommended demand information, the target service list is determined based on the order of the target service in the multiple second service lists, the target service in the target service list is retained with the lowest (most ranked) target service in the target service list, and the target service in other second service lists is deleted.

[0121] In another possible implementation, after deleting the target service in the other second service list, the number of target services in the other second service list may be relatively small. At this time, a default service can be added to the other second service list. The default service is a fallback service, thereby ensuring that a sufficient number of target services are recommended to users.

[0122] The above-mentioned deduplication process is performed after the second service list is determined; in another possible implementation, after the first quality parameter values ​​of the multiple candidate services in the first service list are determined, the candidate services in the first service list of each recommended demand information may be deduplicated based on the first quality parameters of the multiple candidate services of each recommended demand information. Among them, the step of deduplicating the candidate services in the first service list of each recommended demand information based on the first quality parameters of the multiple candidate services of each recommended demand information may be: for the same service among the multiple candidate services in the first service list of each recommended demand information, based on the first quality parameter values ​​corresponding to the service in the multiple first service lists, determine the target service list, the service in the target service list has the highest first quality parameter value, retain the service in the target service list, and delete the service in other first service lists.

[0123] For example, if the first service list of the first recommended demand information and the first service list of the second recommended demand information both include service A, and the first quality parameter value of service A in the first service list of the first recommended demand information is greater than the first quality parameter value of service A in the second service list of the second recommended demand information, then service A in the first service list of the first recommended demand information is retained, and service A in the second service list of the second recommended demand information is deleted.

[0124] For example, please refer to Figure 4 , the service evaluation execution framework sends the evaluation result to the first management module, and the evaluation result includes the intent list and the second service list for each intent information in the intent list. The first management module receives the intent list and the second service list for each intent information in the intent list, and then performs processes such as service deduplication and adding backup services, and finally displays the second service list in the service card of the terminal. In addition, the terminal also maintains the life cycle of the target service in the second service list. The process is: setting a life cycle for the target service in the second service list, and deleting the target service from the second service list when the life cycle of the target service is reached.

[0125] After the terminal determines the second service list, the second service list can be displayed in the service card of the terminal. The schematic diagram of the service card is as shown in FIG. Figure 8 shown.

[0126] In the embodiment of the present application, a comprehensive evaluation of multiple candidate services is performed through multiple evaluation dimensions, which can effectively integrate the results of different evaluation dimensions. This comprehensive processing method can ensure the comprehensiveness of the evaluation, so that the target service finally determined is more in line with the actual needs of the user, thereby improving the accuracy of the target service recommendation and further improving the service recommendation effect; that is, the embodiment of the present application can screen out the most suitable services for users to enhance the user experience.

[0127] Moreover, after the user submits a service recommendation request, the embodiment of the present application can obtain the target service that meets the recommendation requirements in a relatively short time, that is, the embodiment of the present application can quickly and accurately match the recommendation requirements and the target service, thereby enhancing the convenience of using the target service, thereby improving the user's satisfaction and experience. In addition, the efficient evaluation algorithm of the present application promotes the digital transformation and intelligent development of the service industry.

[0128] Please refer to Fig. 9 , which shows a flow chart of a service recommendation method provided by an exemplary embodiment of the present application. This embodiment is described by taking the method used in a terminal as an example, and the method may include the following steps:

[0129] Step 901: determine a first service list of recommended demand information, the first service list including a plurality of candidate services that meet the recommended demand corresponding to the recommended demand information, the recommended demand information being used to indicate the current recommended service demand of the terminal.

[0130] In some embodiments, this step is the same as step 301 and will not be repeated here.

[0131] Step 902: Based on the recommendation requirement information, a plurality of target evaluation nodes are determined, and the plurality of target evaluation nodes are used to perform quality evaluation on a plurality of candidate services from different evaluation dimensions.

[0132] In some embodiments, this step is the same as step 302 and will not be repeated here.

