WO2022141279A1 - 业务体验模型的确定方法及通信装置 - Google Patents

业务体验模型的确定方法及通信装置 Download PDF

Info

Publication number
WO2022141279A1
WO2022141279A1 PCT/CN2020/141751 CN2020141751W WO2022141279A1 WO 2022141279 A1 WO2022141279 A1 WO 2022141279A1 CN 2020141751 W CN2020141751 W CN 2020141751W WO 2022141279 A1 WO2022141279 A1 WO 2022141279A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
data set
network element
association information
information
Prior art date
Application number
PCT/CN2020/141751
Other languages
English (en)
French (fr)
Inventor
辛阳
崇卫微
吴晓波
阎亚丽
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to CN202080103972.7A priority Critical patent/CN116097734A/zh
Priority to EP20967619.6A priority patent/EP4262268A4/en
Priority to PCT/CN2020/141751 priority patent/WO2022141279A1/zh
Publication of WO2022141279A1 publication Critical patent/WO2022141279A1/zh
Priority to US18/344,628 priority patent/US20230353465A1/en

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
    • H04L41/5061—Network service management, e.g. ensuring proper service fulfilment according to agreements characterised by the interaction between service providers and their network customers, e.g. customer relationship management
    • H04L41/5067—Customer-centric QoS measurements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L65/00—Network arrangements, protocols or services for supporting real-time applications in data packet communication
    • H04L65/10—Architectures or entities
    • H04L65/1016—IP multimedia subsystem [IMS]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L65/00—Network arrangements, protocols or services for supporting real-time applications in data packet communication
    • H04L65/80—Responding to QoS
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H04W84/02—Hierarchically pre-organised networks, e.g. paging networks, cellular networks, WLAN [Wireless Local Area Network] or WLL [Wireless Local Loop]
    • H04W84/04—Large scale networks; Deep hierarchical networks
    • H04W84/042—Public Land Mobile systems, e.g. cellular systems

Definitions

  • the embodiments of the present application relate to the field of communication technologies, and in particular, to a method and a communication device for determining a service experience model.
  • service data of terminal equipment is distributed in core network network elements, and also distributed in access network equipment and/or third-party equipment.
  • An implementation method is that a data analysis network element in the core network performs model training based on service data of terminal devices distributed on different devices or network elements.
  • the data analysis network element cannot obtain the complete data on the core network network element, access network equipment and third-party equipment, which makes it impossible to accurately train the model.
  • the embodiments of the present application provide a method and a communication device for determining a service experience model, so as to accurately determine the service experience model of a service.
  • an embodiment of the present application provides a method for determining a service experience model, including: a data analysis network element obtains first associated information corresponding to a first data set, where the first data set includes service data on a core network network element. data; the data analysis network element obtains a second data set according to the first association information, the second data set includes data corresponding to the second association information in the first data set, and the second association information is the first association information and the intersection of the third association information, the third association information corresponds to the third data set, and the third data set includes the data of the service on the first device; the data analysis network element according to the first information and the second data set, determine a fourth data set, the fourth data set is a subset or all of the second data set, the first information includes the capability information of the first device and/or the capability information of the data analysis network element; the The data analysis network element determines the service experience model of the service according to the fourth data set.
  • the data analysis network element obtains the data of the service of the terminal device on the core network element, aligns the data with the first device, and then performs model training based on the aligned data to obtain an accurate service experience model, and analyzes the data.
  • the network element does not need to acquire the data of the service of the terminal device on the first device.
  • the fourth data set includes a training set, the training set corresponds to the fourth associated information, and the data analysis network element determines the service experience model of the service according to the fourth data set, including: the data The analysis network element obtains the training set according to the fourth association information and the fourth data set; the data analysis network element determines at least one candidate service experience model according to the training set; the data analysis network element obtains the at least one candidate service experience model from the at least one candidate service
  • the business experience model of the business is determined in the experience model.
  • the fourth data set further includes a verification set, the verification set corresponds to the fifth association information, and the data analysis network element determines the service experience model of the service from the at least one candidate service experience model, It includes: the data analysis network element obtains the verification set according to the fifth association information and the fourth data set; the data analysis network element determines the verification results corresponding to the at least one candidate service experience model according to the verification set; the The data analysis network element determines the service experience model of the service according to the verification results corresponding to the at least one candidate service experience model respectively.
  • the fourth data set further includes a test set, the test set corresponds to the sixth association information, and the method further includes: the data analysis network element according to the sixth association information and the fourth data set , to obtain the test set; the data analysis network element determines the test result of the service experience model of the service according to the test set.
  • the data analysis network element sends the fourth association information corresponding to the training set to the first device.
  • the data analysis network element sends a first request to the second device, where the first request carries the identification information of the first device, and the first request is used to request the first data set; the A data analysis network element receives the first data set from the second device.
  • the data analysis network element acquiring the second data set according to the first association information includes: the data analysis network element sending the first association information to the first device; the data analysis network element Receive the second association information from the first device; the data analysis network element acquires the second data set according to the second association information and the first data set.
  • the data analysis network element acquiring the second data set according to the first association information includes: the data analysis network element receiving the third association information from the first device; the data analysis network element The second association information is determined according to the third association information and the first association information; the data analysis network element obtains the second data set according to the second association information and the first data set.
  • obtaining the first associated information corresponding to the first data set by the data analysis network element includes: the data analysis network element sends a second request to the second device, where the second request carries the first data set.
  • the identification information of the device, the second request is used to request the first association information corresponding to the first data set; the data analysis network element receives the first association information from the second device.
  • the data analysis network element acquiring the second data set according to the first association information includes: the data analysis network element determines the second association information according to the first association information; the data analysis The network element sends a third request to the second device, the third request carries the second association information, and the third request is used to request the second data set; the data analysis network element receives the second data set from the second device data set.
  • the method can reduce the data transmission volume, thereby reducing the data transmission pressure.
  • the data analysis network element sends a fourth request to the network element storage function network element, where the fourth request is used to request the address information of the second device; the data analysis network element obtains information from the network element The storage function network element receives the address information of the second device.
  • the second device is a data analysis network element supporting a data lake function, or a data analysis network element supporting a data collection coordination function, or a data analysis network element supporting a data collection function.
  • the first device is an access network device or a service device.
  • the first association information includes the following information: identification information of the first device, identification information allocated by the first device to the terminal device, and time stamp.
  • the identification information allocated by the first device to the terminal device is the identification information allocated by the first device to the terminal device on a first interface, where the first interface is a connection between the first device and the terminal device.
  • the data analysis network element determines the service experience model of the service by adopting a vertical federated learning method according to the fourth data set.
  • an embodiment of the present application provides a communication device, and the device may be a data analysis network element or a chip used for the data analysis network element.
  • the apparatus has the function of implementing the above-mentioned first aspect or each possible implementation method based on the first aspect. This function can be implemented by hardware or by executing corresponding software by hardware.
  • the hardware or software includes one or more modules corresponding to the above functions.
  • an embodiment of the present application provides a communication device, including a processor, where the processor is coupled to a memory, and the memory is used to store a program or an instruction, and when the program or instruction is executed by the processor, the device implements the above-mentioned first aspect Or each possible implementation method based on the first aspect.
  • the memory may be located within the device or external to the device.
  • the processor includes one or more.
  • an embodiment of the present application provides a communication apparatus, including units or means for executing the first aspect or each step of each possible implementation method based on the first aspect.
  • an embodiment of the present application provides a communication device, including a processor and an interface circuit, where the processor is configured to control the interface circuit to communicate with other devices, and execute the first aspect or possible implementations based on the first aspect method.
  • the processor includes one or more.
  • embodiments of the present application further provide a computer-readable storage medium, including instructions, which, when executed on a computer, cause the computer to execute the first aspect or each possible implementation method based on the first aspect.
  • the embodiments of the present application further provide a computer program product, which, when running on a computer, enables the computer to execute the above-mentioned first aspect or each possible implementation method based on the first aspect.
  • an embodiment of the present application further provides a chip system, including a processor, the processor is coupled to a memory, and the memory is used to store programs or instructions, and when the programs or instructions are executed by the processor, the chip system enables the above-mentioned first step.
  • a chip system including a processor, the processor is coupled to a memory, and the memory is used to store programs or instructions, and when the programs or instructions are executed by the processor, the chip system enables the above-mentioned first step.
  • the memory may be located within the system-on-chip, or may be located outside the system-on-chip.
  • the processor includes one or more.