[0133] Step 903: For any candidate service, multiple first quality parameter values ​​of the candidate service are determined based on multiple evaluation dimensions corresponding to multiple target evaluation nodes, and different first quality parameter values ​​of the candidate service are determined based on different evaluation dimensions. The first quality parameter value of the candidate service is used to indicate the degree to which the candidate service satisfies the recommendation requirement corresponding to the recommendation requirement information.

[0134] In some embodiments, this step is the same as step 303 and will not be repeated here.

[0135] Step 904: Based on the first quality parameter values ​​of the plurality of candidate services, determine at least one target service from the first service list to obtain a second service list.

[0136] In some embodiments, this step is the same as step 304 and will not be described again herein.

[0137] Step 905: Obtain first adjustment reference information, where the first adjustment reference information is used to adjust the evaluation dimension of the target evaluation node, and the first adjustment reference information includes at least one of first feedback information based on the target service and real-time data of the terminal, where the real-time data is data currently generated by the terminal.

[0138] The first adjustment reference information may also include historical service data, which is data of services used in the past. In addition, the terminal may obtain the first adjustment reference information with the help of a large language model, thereby adjusting the evaluation dimension according to real-time data, historical service data, and user feedback through the large language model. As the usage time increases, the service recommendation capability will continue to improve and can better adapt to changes in user preferences.

[0139] Step 906: Based on the first adjustment reference information, adjust the evaluation dimension of the target evaluation node.

[0140] The evaluation dimension of the target evaluation node is adjusted with the help of a large language model; accordingly, this step may be: inputting the first adjustment reference information and multiple evaluation dimensions of the target evaluation node into the large language model, and outputting the selected evaluation dimension of the target evaluation node, wherein the selected evaluation dimension is the evaluation dimension of the adjusted target evaluation node.

[0141] In a possible implementation, the evaluation dimension supports configurability, and the evaluation dimension of the target evaluation node can also be increased; accordingly, this step can be: inputting the first adjustment reference information and the target evaluation node into the large language model to input the newly added evaluation dimension, and using the newly added evaluation dimension as the evaluation dimension of the adjusted target evaluation node.

[0142] In the embodiments of the present application, there are many factors that affect the service evaluation; and different evaluation dimensions may be added later; therefore, the present application is implemented based on a flexibly expandable service evaluation execution framework, which is not coupled to any specific evaluation dimension.

[0143] Step 907: Adjust the second service list based on the adjusted evaluation dimension of the target evaluation node.

[0144] Based on the evaluation dimension of the adjusted target evaluation node, the third quality parameter values ​​of the plurality of candidate services are determined, and the third quality parameter values ​​of the candidate services are used to indicate the degree to which the candidate services meet the recommendation requirements corresponding to the recommendation requirement information; based on the third quality parameter values ​​of the plurality of candidate services, at least one target service is re-determined from the first service list, and the re-determined at least one target service is formed into an updated second service list. After the terminal determines the updated second service list, the updated second service list is displayed in the service card of the terminal.

[0145] In an embodiment of the present application, by comprehensively analyzing the evaluation dimensions of multiple target evaluation nodes, the reasonable allocation and efficient utilization of resources such as target evaluation nodes are ensured, thereby reducing resource waste; and, service recommendations are dynamically adjusted according to real-time data and user needs, so that the adjusted target services are more in line with user needs, thereby not only improving the operational efficiency of the service provider, but also continuously optimizing service quality and promptly responding to changes in user needs.

[0146] Please refer to Fig.10 , which shows a flow chart of a service recommendation method provided by an exemplary embodiment of the present application. This embodiment is described by taking the method used in a terminal as an example, and the method may include the following steps:

[0147] Step 1001: Determine a first service list of recommended demand information, where the first service list includes a plurality of candidate services that meet the recommended demand corresponding to the recommended demand information, where the recommended demand information is used to indicate the current recommended service demand of the terminal.

[0148] In some embodiments, this step is the same as step 301 and will not be repeated here.