  • FIG. 1 is a schematic diagram of a 5G network architecture to which the embodiments of the present application are applied;
  • FIG. 2 is a schematic diagram of the business experience evaluation process based on NWDAF
  • Figure 3 is a schematic diagram of the process of vertical federated learning training
  • Figure 4 is a schematic diagram of data set division
  • FIG. 5 is a schematic diagram of the NWDAF functional decomposition architecture
  • Fig. 6 is the data schematic diagram that NWDAF obtains UE through pairwise association
  • FIG. 7 provides a method for determining a service experience model according to an embodiment of the present application.
  • FIG. 8 provides a method for determining a service experience model according to an embodiment of the present application
  • FIG. 9 provides a method for determining a service experience model according to an embodiment of the present application.
  • FIG. 10 provides a method for determining a service experience model according to an embodiment of the present application
  • FIG. 11 provides a communication device according to an embodiment of the present application.
  • FIG. 12 provides another communication apparatus according to an embodiment of the present application.
  • FIG. 1 is a schematic diagram of a 5G network architecture to which the embodiments of the present application are applied
  • the 5G network architecture shown in FIG. 1 includes three parts, namely, a terminal device part, a data network (DN) part, and an operator network part. The following briefly describes the functions of some of the network elements.
  • the operator network may include one or more of the following network elements: Authentication Server Function (AUSF) network element, Network Exposure Function (NEF) network element, Policy Control Function (Policy Control Function) Function, PCF) network element, unified data management (unified data management, UDM), unified database (Unified Data Repository, UDR), network storage function (Network Repository Function, NRF) network element, access and mobility management function (Access and Mobility Management Function, AMF) network elements, session management function (session management function, SMF) network elements, radio access network (Radio Access Network, RAN) equipment user plane function (user plane function, UPF) network elements and network data Analysis function (Network Data Analytics Function, NWDAF) network element, etc.
  • the part other than the radio access network part may be referred to as the core network part.
  • the operator network further includes an application function (Application Function, AF) network element.
  • the terminal device in this embodiment of the present application may be a device for implementing a wireless communication function.
  • the terminal equipment may be a user equipment (UE), an access terminal, a terminal unit, a terminal station, a mobile station, a mobile station in a 5G network or a public land mobile network (PLMN) evolved in the future.
  • UE user equipment
  • PLMN public land mobile network
  • remote station remote terminal
  • mobile device wireless communication device
  • terminal agent or terminal device etc.
  • the access terminal may be a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA), a wireless communication Functional handheld devices, computing devices or other processing devices connected to wireless modems, in-vehicle devices or wearable devices, virtual reality (VR) end devices, augmented reality (AR) end devices, industrial control (industrial) wireless terminal in control), wireless terminal in self-driving, wireless terminal in remote medical, wireless terminal in smart grid, wireless terminal in transportation safety Terminals, wireless terminals in smart cities, wireless terminals in smart homes, etc. Terminals can be mobile or stationary.
  • SIP session initiation protocol
  • WLL wireless local loop
  • PDA personal digital assistant
  • a wireless communication Functional handheld devices computing devices or other processing devices connected to wireless modems, in-vehicle devices or wearable devices, virtual reality (VR) end devices, augmented reality (AR) end devices, industrial control (industrial) wireless terminal in control), wireless terminal in self-driving,
  • the above-mentioned terminal device can establish a connection with the operator network through an interface (eg, N1, etc.) provided by the operator network, and use the data and/or voice services provided by the operator network.
  • the terminal device can also access the DN through the operator's network, and use the operator's service deployed on the DN and/or the service provided by a third party.
  • the above-mentioned third party may be a service party other than the operator's network and the terminal device, and may provide other data and/or voice services for the terminal device.
  • the specific expression form of the above third party can be specifically determined according to the actual application scenario, and is not limited here.
  • RAN is a sub-network of an operator's network, and is an implementation system between service nodes and terminal equipment in the operator's network.
  • the terminal device To access the operator's network, the terminal device first passes through the RAN, and then can be connected to the service node of the operator's network through the RAN.
  • the RAN device in this application is a device that provides a wireless communication function for a terminal device, and the RAN device is also called an access network device.
  • the RAN equipment in this application includes but is not limited to: next-generation base station (g nodeB, gNB), evolved node B (evolved node B, eNB), radio network controller (radio network controller, RNC), node B in 5G (node B, NB), base station controller (BSC), base transceiver station (base transceiver station, BTS), home base station (for example, home evolved nodeB, or home node B, HNB), baseband unit (baseBand unit, BBU), transmission point (transmitting and receiving point, TRP), transmitting point (transmitting point, TP), mobile switching center, etc.
  • next-generation base station g nodeB, gNB
  • evolved node B evolved node B
  • eNB evolved node B
  • RNC radio network controller
  • node B in 5G node B, NB
  • base station controller BSC
  • base transceiver station base transceiver station
  • BTS home base station
  • base station for example, home
  • the AMF network element mainly performs functions such as mobility management and access authentication/authorization. In addition, it is also responsible for transferring user policies between UE and PCF.
  • the SMF network element mainly performs functions such as session management, execution of control policies issued by PCF, selection of UPF, and allocation of UE Internet Protocol (IP) addresses.
  • IP Internet Protocol
  • the UPF network element as the interface UPF with the data network, implements functions such as user plane data forwarding, session/flow-level accounting statistics, and bandwidth limitation.
  • the UDM network element is mainly responsible for the management of contract data, user access authorization and other functions.
  • UDR is mainly responsible for the access function of contract data, policy data, application data and other types of data.
  • the NEF network element is mainly used to support the opening of capabilities and events.
  • the AF network element mainly conveys the requirements of the application side to the network side, such as quality of service (Quality of Service, QoS) requirements or user status event subscriptions.
  • the AF may be a third-party functional entity or an application service deployed by an operator, such as an IP Multimedia Subsystem (IP Multimedia Subsystem, IMS) voice call service.
  • IP Multimedia Subsystem IP Multimedia Subsystem
  • the PCF network element is mainly responsible for policy control functions such as charging for sessions and service flow levels, QoS bandwidth guarantee and mobility management, and UE policy decision-making.
  • the NRF network element can be used to provide the network element discovery function, and provide network element information corresponding to the network element type based on the request of other network elements.
  • NRF also provides network element management services, such as network element registration, update, de-registration, and network element status subscription and push.
  • AUSF network element It is mainly responsible for authenticating users to determine whether to allow users or devices to access the network.
  • NWDAF network elements are mainly used to collect network data (including one or more of terminal equipment data, RAN equipment data, core network data and third-party application data), and provide network data analysis services, which can output data analysis results. It is used for network, network management and application execution policy decision-making. NWDAF can utilize machine learning models for data analysis.
  • the functions of NWDAF in 3GPP Release 17 are decomposed, including data collection function (or data collection function), model training function (or machine learning model training logical function), and model inference function (or analytics logical function). In the scenario where the data collection function, training function and inference function are separated, the data collection function, training function and inference function of the same model can be deployed separately in different NWDAF instances.
  • the NWDAF deploying the data collection function (may be referred to as a data collection NWDAF or data lake or data repository function (DRF)) can be used from terminal equipment, RAN equipment, core network equipment and third parties Data is collected in the application, and the NWDAF that deploys the training function (can be called training NWDAF) can perform model training according to the collected data to obtain the trained model, and the NWDAF that deploys the inference function (referred to as inference NWDAF) obtains the model provided by training NWDAF. Model inference, providing data analysis services.
  • the NWDAF can be an independent network element, or can be set up together with other network elements, for example, set the NWDAF to the PCF network element.
  • a DN is a network outside the operator's network.
  • the operator's network can access multiple DNs, and multiple services can be deployed on the DNs, which can provide data and/or voice services for terminal devices.
  • DN is the private network of a smart factory.
  • the sensors installed in the workshop of the smart factory can be terminal devices, and the control server of the sensor is deployed in the DN, and the control server can provide services for the sensor.
  • the sensor can communicate with the control server, obtain the instruction of the control server, and transmit the collected sensor data to the control server according to the instruction.
  • the DN is an internal office network of a company.
  • the mobile phones or computers of employees of the company can be terminal devices, and the mobile phones or computers of employees can access information and data resources on the internal office network of the company.
  • Nnwdaf, Nausf, Nnef, Npcf, Nudm, Naf, Namf, Nsmf, N1, N2, N3, N4, and N6 are interface serial numbers.
  • interface serial numbers refer to the meanings defined in the 3GPP standard protocol, which is not limited here.
  • the data analysis network element may be the NWDAF network element shown in FIG. 1 , or may be other network elements in the future communication system that have the functions of the NWDAF network element in the present application.
  • Mobility management The network element may be the AMF network element shown in FIG. 1, or may be other network elements in the future communication system having the functions of the AMF network element in this application, and the policy control network element may be the PCF network element shown in FIG. It can be other network elements in the future communication system that have the functions of the PCF network element in this application, and the user plane network element can be the UPF network element shown in FIG.