[0149] Step 1002: Based on the recommendation requirement information, a plurality of target evaluation nodes are determined, and the plurality of target evaluation nodes are used to perform quality evaluation on a plurality of candidate services from different evaluation dimensions.

[0150] In some embodiments, this step is the same as step 302 and will not be repeated here.

[0151] Step 1003: For any candidate service, multiple first quality parameter values ​​of the candidate service are determined based on multiple evaluation dimensions corresponding to multiple target evaluation nodes, and different first quality parameter values ​​of the candidate service are determined based on different evaluation dimensions. The first quality parameter value of the candidate service is used to indicate the degree to which the candidate service satisfies the recommendation requirement corresponding to the recommendation requirement information.

[0152] In some embodiments, this step is the same as step 303 and will not be repeated here.

[0153] Step 1004: Based on the first quality parameter values ​​of the plurality of candidate services, determine at least one target service from the first service list to obtain a second service list.

[0154] In some embodiments, this step is the same as step 304 and will not be described again herein.

[0155] Step 1005: Obtain second adjustment reference information, where the second adjustment reference information is used to adjust multiple target evaluation nodes, and the second adjustment reference information includes at least one of second feedback information based on the target service and business change information of the terminal, where the business change information is used to indicate business change status of the terminal.

[0156] The terminal can obtain the second adjustment reference information by means of the large language model, so as to adjust the evaluation node according to user feedback and business changes through the large language model.

[0157] Step 1006: re-determine multiple target evaluation nodes based on the second adjustment reference information.

[0158] Re-determine multiple target evaluation nodes with the help of a large language model; accordingly, this step can be: input the second adjustment reference information and a set of resource configuration files into the large language model, output the selected target resource configuration file, determine multiple target evaluation nodes from the selected target resource configuration files, and the determined target evaluation nodes are the re-determined multiple target evaluation nodes.

[0159] Step 1007: Adjust the second service list based on the re-determined multiple target evaluation nodes.

[0160] Based on the multiple evaluation dimensions corresponding to the re-determined multiple target evaluation nodes, the fourth quality parameter values ​​of the multiple candidate services are determined, and the fourth quality parameter values ​​of the candidate services are used to indicate the degree to which the candidate services meet the recommendation requirements corresponding to the recommendation requirement information; based on the fourth quality parameter values ​​of the multiple candidate services, at least one target service is re-determined from the first service list, and the at least one re-determined target service is formed into an updated second service list. After the terminal determines the updated second service list, the updated second service list is displayed in the service card of the terminal.

[0161] In the embodiment of the present application, the target evaluation node can be quickly adjusted according to market changes and user feedback. This flexibility enables the solution to adapt to the rapidly changing business environment and maintain competitive advantage. In addition, the solution can also adapt to different business scenarios and business needs; whether in the taxi field, food delivery field or other service fields, the solution can quickly adjust the target evaluation node to meet the diverse user needs.

[0162] Please refer to Fig.11 , which shows a structural block diagram of a service recommendation device provided by an exemplary embodiment of the present application. The device includes:

[0163] A first determining module 1101 is used to determine a first service list of recommended demand information, wherein the first service list includes a plurality of candidate services that meet the recommended demand corresponding to the recommended demand information, wherein the recommended demand information is used to indicate the current recommended service demand of the terminal;

[0164] A second determination module 1102 is used to determine a plurality of target evaluation nodes based on the recommendation requirement information, wherein the plurality of target evaluation nodes are used to perform quality evaluation on the plurality of candidate services from different evaluation dimensions;

[0165] A third determination module 1103 is used to determine, for any candidate service, multiple first quality parameter values ​​of the candidate service based on multiple evaluation dimensions corresponding to the multiple target evaluation nodes, and determine different first quality parameter values ​​of the candidate service based on different evaluation dimensions, where the first quality parameter value of the candidate service is used to indicate the degree to which the candidate service satisfies the recommendation requirement corresponding to the recommendation requirement information;

[0166] The fourth determining module 1104 is configured to determine at least one target service from the first service list based on the first quality parameter values ​​of the plurality of candidate services to obtain a second service list.