  • the application function network element may be the AF network element shown in Figure 1, or may be other network elements in the future communication system with the function of the AF network element in this application, and the access network device may be Figure 1
  • the RAN equipment shown may also be other network elements in the future communication system having the functions of the RAN equipment in this application.
  • the data analysis network element is the NWDAF network element
  • the mobility management network element is the AMF network element
  • the policy control network element is the PCF network element
  • the user plane network element is the PCF network element
  • the application function is an AF network element
  • NWDAF network elements are further divided into data lake (also called data collection NWDAF), training NWDAF network elements and inference NWDAF network elements.
  • the terminal device is a UE as an example for description.
  • Service providers (such as service equipment or AF) are most concerned about the service experience of their services in the 5G network, and can accurately understand their own service characteristics. Therefore, service providers can accurately measure their service experience and effectively monitor service quality.
  • the current 5G network lacks a service experience evaluation mechanism, trying to ensure rich and changeable 5G services through fixed QoS parameters, resulting in the inability to accurately match service experience requirements and network resources.
  • FIG. 2 it is a schematic diagram of the service experience evaluation process based on NWDAF.
  • the NWDAF determines the service experience data analysis result of the service flow, and sends the service experience data analysis result to the PCF.
  • the PCF judges whether the service experience of the service flow can be satisfied according to the relationship between the service experience data analysis result and the service experience requirement. If it cannot be satisfied, the PCF can re-determine the QoS parameters of the service flow.
  • the network can improve the service experience of the service flow, so that the NWDAF can obtain the service experience data analysis result based on the new service experience of the service flow. Able to meet business experience requirements.
  • NWDAF can train to obtain the relationship model between service experience and RAN data, core network (CN) data, and AF data, that is, the service experience model. Based on the service experience model, NWDAF can determine the service experience data analysis results corresponding to the new data of the RAN, core network and AF. Taking linear regression as an example, the business experience model is as follows:
  • h(x) represents the service experience
  • x 0 1
  • x i 1
  • D the dimension of the data
  • the private data of RAN, CN, and AF cannot be centralized in the same NWDAF for centralized training. This is because the private data of RAN, CN, and AF come from different domains, and the private data of RAN comes from the access network. domain, CN's private data comes from the core network domain, and AF's private data comes from the third-party device domain.
  • VFL Vertical Federated Learning
  • VFL solves the model training and reasoning when each participant is unwilling to share the original data. with less overlap.
  • VFL combines different data features of common samples of multiple participants to perform federated learning, that is, the training data of each participant is divided vertically, so it is called vertical federated learning.
  • the business experience model based on vertical federated learning is as follows:
  • x i represents the ith sample data, where is the private data distributed on the RAN in the i-th sample data, is the private data distributed on CN in the ith sample data, is the private data distributed on AF in the ith sample data, is the public data actively reported by RAN, CN and AF in the i-th sample data (such as RSRP/RSRQ/SINR, QoS flow Bit Rate/QoS flow Packet Delay/QoS flow Packet Error Rate, Buffer Size in Table 1), ⁇ A , ⁇ B , ⁇ C , ⁇ D are respectively the corresponding model parameters.
  • ⁇ A , ⁇ B , ⁇ C , ⁇ D are parameter vectors composed of one or more model parameters.
  • Client A has a dataset
  • Client B has a dataset
  • y i is the label data
  • L is the loss function
  • model parameters are updated as follows:
  • Step 1 Client A and Client B initialize model parameters ⁇ A and ⁇ B respectively;
  • Step 2 Client A calculates based on ⁇ A and L A , which is then sent to Client B;
  • Step 3 Client B calculates based on ⁇ B further based on And yi calculates di , L AB , and L B , and finally calculates L based on L A , L AB , and L B .
  • Client B sends d i to Client A;
  • Step 4 Client A and Client B respectively calculate based on d i as well as Then based on as well as Update model parameters ⁇ A and ⁇ B .
  • steps 2 to 4 are executed cyclically until the end condition of model training is reached, such as the number of iterations reaches a set threshold (such as 10,000 times) or the value of the loss function L is less than the set threshold (such as 0.001).
  • a set threshold such as 10,000 times
  • the value of the loss function L is less than the set threshold (such as 0.001).
  • Client A and Client B calculate the local inference results based on the trained model parameters ⁇ A and ⁇ B respectively as well as Then Client A will infer the result locally Send it to Client B, and Client B will finally determine the inference result
  • FIG. 4 it is a schematic diagram of data set division.
  • the dataset needs to be divided into: training set, validation set and test set.
  • the training set is used to train the models corresponding to each algorithm according to different algorithms
  • the validation set is used to verify the results of each model
  • the algorithm can be continuously updated to adjust the model during the training process, and the best among them can be selected according to the verification results.
  • the test results of the model are determined.
  • NWDAF can be decomposed into training NWDAF (responsible for model training), inference NWDAF (responsible for inference of data analysis results), and data lake (collection and control of training data or inference data).
  • NWDAF needs to do end-to-end UE granularity (per UE) data analysis between RAN, CN and AF, and then needs to consider the association of UE data distributed on RAN, CN and AF.
  • the NWDAF can determine the data of the UE through the pairwise association of the association information.
  • the embodiment of the present application adopts the following idea: RAN-AMF-SMF-UPF-AF.
  • FIG. 6 it is a schematic diagram showing that the NWDAF obtains UE data through pairwise association.
  • the RAN and AMF report data they both carry the timestamp, the RAN UE NGAP ID assigned to the UE on the N2 interface, and the RAN Global RAN Node ID, that is, each piece of data is carried out with associated information.
  • the associated information includes Timestamp, RAN UE NGAP ID and Global RAN Node ID.
  • NGAP is the abbreviation of Next Generation Application Protocol (Next Generation Application Protocol). Therefore, NWDAF correlates UE data on RAN and AMF by Timestamp, RAN UE NGAP ID and Global RAN Node ID.
  • NWDAF associates UE data on RAN and UPF through Timestamp and AN Tunnel Info, associates UE data on SMF and PCF through Timestamp and Subscription Permanent Identifier (SUPI), and associates UE data on SMF and PCF through Timestamp
  • SUPI Timestamp and Subscription Permanent Identifier
  • SUPI Subscription Permanent Identifier
  • associate UE's data on AMF and SMF with SUPI associate UE's data on SMF and UPF through Timestamp and UE IP
  • Internet Protocol 5-tuple Internet Protocol 5-tuple, IP 5-tuple
  • association information between RAN and AMF is denoted by a
  • association information between AMF and SMF is denoted by b
  • association information between SMF and UPF is denoted by c
  • association information between UPF and AF is denoted by d means, then:
  • the format of the sample data of the UE collected by the RAN is (a, the data of the UE);
  • the format of the sample data of the UE collected by the AMF is (a, b, the data of the UE), and then the AMF reports to the NWDAF (a, b, the data of the UE);
  • the format of the sample data of the UE collected by the SMF is (b, c, the data of the UE), and then the SMF reports to the NWDAF (b, c, the data of the UE);
  • the format of the UE's sample data collected by the UPF is (c, d, the UE's data), and then the UPF reports to the NWDAF (c, d, the UE's data);
  • the format of the sample data of the UE collected by the AF is (d, the data of the UE), and then the AF reports (d, the data of the UE) to the NWDAF.
  • the sample data reported by AMF, SMF, UPF, and AF received by NWDAF are: (a, UE data), (a, b, UE data), (b, c, UE data), (c, d, UE data) data), (d, UE data), then NWDAF can convert the format of the sample data according to the correspondence between a and b, the correspondence between b and c, and the correspondence between c and d into (a, UE's data).
  • the embodiment of the present application provides a method for determining a service experience model, and the method can be executed by an NWDAF in a core network or a chip used for NWDAF.
  • the first data set includes the data of the service of the terminal device on the core network element, that is, the first data set is used to represent the set of data of the service of the terminal device on the core network element.
  • the core network element here may be one or more of UPF, SMF, AMF, PCF, and so on.
  • the acquired data of the network element of the core network may include the data of the UPF, the data of the AMF, and the like shown in Table 1.
  • the first association information is used to associate the first data set with the third data set on the first device (which may be an access network device or a service device (also referred to as AF)).
  • the first association information may also be used to identify data of the first data set.
  • the first association information corresponds to the first data set.
  • the first association information may include the following information: identification information of the first device, identification information allocated by the first device to the terminal device, and timestamp (Timestamp).
  • the identification information allocated by the first device to the terminal device may be identification information allocated by the first device to the terminal device on a first interface, and the first interface may be an interface between the first device and a core network element.
  • the first device here may be an access network device or an AF.
  • the first association information may include Timestamp, RAN UE NGAP ID and Global RAN Node ID. Among them, Timestamp is the timestamp, RAN UE NGAP ID is the identification information allocated by the access network device to the terminal device, and the Global RAN Node ID is the identification information of the access network device.
  • the third data set includes the data of the service of the terminal device on the first device (which may be an access network device or AF), that is, the third data set is used to indicate that the service of the terminal device is on the first device collection of data.
  • the acquired data on the first device may include RAN data or AF data shown in Table 1.
  • the third data set corresponds to the same service as the first data set.
  • the third association information is used to associate the third dataset with the first dataset on NWDAF.
  • the third association information may also be used to identify the data of the third data set.
  • the third association information corresponds to the third data set.
  • the third association information may include the following information: identification information of the first device, identification information allocated by the first device to the terminal device, and timestamp (Timestamp).
  • the identification information allocated by the first device to the terminal device may be identification information allocated by the first device to the terminal device on a first interface, and the first interface may be an interface between the first device and a core network element.
  • the third association information may include: Timestamp, RAN UE NGAP ID and Global RAN Node ID. Among them, Timestamp is the timestamp, RAN UE NGAP ID is the identification information allocated by the access network device to the terminal device, and the Global RAN Node ID is the identification information of the access network device.
  • the second data set includes data corresponding to the second association information in the first data set, and the second association information is an intersection of the first association information and the third association information. Therefore, the second dataset is a proper subset of the first dataset or the second dataset is the same as the first dataset.
  • the first data set includes 10,000 pieces of data, and the corresponding first associated information is represented as ID-1 to ID-10000, respectively.
  • the third data set includes 9000 pieces of data, and the corresponding third associated information is represented as ID-5001 to ID-14000, respectively. Therefore, the second data set includes 5000 pieces of data, and the corresponding second association information is represented as ID-5001 to ID-10000, respectively.
  • the fourth data set is a subset of the second data set or the fourth data set is the same as the second data set.
  • the fourth data set includes a training set, and the training set corresponds to the fourth associated information.
  • the fourth data set further includes a validation set and a test set. The verification set corresponds to the fifth associated information, and the test set corresponds to the sixth associated information.
  • an embodiment of the present application provides a method for determining a service experience model, and the method includes the following steps:
  • Step 701 NWDAF acquires first association information corresponding to the first data set.
  • Step 702 NWDAF acquires a second data set according to the first association information.
  • Step 703 the NWDAF determines a fourth data set according to the first information and the second data set, where the first information includes capability information of the first device and/or capability information of the NWDAF.
  • the capability information here includes the computing power of the central processing unit (CPU), the computing power resources of the graphics processing unit (Graphics Processing Unit, GPU), the memory resources, the hard disk resources, the NWDAF and the first device (RAN device or AF). ), one or more of bandwidth resources or time delays are transmitted.
  • Step 704 NWDAF determines the service experience model of the service according to the fourth data set.
  • NWDAF can perform vertical federated learning based on the fourth dataset to obtain a business experience model of the business.
  • the NWDAF may be client A or client B in FIG. 3 , and the specific model training method may refer to the foregoing description.
  • NWDAF obtains the data of the service of the terminal device on the core network element, aligns the data with the first device, and then performs model training based on the aligned data to obtain an accurate service experience model. NWDAF does not need to obtain the data of the terminal device. Business data on the first device.
  • the NWDAF may also acquire the first data set from the second device. For example, the NWDAF sends a first request to the second device, where the first request carries the identification information of the first device, and the first request is used to request the first data set, or it can be understood that the first request is used to request the acquisition of core network elements
  • the second device then sends the first data set to the NWDAF.
  • the method for the NWDAF to acquire the second data set may be, for example: the NWDAF sends the first association information to the first device, and the first device sends the first association information to the first device according to the third association information and the third association information.
  • the first association information determines the second association information, that is, the intersection of the third association information and the first association information is determined as the second association information, and then the first device sends the second association information to the NWDAF, and then the NWDAF A data set acquires the second data set, that is, the NWDAF acquires the data corresponding to the second associated information in the first data set to form the second data set.
  • the method for the NWDAF to acquire the second data set may, for example, be: the first device sends the third association information to the NWDAF, and the NWDAF sends the third association information to the NWDAF according to the third association information and the first data set.
  • the association information determines the second association information, that is, NWDAF determines the intersection of the third association information and the first association information as the second association information, and then NWDAF obtains the second data set according to the second association information and the first data set, that is The NWDAF acquires data corresponding to the second associated information in the first data set to form a second data set.
  • the NWDAF may not need to acquire the first data set from the second device, but directly acquire the second data set from the second device.
  • the method can reduce the data transmission volume, thereby reducing the data transmission pressure.
  • the above step 701 may be: NWDAF sends a second request to the second device, the second request carries the identification information of the first device, and the second request is used to request the first association information corresponding to the first data set, or It is understood that the second request is used to request to obtain the association information corresponding to the service data of the terminal device related to the first device on the core network element, and then the second device sends the first association information to the NWDAF.
  • the NWDAF After the NWDAF acquires the first association information from the second device, the NWDAF aligns the association information with the first device, that is, the NWDAF determines the second association information according to the first association information, and the specific method may refer to the foregoing description. Then, the NWDAF sends a third request to the second device, where the third request carries the second association information, the third request is used to request the second data set corresponding to the second association information, and the second device obtains the first data set according to the second association information Second dataset and send the second dataset to NWDAF. That is, in the above step 702, the NWDAF determines the second association information according to the first association information, and then acquires the second data set from the second device according to the second association information.
  • the NWDAF before the NWDAF acquires the first association information, the first data set or the second data set from the second device, the NWDAF also sends a fourth request, the fourth request is used to request the address information of the second device, and then the network element storage function network element sends the address information of the second device to the NWDAF.
  • the second device may be an NWDAF supporting a data lake function (also referred to as a data lake), or an NWDAF supporting a data collection coordination function, or an NWDAF supporting a data collection function.
  • the above-mentioned fourth data set includes a training set, and the training set corresponds to the fourth associated information, that is, the data of the training set consists of the fourth associated information.
  • the training set is a proper subset of the fourth data set or the training set is the same as the fourth data set.
  • the training set is used to train models corresponding to each algorithm according to different algorithms. Therefore, the above step 704 may be: NWDAF obtains a training set according to the fourth association information and the fourth data set, NWDAF determines at least one candidate service experience model according to the training set, and NWDAF determines a service experience model of the service from the at least one candidate service experience model.
  • NWDAF performs model training based on Algorithm 1 and the training set to obtain candidate service experience model 1, performs model training based on Algorithm 2 and the training set to obtain candidate service experience model 2, and performs model training based on Algorithm 3 and the training set to obtain candidate service experience.
  • Model 3 The NWDAF then determines the service experience model of the service from the candidate service experience model 1 , the candidate service experience model 2 and the candidate service experience model 3 .
  • the above-mentioned fourth data set further includes a verification set, and the verification set corresponds to the fifth association information, that is, the data of the verification set is determined by the fifth association information.
  • the validation set is a proper subset of the fourth dataset or the validation set is the same as the fourth dataset.
  • the verification set is used to verify the results of each model, and the algorithm can be continuously updated to adjust the model during the training process, and the best model can be selected according to the verification results.
  • NWDAF determines the service experience model of the service from at least one candidate service experience model. For example, it may be: NWDAF obtains a verification set according to the fifth association information and the fourth data set, and NWDAF determines the corresponding verification of at least one candidate service experience model according to the verification set.
  • the NWDAF determines the service experience model of the service according to the verification results corresponding to the at least one candidate service experience model respectively.
  • the candidate service experience model corresponding to the best verification result is determined as the service experience model.
  • sort the verification results take the top N verification results (N is an integer greater than 1), randomly select a verification result from the N verification results, and assign the candidate service experience model corresponding to the selected verification result. Determined as a business experience model.
  • the above-mentioned fourth data set further includes a test set, and the test set corresponds to the sixth association information, that is, the data of the test set is determined by the sixth association information.
  • the test set is a proper subset of the fourth data set or the test set is the same as the fourth data set.
  • the test set is used to determine the test results of the business experience model.
  • the NWDAF may acquire the test set according to the sixth association information and the fourth data set, and then the NWDAF determines the test result of the service experience model of the service according to the test set.
  • NWDAF can divide the fourth data set into training set, validation set and test set, and the training set, validation set and test set have no intersection with each other.
  • the union of the training set, the validation set and the test set is equal to the fourth data set.
  • NWDAF may further divide the fourth data set into a training set, a validation set and a test set, and the training set, the validation set and the test set may have intersections with each other.