[0167] In another possible implementation, the second determination module 1102 is used to determine selection strategy information of the recommended demand information, the selection strategy information is used to represent a strategy for selecting a resource profile, and different resource profiles include different candidate evaluation nodes; based on the selection strategy information, determine a target resource profile from a resource profile set, the target resource profile includes multiple candidate evaluation nodes; and determine the multiple target evaluation nodes from the target resource profile.

[0168] In another possible implementation, the second determining module 1102 is configured to determine, from the resource profile set, a target resource profile of the same input source based on the input source of the recommended requirement information, when the selection strategy information is used to indicate selection of the resource profile based on an input source, wherein the input source is used to indicate a source of the recommended requirement information;

[0169] The second determining module 1102 is configured to determine, when the selection strategy information is used to indicate selecting a resource configuration file based on a recommended requirement, a target resource configuration file corresponding to the recommended requirement from the resource configuration file set based on the recommended requirement corresponding to the recommended requirement information, wherein different recommended requirements correspond to different resource configuration files;

[0170] The second determining module 1102 is configured to, when the selection strategy information is used to indicate selecting a resource configuration file based on a recommended requirement set to which the recommended requirement belongs, determine the recommended requirement set to which the recommended requirement corresponding to the recommended requirement information belongs, and determine, based on the recommended requirement set, a target resource configuration file corresponding to the recommended requirement set from the resource configuration file set, wherein different recommended requirement sets correspond to different resource configuration files, and the resource configuration file corresponding to the recommended requirement set is adapted to at least one recommended requirement included in the recommended requirement set;

[0171] The second determination module 1102 is configured to determine a default resource configuration file from the resource configuration file set and determine the default resource configuration file as the target resource configuration file when the selection policy information is used to indicate selection of a default resource configuration file.

[0172] In another possible implementation, the file name of the resource configuration file in the resource configuration file set includes an input source;

[0173] The second determination module 1102 is used to determine, based on the input source of the recommendation requirement information, a resource configuration file whose file name includes the input source from the resource configuration file set, and determine the determined resource configuration file as the target resource configuration file.

[0174] In another possible implementation, the file name of the resource configuration file in the resource configuration file set includes the first identification information of the recommended requirement;

[0175] The second determination module 1102 is used to determine a resource configuration file including the first identification information of the recommended requirement corresponding to the recommended requirement information from the resource configuration file set based on the first identification information of the recommended requirement, and determine the determined resource configuration file as the target resource configuration file.

[0176] In another possible implementation, the file name of the resource configuration file in the resource configuration file set includes the second identification information of the recommended requirement set;

[0177] The second determination module 1102 is used to determine a resource configuration file including the second identification information from the resource configuration file set based on the second identification information of the recommended requirement set, and determine the determined resource configuration file as the target resource configuration file.

[0178] In another possible implementation, the device further includes:

[0179] The configuration module is used to obtain the configured candidate resource configuration files and add the candidate resource configuration files to the resource configuration file set.

[0180] In another possible implementation, the configuration module is configured to display a first configuration interface, where the first configuration interface displays a plurality of resource configuration files obtained from a server; and obtain a candidate resource configuration file selected from the plurality of resource configuration files.

[0181] In another possible implementation, the configuration module is used to display a second configuration interface, which displays multiple evaluation dimensions; obtain a selected candidate evaluation dimension from the multiple evaluation dimensions, configure the evaluation node based on the selected candidate evaluation dimension, and obtain a configured candidate resource configuration file.

[0182] In another possible implementation, the second determination module 1102 is used to determine the available status and priority level of the multiple candidate evaluation nodes, where the available status of the candidate evaluation node is used to indicate whether the candidate evaluation node is available; based on the available status and priority level of the multiple candidate evaluation nodes, a preset number of candidate evaluation nodes with an available status and the highest priority level are determined from the multiple candidate evaluation nodes to obtain the multiple target evaluation nodes.