  • the union of the training set, the validation set and the test set is equal to the fourth data set.
  • the NWDAF may also send the fourth association information corresponding to the training set to the first device, and the first device may determine the information on the first device according to the fourth association information and the third data set.
  • a training set also referred to as a fifth data set
  • model training is performed based on the fifth data set to obtain at least one candidate service experience model corresponding to the above service on the first device.
  • the NWDAF may also send the fifth association information corresponding to the verification set to the first device, and the first device may determine the information on the first device according to the fifth association information and the third data set.
  • a verification set also referred to as a sixth data set
  • the at least one candidate business experience model is verified, and based on the verification result, a business experience model is selected from the at least one candidate business experience model, as A service experience model corresponding to the above service on the first device.
  • the NWDAF may also send the sixth association information corresponding to the test set to the first device, and the first device may determine the information on the first device according to the sixth association information and the third data set.
  • a test set also referred to as a seventh data set
  • the above-mentioned service experience model is tested based on the seventh data set, and a test result of the service experience model corresponding to the above-mentioned service on the first device is obtained.
  • the service experience model determined by the first device may be referred to as a service experience sub-model of the first device, and the service experience model determined by the NWDAF may be referred to as a service experience sub-model of the core network.
  • the first device acquires the service data of the terminal device on the first device, aligns the data with the NWDAF, and then performs model training based on the aligned data to obtain an accurate service experience model, and the first device does not need to acquire the terminal device The data of the service on the core network element.
  • FIG. 7 The method shown in FIG. 7 will be described below with reference to the specific embodiments corresponding to FIG. 8 to FIG. 10 .
  • FIG. 8 it is a schematic diagram of a method for determining a service experience model according to an embodiment of the present application. Based on this method, there is a data lake on the core network side, and NWDAF obtains the data of the UE in the core network corresponding to the RAN device from the data lake according to the address information of the RAN device. data to implement longitudinal federated learning.
  • the method includes the following steps:
  • Step 801a the RAN device collects and stores the data of the UE in the access network.
  • the RAN device collection data includes a third data set.
  • the third data set may include data of multiple UEs, and the data of each UE corresponds to one piece of association information (also referred to as third association information).
  • the third association information is association information between RAN and AMF, and the third association information includes Global RAN Node ID, RAN UE NGAP ID and Timestamp.
  • Global RAN Node ID is the identity of the RAN
  • RAN UE NGAP ID is the identity allocated by the RAN device on the N2 interface for the UE
  • Timestamp is the timestamp corresponding to the data collected by the RAN device from the UE.
  • Step 801b the data lake collects and stores UE data in the core network.
  • the data lake collects UE data on core network elements such as UPF, AMF, SMF, and PCF.
  • core network elements such as UPF, AMF, SMF, and PCF.
  • the UE data collected by the data lake may refer to Table 1.
  • the core network element is an AMF for illustration.
  • the data lake subscribes the data of the UE on the AMF to the AMF through the Namf_EventExposure_Subscribe service operation, it can carry the indication information in the service operation.
  • the information is used to indicate the type of association information included in the UE data reported by the AMF.
  • the association information includes Global RAN Node ID, RAN UE NGAP ID and Timestamp.
  • the association information includes Global RAN Node ID, AMF UE NGAP ID, Timestamp.
  • the data lake can obtain UE data on the core network according to the pairwise association method described above.
  • the association information corresponding to the data of the UE in the core network includes the Global RAN Node ID, RAN UE NGAP ID and Timestamp.
  • Step 802 NWDAF sends a request message to the data lake. Accordingly, the data lake receives the request message.
  • the request message carries the Global RAN node ID and a specific time period, and the request message is used to request the data lake to feed back the data of the UE on the core network related to the RAN device in a specific time period.
  • the Global RAN node ID is the identification information of the RAN device.
  • the request message further includes second indication information, where the second indication information is used to instruct the data lake to feed back data of the associated UE on the core network related to the RAN device in a specific time period, wherein the data lake is associated
  • the second indication information is used to instruct the data lake to feed back data of the associated UE on the core network related to the RAN device in a specific time period, wherein the data lake is associated
  • the NWDAF determines that the service experience model needs to be trained for the RAN device according to the local operator policy, and then the NWDAF sends the above request message to the data lake.
  • the service provider that is, AF
  • the NWDAF decides based on the vertical Federated learning method, joint RAN equipment and core network, or joint core network and AF, or joint core network, RAN equipment and AF, training to obtain a service experience model.
  • the method for the NWDAF to obtain the identifier of the RAN device may be, for example: the NWDAF queries the operation, maintenance, operation and maintenance (Operation, Administration and Management, OAM) device for the identifier of the RAN device deployed in the specific area. Since the NWDAF obtains the poor service experience information from the service provider, the NWDAF can learn the address information of the AF.
  • OAM Operaation, Administration and Management
  • Step 803 the data lake sends a response message to the NWDAF, and accordingly, the NWDAF receives the response message.
  • the response message carries data of the UE in the core network related to the RAN device and association information corresponding to the data of the UE, and the association information includes Global RAN Node ID, RAN UE NGAP ID and Timestamp.
  • the UE data acquired by the NWDAF from the data lake is called the first data set, and the association information corresponding to the UE data acquired by the NWDAF from the data lake is called the first association information.
  • the first data set may include data of multiple UEs, and the data of each UE corresponds to one piece of association information (also referred to as first association information).
  • Step 804 the NWDAF performs data alignment with the RAN device to obtain a second data set.
  • the NWDAF may send the acquired first association information to the RAN device, and the RAN device determines the second association information according to the first association information and the third association information, and the second association information is the first association information and the third association information. The intersection of three related information.
  • the RAN device then sends the second association information to the NWDAF. Therefore, the RAN device can determine the candidate data set in the access network according to the second association information and the third data set, and the NWDAF can determine the candidate data set (also referred to as the first data set in the core network) according to the second association information and the first data set. two data sets), the second data set corresponds to the second associated information.
  • the RAN device may send the acquired third association information to the NWDAF, and the NWDAF determines the second association information according to the first association information and the third association information.
  • the second association information is the first association information and the third association information.
  • the intersection of the three association information, and then the second association information is sent to the RAN device. Therefore, the RAN device can determine the candidate data set in the access network according to the second association information, and the NWDAF can determine the candidate data set (also referred to as the second data set) in the core network according to the second association information and the first data set.
  • the candidate data set on the RAN side includes the data of 10 million UEs as follows:
  • the candidate data set (ie the second data set) on the NWDAF side includes the data of 10 million UEs as follows:
  • Data b2 of UE2 corresponding to association information 2;
  • Data b4 of UE4 corresponding to association information 4;
  • Step 805a the NWDAF sends a request message to the RAN device. Accordingly, the RAN device receives the request message.
  • the request message is used to request to obtain the capability information of the RAN device that can participate in the vertical federation training within a specific time period.
  • the request message carries specific time period information.
  • the request message carries indication information, where the indication information is used to instruct to acquire capability information of the RAN device that can participate in the vertical federation training within a specific time period.
  • Step 805b the RAN device sends to the NWDAF the capability information of the RAN device that can participate in the vertical federation training within a specific time period.
  • the NWDAF receives the capability information of the RAN equipment that can participate in the vertical federation training within a specific time period.
  • Step 806 the NWDAF determines the fourth data set according to the capability information of the RAN, the capability information of the NWDAF, and the size of the candidate data set (ie, the second data set) on the NWDAF side.
  • NWDAF determines whether the second data set can all participate in training, if not, selects some UE data from the second data set to form the fourth data set .
  • the second data set that can participate in vertical federated learning after step 804 contains data of 10 million UEs, and every 10% of the RAN equipment can accommodate data of 1 million UEs to participate in training, assuming that NWDAF determines that the RAN The remaining 80% of the capacity on the side can be used for training, then NWDAF can randomly select 8 million UE data from the 10 million UE data to participate in federated training, that is, the selected fourth data set contains the 8 million UE data. data.
  • Step 807 NWDAF divides the fourth data set to obtain a training data set, a verification data set and a test data set.
  • NWDAF can randomly divide it into 7 million - 500,000 - 500,000, as training data set, validation data set and test data set respectively.
  • the training data set includes one or more pieces of data, and the training data set corresponds to the fourth associated information.
  • the verification data set contains one or more pieces of data, and the verification data set corresponds to the fifth associated information.
  • the test data set contains one or more pieces of data, and the test data set corresponds to the sixth associated information.
  • Step 808 the NWDAF sends configuration information to the RAN. Accordingly, the RAN receives the configuration information.