[0183] In another possible implementation, the third determination module 1103 is used to determine an evaluation dimension set based on multiple evaluation dimensions corresponding to the multiple target evaluation nodes, the target evaluation node corresponds to at least one evaluation dimension, and the evaluation dimension set includes multiple evaluation dimensions corresponding to the multiple target evaluation nodes; for any evaluation dimension in the evaluation dimension set, the target evaluation node is called, and a quality evaluation is performed on the candidate service based on the evaluation dimension to obtain a second quality parameter value; based on the weights of the multiple evaluation dimensions in the evaluation dimension set, a weighted sum is performed on the multiple second quality parameter values ​​of the candidate service to obtain a first quality parameter value of the candidate service, and the multiple second quality parameter values ​​of the candidate service are obtained by performing a quality evaluation on the candidate service based on the multiple evaluation dimensions.

[0184] In another possible implementation, the device further includes:

[0185] a first acquisition module, configured to acquire first adjustment reference information, where the first adjustment reference information is used to adjust the evaluation dimension of the target evaluation node, and the first adjustment reference information includes at least one of first feedback information based on the target service and real-time data of the terminal, where the real-time data is data currently generated by the terminal;

[0186] A first adjustment module, configured to adjust the evaluation dimension of the target evaluation node based on the first adjustment reference information;

[0187] The second adjustment module is used to adjust the second service list based on the adjusted evaluation dimension of the target evaluation node.

[0188] In another possible implementation, the device further includes:

[0189] a second acquisition module, configured to acquire second adjustment reference information, where the second adjustment reference information is used to adjust the multiple target evaluation nodes, and the second adjustment reference information includes at least one of second feedback information based on the target service and service change information of the terminal, where the service change information is used to indicate a service change situation of the terminal;

[0190] a fifth determining module, configured to re-determine a plurality of target evaluation nodes based on the second adjustment reference information;

[0191] The third adjustment module is used to adjust the second service list based on the re-determined multiple target evaluation nodes.

[0192] In another possible implementation, the third determination module 1103 is used to detect the service level agreement SLA of the multiple target evaluation nodes; when it is detected that the SLA of the multiple target evaluation nodes meets the agreement requirements, the first quality parameter values ​​of the multiple candidate services are determined based on the multiple evaluation dimensions corresponding to the multiple target evaluation nodes.

[0193] In another possible implementation, the first determination module 1101 is used to determine the recommendation requirement information in response to a service recommendation request; based on the recommendation requirement information, determine through a service discovery system a plurality of candidate services that meet the recommendation requirements corresponding to the recommendation requirement information; and form the plurality of candidate services into the first service list.

[0194] In the embodiment of the present application, a comprehensive evaluation of multiple candidate services is performed through multiple evaluation dimensions, which can effectively integrate the results of different evaluation dimensions. This comprehensive processing method can ensure the comprehensiveness of the evaluation, so that the target service finally determined is more in line with the actual needs of the user, thereby improving the accuracy of the target service recommendation and further improving the service recommendation effect.

[0195] It should be noted that the service recommendation device provided in the above embodiment only uses the division of the above functional modules as an example when making service recommendations. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the terminal is divided into different functional modules to complete all or part of the functions described above. In addition, the service recommendation device provided in the above embodiment and the service recommendation method embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0196] See also Fig.12 , Fig.12 12 is a schematic diagram of a terminal structure provided by an exemplary embodiment of the present application. The terminal may also include one or more of the following components: a processor 1210 , a memory 1220 , and a display screen 1230 .

[0197] The processor 1210 uses various interfaces and lines to connect the various parts of the entire terminal 1200, and executes various functions and processes data of the terminal 1200 by running or executing instructions, programs, code sets or instruction sets stored in the memory 1220, and calling data stored in the memory 1220. Optionally, the processor 1210 can be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 1210 can integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), a neural network processor (NPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen 1230; the NPU is used to implement artificial intelligence (AI) functions; and the modem is used to process wireless communications. It is understandable that the above-mentioned modem may not be integrated into the processor 1210, but may be implemented separately through a computer program product.