  • the configuration information is used to indicate the association information respectively corresponding to the training data set, the verification data set and the test data set used for the vertical federated learning on the RAN side, that is, the above-mentioned fourth association information, fifth association information and sixth association information. Therefore, according to the configuration information, the RAN can select some UE data from the RAN side candidate data set to form the RAN side training data set, select some UE data to form the RAN side verification data set, and select some UE data Constitute the test dataset on the RAN side.
  • the configuration information also includes the type of algorithm used in the vertical federated learning, and parameter information set inside each algorithm (such as the number of layers of the neural network algorithm, the activation function used by each layer, the type of loss function, etc.).
  • both NWDFA and RAN have determined the training data set, verification data set and test data set for vertical federated learning, the two can subsequently train models based on the vertical federated learning method to obtain business experience models respectively.
  • the two perform model training based on their respective training datasets, validation datasets, and test datasets, respectively, to obtain service experience models.
  • NWDAF obtains the data of the terminal equipment's services on the core network elements, aligns the data with the RAN equipment, and then performs model training based on the aligned data to obtain an accurate service experience model. NWDAF does not need to obtain the terminal equipment's services. data on RAN equipment.
  • NWDAF can also send the associated information corresponding to the training data set, the verification data set and the test data set to the RAN device in different stages (such as the training stage, the verification stage and the test stage), to instruct the RAN device to perform the model. Training or model validation or model testing.
  • the above-mentioned NWDAF may be a training NWDAF.
  • the training NWDAF obtains a service experience model after performing model training according to the vertical federation method, and then transmits the service experience model to the inference NWDAF for generating data analysis results on the core network side.
  • the reasoning NWDAF is mainly based on the trained business experience model and new data to determine the data analysis results.
  • FIG. 9 it is a schematic diagram of a method for determining a service experience model provided by an embodiment of the present application. Based on this method, there is a data lake on the core network side, and NWDAF obtains the data of the UE in the core network corresponding to the RAN device from the data lake according to the address information of the RAN device. data to implement longitudinal federated learning.
  • the method includes the following steps:
  • Steps 901a to 901b are the same as the above-mentioned steps 801a to 801b, and reference may be made to the foregoing description.
  • Step 902 NWDAF sends a request message to the data lake. Accordingly, the data lake receives the request message.
  • the request message carries the Global RAN node ID and a specific time period, and the request message is used to request the data lake to feed back the association information corresponding to the data of the UE in the core network related to the RAN device in the specific time period.
  • the Global RAN node ID is the identification information of the RAN device.
  • the request message further includes second indication information, where the second indication information is used to instruct the data lake to feed back the association information corresponding to the data of the associated UE in the core network related to the RAN device in a specific time period,
  • second indication information is used to instruct the data lake to feed back the association information corresponding to the data of the associated UE in the core network related to the RAN device in a specific time period
  • the NWDAF determines that the service experience model needs to be trained for the RAN device according to the local operator policy, and then the NWDAF sends the above request message to the data lake.
  • the service provider that is, AF
  • the NWDAF decides based on the vertical Federated learning method, joint RAN equipment and core network, or joint core network and AF, or joint core network, RAN equipment and AF, training to obtain a service experience model.
  • the method for the NWDAF to acquire the identifier of the RAN device may be, for example, the NWDAF may query the OAM device for the identifier of the RAN device deployed in the specific area. Since the NWDAF obtains the poor service experience information from the service provider, the NWDAF can learn the address information of the AF.
  • Step 903 the data lake sends a response message to the NWDAF, and accordingly, the NWDAF receives the response message.
  • the response message carries association information corresponding to the data of the UE in the core network related to the RAN device, and the association information includes Global RAN Node ID, RAN UE NGAP ID and Timestamp.
  • the response message also carries the total size of data of UEs in the core network related to the RAN device, or carries the size of data corresponding to each UE in the data in the core network related to the RAN device.
  • the association information obtained by NWDAF from the data lake is called the first association information.
  • Step 904 the NWDAF performs data alignment with the RAN device to obtain second association information.
  • the NWDAF may send the acquired first association information to the RAN device, and the RAN device determines the second association information according to the first association information and the third association information, and the second association information is the first association information and the third association information. The intersection of three related information.
  • the RAN then sends the second association information to the NWDAF.
  • the RAN device can determine the candidate data set in the access network according to the second association information and the third data set.
  • the RAN may send the acquired third association information to the NWDAF, and the NWDAF determines the second association information according to the first association information and the third association information, and the second association information is the first association information and the third association information.
  • the intersection of the association information, and then the second association information is sent to the RAN device.
  • the RAN device can determine the candidate data set in the access network according to the second association information and the third data set.
  • Step 905 NWDAF sends a request message to the data lake. Accordingly, the data lake receives the request message.
  • the request message carries the second association information, and the request message is used to request to obtain data of the UE corresponding to the second association information.
  • Step 906 the data lake sends a response message to the NWDAF, and accordingly, the NWDAF receives the response message.
  • the response message carries the data of the UE in the core network related to the RAN device corresponding to the second association information, where the second association information includes Global RAN Node ID, RAN UE NGAP ID and Timestamp.
  • the set formed by the data of the UE in the core network related to the RAN device corresponding to the second association information is referred to as the second data set.
  • the NWDAF can obtain the association information (ie, the first association information) corresponding to the data of the UE in the core network from the data lake.
  • the NWDAF can obtain the alignment information.
  • the UE data ie, the second data set
  • the UE data ie, the second data set
  • Steps 907a to 910 are the same as the above-mentioned steps 805a to 808, and reference may be made to the foregoing description.
  • both NWDFA and RAN have determined the training data set, verification data set and test data set for vertical federated learning, the two can subsequently train models based on the vertical federated learning method, respectively, to obtain business experience models.
  • you can refer to the method described above use RAN as Client A, and use NWDAF as Client B.
  • the two perform model training based on their respective training datasets, validation datasets, and test datasets, and then Client A and Client B pass through the middle of the interaction. result d i and other information, the two Clients jointly obtain the service experience model.
  • the parameter ⁇ A corresponding to the service experience model is reserved on Client A
  • the parameter ⁇ B is reserved on Client B, which are respectively used for local reasoning.
  • NWDAF obtains the data of the terminal equipment's services on the core network elements, aligns the data with the RAN equipment, and then performs model training based on the aligned data to obtain an accurate service experience model. NWDAF does not need to obtain the terminal equipment's services. data on RAN equipment. Moreover, in this solution, the NWDAF does not obtain the data of all the UEs in the core network from the data lake, but obtains the data of the UEs corresponding to the aligned second association information, thus reducing the transmission pressure between the NWDAF and the data lake.
  • the above-mentioned NWDAF may be a training NWDAF.
  • the training NWDAF obtains a service experience model after performing model training according to the vertical federation method, and then transmits the service experience model to the inference NWDAF for generating data analysis results on the core network side.
  • the reasoning NWDAF is mainly based on the trained business experience model and new data to determine the data analysis results.
  • FIG. 10 it is a schematic diagram of a method for determining a service experience model provided by an embodiment of the present application.
  • NWDAFs with data collection function.
  • the information of the RAN device corresponding to the data is registered on the NRF, and the NWDAF that assists the federated learning and training addresses these NWDAFs through the NRF, and then requests the NWDAF for the data of the UE of the core network related to a certain RAN device.
  • the method includes the following steps:
  • Step 1001 is the same as the above-mentioned step 801a, and reference may be made to the foregoing description.
  • step 1002 different NWDAFs on the core network side collect and store data of UEs related to the same RAN device in the core network.
  • NWDAFs can collect data of UEs related to the same RAN device on the same or different core network elements.
  • NWDAF1 to NWDAF3 collect data of UEs related to RAN devices on the AMF
  • NWDAF4 to NWDAF7 collect data on the SMF.
  • RAN equipment-related UE data, NWDAF8 to NWDAF10 collect RAN equipment-related UE data on the UPF, and so on.
  • Step 1003 different NWDAFs on the core network side send registration requests to the NRF, and accordingly, the NRF receives the registration requests.
  • the registration request is used to request the identification information of the NWDAF to be registered to the NRF, and the NWDAF is the NWDAF that stores the data of the UE corresponding to the association information.
  • the registration request carries the Global RAN Node ID, Timestamp and NWDAF ID.
  • the registration request carries Global RAN Node ID, RAN UE NGAP ID, Timestamp and NWDAF ID.
  • Step 1004 train the NWDAF to send a request message to the NRF. Accordingly, the NRF receives the request message.
  • the request message carries the Global RAN node ID and a specific time period, and the request message is used to request the NWDAF ID corresponding to the Global RAN node ID and the specific time period, that is, the NWDAF indicated by the NWDAF ID is stored in the specific time period.
  • Step 1005 the NRF sends a response message to the training NWDAF. Accordingly, the training NWDAF receives the response message.
  • the response message carries one or more NWDAF IDs.
  • Step 1006 Train the NWDAF to send request messages to different NWDAFs. Correspondingly, different NWDAFs receive request messages.
  • the request message carries the Global RAN node ID and a specific time period, and the request message is used to request to feed back the data of the UE related to the RAN device in the specific time period.
  • the request message carries indication information, where the indication information is used to indicate to feed back data of the UE related to the RAN device in a specific time period.
  • step 1007 different NWDAFs send response messages to the training NWDAF, and accordingly, the training NWDAF receives the response message.
  • the response message carries UE data related to the RAN device and association information corresponding to the data of each UE.
  • the associated information includes, for example, Global RAN Node ID, RAN UE NGAP ID, and Timestamp, or includes Global RAN Node ID, AN Tunnel Info, and Timestamp, or includes Global RAN Node ID, SUPI, and Timestamp, etc.
  • Step 1008 train the NWDAF to process data from UEs of different NWDAFs.
  • the set of UE data obtained after the training NWDAF process is called the first data set, and the association information corresponding to the first data set is called the first association information.
  • Steps 1009 to 1013 are the same as the above-mentioned steps 804 to 808, and reference may be made to the foregoing description.
  • NWDAF obtains the data of the terminal equipment's services on the core network elements, aligns the data with the RAN equipment, and then performs model training based on the aligned data to obtain an accurate service experience model. NWDAF does not need to obtain the terminal equipment's services. data on RAN equipment.
  • the above solution is for the scenario where there is no data lake, and the data of the UE associated with a certain RAN device may exist in different NWDAFs, so that the training NWDAF supporting federated learning can separately request these NWDAFs according to the identity of the RAN device. UE data.
  • the communication device 1100 includes an association information acquisition unit 1110 , a data set acquisition unit 1120 , a data set determination unit 1130 and a service The experience model determination unit 1140 .
  • a transceiver unit 1150 is also included.
  • the associated information obtaining unit 1110 is configured to obtain the first associated information corresponding to the first data set, the first data set includes the data of the service on the core network element; the data set obtaining unit 1120 is configured to obtain the first associated information according to the first data set.
  • the association information obtains a second data set, the second data set includes data corresponding to the second association information in the first data set, and the second association information is the intersection of the first association information and the third association information , the third association information corresponds to a third data set, and the third data set includes the data of the service on the first device;
  • the data set determining unit 1130 is configured to determine according to the first information and the second data set data set, determine a fourth data set, the fourth data set is a subset or all of the second data set, and the first information includes the capability information of the first device and/or the data analysis network meta-capability information; the service experience model determining unit 1140 is configured to determine the service experience model of the service according to the fourth data set.
  • the fourth data set includes a training set, the training set corresponds to the fourth association information, and the service experience model determining unit 1140 is specifically configured to: according to the fourth association information and For the fourth data set, the training set is obtained; according to the training set, at least one candidate service experience model is determined; and a service experience model of the service is determined from the at least one candidate service experience model.
  • the fourth data set further includes a verification set, where the verification set corresponds to fifth association information
  • the service experience model determining unit 1140 is specifically configured to: according to the fifth association information and the fourth data set, obtaining the verification set; according to the verification set, determining the verification results corresponding to the at least one candidate service experience model respectively; according to the verification results corresponding to the at least one candidate service experience model, A business experience model for the business is determined.
  • the fourth data set further includes a test set, where the test set corresponds to sixth association information
  • the service experience model determining unit 1140 is further configured to: according to the sixth association information and the fourth data set, obtaining the test set; and determining the test result of the service experience model of the service according to the test set.
  • the transceiver unit 1150 is configured to send the fourth association information corresponding to the training set to the first device.
  • the transceiver unit 1150 is configured to: send a first request to a second device, where the first request carries the identification information of the first device, and the first request is used to request the a first data set; receiving the first data set from the second device.
  • the data set obtaining unit 1120 is specifically configured to: send the first associated information to the first device through the transceiver unit 1150; and obtaining the second data set according to the second association information and the first data set.
  • the data set obtaining unit 1120 is specifically configured to: receive the third association information from the first device through the transceiver unit 1150; The first association information is determined, and the second association information is determined; and the second data set is acquired according to the second association information and the first data set.
  • the associated information obtaining unit 1110 is specifically configured to: send a second request to the second device through the transceiver unit 1150, where the second request carries the identification information of the first device , the second request is used to request the first association information corresponding to the first data set; the first association information is received from the second device through the transceiver unit 1150 .
  • the data set obtaining unit 1120 is specifically configured to: determine the second association information according to the first association information; send the data set to the second device through the transceiver unit 1150 a third request, where the third request carries the second association information, and the third request is used to request the second data set; the second data is received from the second device through the transceiver unit 1150 set.
  • the transceiver unit 1150 is further configured to: send a fourth request to the network element storage function network element, where the fourth request is used to request the address information of the second device;
  • the network element storage function network element receives the address information of the second device.
  • the second device is a data analysis network element supporting a data lake function, or a data analysis network element supporting a data collection coordination function, or a data analysis network element supporting a data collection function.
  • the first device is an access network device or a service device.
  • the first association information includes the following information: identification information of the first device, identification information allocated by the first device to the terminal device, and a timestamp.
  • the identification information allocated by the first device to the terminal device is the identification information allocated by the first device to the terminal device on a first interface
  • the first interface is An interface between the first device and the core network element.
  • the above-mentioned communication device may further include a storage unit, which is used to store data or instructions (also referred to as codes or programs), and each of the above-mentioned units may interact or be coupled with the storage unit to implement corresponding methods or functions.
  • the processing unit 1120 may read data or instructions in the storage unit, so that the communication apparatus implements the methods in the above embodiments.
  • each unit in the above communication apparatus can all be implemented in the form of software calling through the processing element; also all can be implemented in the form of hardware; some units can also be implemented in the form of software calling through the processing element, and some units can be implemented in the form of hardware.
  • each unit may be a separately established processing element, or may be integrated in a certain chip of the communication device to realize, in addition, it may also be stored in the memory in the form of a program, which can be called and executed by a certain processing element of the communication device. function of the unit.
  • each step of the above method or each of the above units may be implemented by an integrated logic circuit of hardware in the processor element or implemented in the form of software being invoked by the processing element.
  • a unit in any of the above communication devices may be one or more integrated circuits configured to implement the above method, eg, one or more application specific integrated circuits (ASICs), or, an or multiple microprocessors (digital singnal processors, DSP), or, one or more field programmable gate arrays (FPGA), or a combination of at least two of these integrated circuit forms.
  • ASICs application specific integrated circuits
  • DSP digital singnal processors
  • FPGA field programmable gate arrays
  • a unit in the communication device can be implemented in the form of a processing element scheduler
  • the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processors that can invoke programs.
  • CPU central processing unit
  • these units can be integrated together and implemented in the form of a system-on-a-chip (SOC).
  • the communication apparatus includes: a processor 1210 and an interface 1230 , and optionally, the communication apparatus further includes a memory 1220 .
  • the interface 1230 is used to enable communication with other devices.
  • the method performed by the data analysis network element in the above embodiments may be implemented by the processor 1210 calling a program stored in a memory (which may be the memory 1220 in the data analysis network element, or an external memory). That is, the data analysis network element may include a processor 1210, and the processor 1210 executes the method performed by the data analysis network element in the above method embodiments by invoking the program in the memory.
  • the processor here may be an integrated circuit with signal processing capability, such as a CPU.
  • the data analysis network element may be implemented by one or more integrated circuits configured to implement the above methods. For example: one or more ASICs, or, one or more microprocessor DSPs, or, one or more FPGAs, etc., or a combination of at least two of these integrated circuit forms. Alternatively, the above implementations may be combined.
  • the functions/implementation process of the associated information obtaining unit 1110, the data set obtaining unit 1120, the data set determining unit 1130, the service experience model determining unit 1140 and the transceiver unit 1150 in FIG. 11 can be implemented through the communication device 1200 shown in FIG.
  • the processor 1210 in the memory 1220 invokes computer-executable instructions stored in the memory 1220 for implementation.
  • the functions/implementation process of the association information acquiring unit 1110, the data set acquiring unit 1120, the data set determining unit 1130 and the service experience model determining unit 1140 in FIG. 11 can be implemented by the processor 1210 in the communication device 1200 shown in FIG. 12 It is implemented by calling the computer-executed instructions stored in the memory 1220.
  • the function/implementation process of the transceiver unit 1150 in FIG. 11 can be implemented through the interface 1230 in the communication device 1200 shown in FIG.
  • the function/implementation process may be implemented by the processor calling program instructions in the memory to drive the interface 1230 .