[0198] The memory 1220 may include a random access memory (RAM) or a read-only memory (ROM). Optionally, the memory 1220 includes a non-transitory computer-readable storage medium. The memory 1220 may be used to store instructions, programs, codes, code sets, or instruction sets. The memory 1220 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the following various method embodiments, etc.; the data storage area may store data (such as audio data, a phone book), etc., created according to the use of the terminal 1200.

[0199] The display screen 1230 is a display component for displaying a user interface. Optionally, the display screen 1230 is a display screen with a touch function, through which the user can use any suitable object such as a finger, a touch pen, etc. to perform touch operations on the display screen 1230.

[0200] The display screen 1230 is usually arranged on the front panel of the terminal 1200. The display screen 1230 can be designed as a full screen, a curved screen, a special-shaped screen, a double-sided screen or a folding screen. The display screen 1230 can also be designed as a combination of a full screen and a curved screen, a combination of a special-shaped screen and a curved screen, etc., which is not limited in this embodiment.

[0201] In addition, those skilled in the art can understand that the structure of the terminal 1200 shown in the above drawings does not constitute a limitation on the terminal 1200, and the terminal 1200 may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently. For example, the terminal 1200 also includes a Wifi module, an audio acquisition device, a speaker, a radio frequency circuit, an input unit, a sensor, an audio circuit, a Bluetooth module, a power supply and other components, which will not be described in detail here.

[0202] An embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores at least one computer instruction, and the at least one computer instruction is used to be executed by a processor to implement the service recommendation method as described in the above embodiment.

[0203] On the other hand, an embodiment of the present application provides a computer program product, which includes computer instructions, and the computer instructions are stored in a computer-readable storage medium; a processor reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to implement the service recommendation method as described in the above embodiment.

[0204] Those skilled in the art should be aware that in one or more of the above examples, the functions described in the embodiments of the present application can be implemented with hardware, software, firmware, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein the communication media include any media that facilitates the transmission of a computer program from one place to another. The storage medium can be any available medium that a general or special-purpose computer can access.

[0205] The above description is only an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A service recommendation method, characterized in that: The method comprises: Determine a first service list of recommended demand information, the first service list including a plurality of candidate services that meet the recommended demand corresponding to the recommended demand information, the recommended demand information being used to indicate the current recommended service demand of the terminal; Based on the recommendation requirement information, determining a plurality of target evaluation nodes, wherein the plurality of target evaluation nodes are used to perform quality evaluation on the plurality of candidate services from different evaluation dimensions; For any candidate service, based on the multiple evaluation dimensions corresponding to the multiple target evaluation nodes, determine multiple first quality parameter values ​​of the candidate service, and determine different first quality parameter values ​​of the candidate service based on different evaluation dimensions, wherein the first quality parameter value of the candidate service is used to indicate the degree to which the candidate service satisfies the recommendation requirement corresponding to the recommendation requirement information; Based on the first quality parameter values ​​of the multiple candidate services, at least one target service is determined from the first service list to obtain a second service list.

2. The method according to claim 1, characterized in that: The step of determining a plurality of target evaluation nodes based on the recommendation requirement information includes: Determine selection strategy information of the recommended demand information, wherein the selection strategy information is used to represent a strategy for selecting a resource configuration file, and different resource configuration files include different candidate evaluation nodes; Based on the selection strategy information, determining a target resource profile from a set of resource profiles, the target resource profile including a plurality of candidate evaluation nodes; The plurality of target evaluation nodes are determined from the target resource configuration file.

3. The method according to claim 2, characterized in that The step of determining a target resource configuration file from a set of resource configuration files based on the selection policy information includes: In the case where the selection strategy information is used to indicate selection of a resource profile based on an input source, determining a target resource profile of the same input source from the resource profile set based on the input source of the recommended requirement information, the input source being used to indicate a source of the recommended requirement information; In the case where the selection strategy information is used to indicate selection of a resource configuration file based on a recommended requirement, based on the recommended requirement corresponding to the recommended requirement information, determining a target resource configuration file corresponding to the recommended requirement from the resource configuration file set, where different recommended requirements correspond to different resource configuration files; In the case where the selection strategy information is used to indicate selecting a resource configuration file based on a recommended requirement set to which the recommended requirement belongs, determining the recommended requirement set to which the recommended requirement corresponding to the recommended requirement information belongs, and based on the recommended requirement set, determining a target resource configuration file corresponding to the recommended requirement set from the resource configuration file set, different recommended requirement sets correspond to different resource configuration files, and the resource configuration file corresponding to the recommended requirement set is adapted to at least one recommended requirement included in the recommended requirement set; In the case where the selection policy information is used to indicate the selection of a default resource configuration file, a default resource configuration file is determined from the resource configuration file set, and the default resource configuration file is determined as the target resource configuration file.

4. The method according to claim 3, characterized in that The file names of the resource configuration files in the resource configuration file set include input sources; The step of determining a target resource configuration file of the same input source from the resource configuration file set based on the input source of the recommended demand information comprises: Based on the input source of the recommendation requirement information, a resource configuration file whose file name includes the input source is determined from the resource configuration file set, and the determined resource configuration file is determined as the target resource configuration file.

5. The method according to claim 3, characterized in that: The file name of the resource configuration file in the resource configuration file set includes the first identification information of the recommended requirement; The determining, based on the recommended requirement corresponding to the recommended requirement information, a target resource configuration file corresponding to the recommended requirement from the resource configuration file set includes: Based on the first identification information of the recommended requirement corresponding to the recommended requirement information, a resource configuration file including the first identification information is determined from the resource configuration file set, and the determined resource configuration file is determined as the target resource configuration file.

6. The method according to claim 3, characterized in that The file name of the resource configuration file in the resource configuration file set includes the second identification information of the recommended requirement set; The determining, based on the recommended requirement set, from the resource configuration file set, a target resource configuration file corresponding to the recommended requirement set, comprises: Based on the second identification information of the recommended requirement set, a resource configuration file including the second identification information is determined from the resource configuration file set, and the determined resource configuration file is determined as the target resource configuration file.

7. The method according to claim 2, characterized in that The method further comprises: The configured candidate resource configuration file is obtained, and the candidate resource configuration file is added to the resource configuration file set.

8. The method according to claim 7, characterized in that The step of obtaining the configured candidate resource configuration file includes: Displaying a first configuration interface, wherein the first configuration interface displays a plurality of resource configuration files obtained from a server; A candidate resource configuration file selected from the multiple resource configuration files is obtained.

9. The method according to claim 7, characterized in that: The step of obtaining the configured candidate resource configuration file includes: Displaying a second configuration interface, wherein the second configuration interface displays multiple evaluation dimensions; A candidate evaluation dimension selected from the multiple evaluation dimensions is obtained, and an evaluation node is configured based on the selected candidate evaluation dimension to obtain a configured candidate resource configuration file.

10. The method according to claim 2, characterized in that The step of determining the plurality of target evaluation nodes from the target resource configuration file comprises: Determining the availability status and priority levels of the multiple candidate evaluation nodes, wherein the availability status of the candidate evaluation node is used to indicate whether the candidate evaluation node is available; Based on the availability status and priority levels of the multiple candidate evaluation nodes, a preset number of candidate evaluation nodes with an availability status of available and the highest priority levels are determined from the multiple candidate evaluation nodes to obtain the multiple target evaluation nodes.

11. The method according to claim 1, characterized in that The determining, based on the multiple evaluation dimensions corresponding to the multiple target evaluation nodes, the first quality parameter values ​​of the multiple candidate services includes: Determine an evaluation dimension set based on the multiple evaluation dimensions corresponding to the multiple target evaluation nodes, the target evaluation node corresponds to at least one evaluation dimension, and the evaluation dimension set includes the multiple evaluation dimensions corresponding to the multiple target evaluation nodes; For any evaluation dimension in the evaluation dimension set, calling the target evaluation node, and performing quality evaluation on the candidate service based on the evaluation dimension to obtain a second quality parameter value; Based on the weights of multiple evaluation dimensions in the evaluation dimension set, a weighted sum is performed on multiple second quality parameter values ​​of the candidate service to obtain the first quality parameter value of the candidate service, and the multiple second quality parameter values ​​of the candidate service are obtained by performing quality evaluation on the candidate service based on the multiple evaluation dimensions.

12. The method according to claim 1, characterized in that The method further comprises: Acquire first adjustment reference information, where the first adjustment reference information is used to adjust the evaluation dimension of the target evaluation node, and the first adjustment reference information includes at least one of first feedback information based on the target service and real-time data of the terminal, where the real-time data is data currently generated by the terminal; Adjusting the evaluation dimension of the target evaluation node based on the first adjustment reference information; The second service list is adjusted based on the adjusted evaluation dimension of the target evaluation node.

13. The method according to claim 1, characterized in that The method further comprises: Acquire second adjustment reference information, where the second adjustment reference information is used to adjust the multiple target evaluation nodes, and the second adjustment reference information includes at least one of second feedback information based on the target service and service change information of the terminal, where the service change information is used to indicate a service change situation of the terminal; re-determining a plurality of target evaluation nodes based on the second adjustment reference information; The second service list is adjusted based on the re-determined plurality of target evaluation nodes.

14. The method according to claim 1, characterized in that The determining, based on the multiple evaluation dimensions corresponding to the multiple target evaluation nodes, the first quality parameter values ​​of the multiple candidate services includes: Testing the service level agreements (SLAs) of the multiple target evaluation nodes; In a case where it is detected that the SLAs of the multiple target evaluation nodes meet protocol requirements, first quality parameter values ​​of the multiple candidate services are determined based on multiple evaluation dimensions corresponding to the multiple target evaluation nodes.

15. The method according to claim 1, characterized in that The first service list of determining the recommended demand information includes: In response to the service recommendation request, determining the recommendation requirement information; Based on the recommended requirement information, determining, through a service discovery system, a plurality of candidate services that meet the recommended requirement corresponding to the recommended requirement information; The plurality of candidate services are combined into the first service list.

16. A service recommendation device, characterized in that: The device comprises: A first determining module, configured to determine a first service list of recommended demand information, wherein the first service list includes a plurality of candidate services that meet the recommended demand corresponding to the recommended demand information, wherein the recommended demand information is used to indicate the current recommended service demand of the terminal; A second determination module is used to determine a plurality of target evaluation nodes based on the recommendation requirement information, wherein the plurality of target evaluation nodes are used to perform quality evaluation on the plurality of candidate services from different evaluation dimensions; A third determination module is used to determine, for any candidate service, multiple first quality parameter values ​​of the candidate service based on multiple evaluation dimensions corresponding to the multiple target evaluation nodes, and determine different first quality parameter values ​​of the candidate service based on different evaluation dimensions, wherein the first quality parameter value of the candidate service is used to indicate the degree to which the candidate service satisfies the recommendation requirement corresponding to the recommendation requirement information; The fourth determining module is configured to determine at least one target service from the first service list based on the first quality parameter values ​​of the plurality of candidate services to obtain a second service list.

17. A terminal, characterized in that: The terminal includes a processor and a memory, wherein the memory stores at least one computer instruction, and the at least one computer instruction is loaded and executed by the processor to implement the service recommendation method according to any one of claims 1 to 15.

18. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one computer instruction, and the at least one computer instruction is used to be executed by a processor to implement the service recommendation method according to any one of claims 1 to 15.

19. A computer program product, characterized in that The computer program product includes computer instructions, which are stored in a computer-readable storage medium; a processor reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to implement the service recommendation method as described in any one of claims 1 to 15.

Images