  • At least one item (single, species) of a, b, or c can represent: a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, and c can be single or multiple.
  • “Plurality" means two or more, and other quantifiers are similar.
  • the above-mentioned embodiments it may be implemented in whole or in part by software, hardware, firmware or any combination thereof.
  • software it can be implemented in whole or in part in the form of a computer program product.
  • the computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated.
  • the computer may be a general purpose computer, special purpose computer, computer network, or other programmable device.
  • the computer instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be downloaded from a website site, computer, server, or data center Transmission to another website site, computer, server, or data center is by wire (eg, coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (eg, infrared, wireless, microwave, etc.).
  • the computer-readable storage medium may be any available medium that a computer can access, or a data storage device such as a server, a data center, or the like that includes an integration of one or more available media.
  • the usable media may be magnetic media (eg, floppy disks, hard disks, magnetic tapes), optical media (eg, DVDs), or semiconductor media (eg, solid state disks (SSDs)), and the like.
  • a general-purpose processor may be a microprocessor, or alternatively, the general-purpose processor may be any conventional processor, controller, microcontroller, or state machine.
  • a processor may also be implemented by a combination of computing devices, such as a digital signal processor and a microprocessor, multiple microprocessors, one or more microprocessors in combination with a digital signal processor core, or any other similar configuration. accomplish.
  • a software unit executed by a processor, or a combination of the two.
  • Software units can be stored in random access memory (Random Access Memory, RAM), flash memory, read-only memory (Read-Only Memory, ROM), EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM or this In any other form of storage media in the field.
  • a storage medium may be coupled to the processor such that the processor may read information from, and store information in, the storage medium.
  • the storage medium can also be integrated into the processor.
  • the processor and storage medium may be provided in the ASIC.
  • the above-described functions described herein may be implemented in hardware, software, firmware, or any combination of the three. If implemented in software, the functions may be stored on, or transmitted over, a computer-readable medium in the form of one or more instructions or code.
  • Computer-readable media includes computer storage media and communication media that facilitate the transfer of a computer program from one place to another. Storage media can be any available media that a general-purpose or special-purpose computer can access.
  • Such computer-readable media may include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other device that can be used to carry or store instructions or data structures and Other media in the form of program code that can be read by a general-purpose or special-purpose computer, or a general-purpose or special-purpose processor.
  • any connection is properly defined as a computer-readable medium, for example, if software is transmitted from a web site, server or other remote source over a coaxial cable, fiber optic computer, twisted pair, digital subscriber line (DSL) Or transmitted by wireless means such as infrared, wireless, and microwave are also included in the definition of computer-readable media.
  • DSL digital subscriber line
  • the discs and magnetic discs include compact discs, laser discs, optical discs, digital versatile discs (English: Digital Versatile Disc, DVD for short), floppy discs and Blu-ray discs. Disks usually reproduce data magnetically, while Discs usually use lasers to optically reproduce data. Combinations of the above can also be included in computer readable media.
  • the functions described in this application may be implemented in hardware, software, firmware, or any combination thereof.
  • the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium.
  • Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another.
  • a storage medium can be any available medium that can be accessed by a general purpose or special purpose computer.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Multimedia (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • General Business, Economics & Management (AREA)
  • Artificial Intelligence (AREA)
  • Medical Informatics (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Telephonic Communication Services (AREA)

Abstract

本申请实施例提供业务体验模型的确定方法及通信装置。该方法包括：数据分析网元获取第一数据集对应的第一关联信息，第一数据集包括业务在核心网网元上的数据；数据分析网元根据第一关联信息获取第二数据集；数据分析网元根据第一信息以及第二数据集确定第四数据集，第四数据集为第二数据集的子集或者全部，第一信息包括第一设备的能力信息和/或数据分析网元的能力信息；数据分析网元根据第四数据集确定业务的业务体验模型。该方案，数据分析网元获取终端设备的业务在核心网网元上的数据，并与第一设备进行数据对齐，然后基于对齐后的数据进行模型训练得到准确的业务体验模型，数据分析网元无需获取终端设备的业务在第一设备上的数据。

Description

业务体验模型的确定方法及通信装置 技术领域
本申请实施例涉及通信技术领域，尤其涉及业务体验模型的确定方法及通信装置。
背景技术
在第五代(5th generation，5G)网络通信中，终端设备的业务数据分布在核心网网元，以及还分布在接入网设备和/或第三方设备。
为了对业务体验进行评估，需要基于分布在不同设备或网元上的终端设备的业务数据进行模型训练，得到业务体验模型，并根据该业务体验模型对业务体验进行评估。
一种实现方法是，由核心网内的数据分析网元基于分布在不同设备或网元上的终端设备的业务数据进行模型训练。然而，基于数据保密的需求，数据分析网元无法获得核心网网元、接入网设备和第三方设备上的完整数据，进而导致无法准确进行模型训练。
发明内容
本申请实施例提供业务体验模型的确定方法及通信装置，用以准确确定业务的业务体验模型。
第一方面，本申请实施例提供一种业务体验模型的确定方法，包括：数据分析网元获取第一数据集对应的第一关联信息，该第一数据集包括业务在核心网网元上的数据；该数据分析网元根据该第一关联信息获取第二数据集，该第二数据集包括第二关联信息在该第一数据集中对应的数据，该第二关联信息是该第一关联信息和第三关联信息的交集，该第三关联信息对应第三数据集，该第三数据集包括该业务在该第一设备上的数据；该数据分析网元根据第一信息以及该第二数据集，确定第四数据集，该第四数据集为该第二数据集的子集或者全部，该第一信息包括该第一设备的能力信息和/或该数据分析网元的能力信息；该数据分析网元根据该第四数据集，确定该业务的业务体验模型。
基于上述方案，数据分析网元获取终端设备的业务在核心网网元上的数据，并与第一设备进行数据对齐，然后基于对齐后的数据进行模型训练得到准确的业务体验模型，并且数据分析网元无需获取终端设备的业务在第一设备上的数据。
在一种可能的实现方法中，该第四数据集包括训练集，该训练集对应第四关联信息，该数据分析网元根据该第四数据集确定该业务的业务体验模型，包括：该数据分析网元根据该第四关联信息以及该第四数据集，获取该训练集；该数据分析网元根据该训练集，确定至少一个候选业务体验模型；该数据分析网元从该至少一个候选业务体验模型中确定该业务的业务体验模型。
在一种可能的实现方法中，该第四数据集还包括验证集，该验证集对应第五关联信息，该数据分析网元从该至少一个候选业务体验模型中确定该业务的业务体验模型，包括：该数据分析网元根据该第五关联信息以及该第四数据集，获取该验证集；该数据分析网元根据该验证集，确定该至少一个候选业务体验模型分别对应的验证结果；该数据分析网元根据该至少一个候选业务体验模型分别对应的验证结果，确定该业务的业务体验模型。
在一种可能的实现方法中，该第四数据集还包括测试集，该测试集对应第六关联信息，该方法还包括：该数据分析网元根据该第六关联信息以及该第四数据集，获取该测试集；该数据分析网元根据该测试集，确定该业务的业务体验模型的测试结果。
在一种可能的实现方法中，该数据分析网元向该第一设备发送该训练集对应的该第四关联信息。
在一种可能的实现方法中，该数据分析网元向第二设备发送第一请求，该第一请求携带该第一设备的标识信息，该第一请求用于请求该第一数据集；该数据分析网元从该第二设备接收该第一数据集。
在一种可能的实现方法中，该数据分析网元根据该第一关联信息获取第二数据集，包括：该数据分析网元向该第一设备发送该第一关联信息；该数据分析网元接收来自该第一设备的该第二关联信息；该数据分析网元根据该第二关联信息以及该第一数据集，获取该第二数据集。
在一种可能的实现方法中，该数据分析网元根据该第一关联信息获取第二数据集，包括：该数据分析网元从该第一设备接收该第三关联信息；该数据分析网元根据该第三关联信息和该第一关联信息，确定该第二关联信息；该数据分析网元根据该第二关联信息以及该第一数据集，获取该第二数据集。
在一种可能的实现方法中，该数据分析网元获取第一数据集对应的第一关联信息，包括：该数据分析网元向第二设备发送第二请求，该第二请求携带该第一设备的标识信息，该第二请求用于请求该第一数据集对应的第一关联信息；该数据分析网元从该第二设备接收该第一关联信息。
在一种可能的实现方法中，该数据分析网元根据该第一关联信息获取第二数据集，包括：该数据分析网元根据该第一关联信息，确定该第二关联信息；该数据分析网元向该第二设备发送第三请求，该第三请求携带该第二关联信息，该第三请求用于请求该第二数据集；该数据分析网元从该第二设备接收该第二数据集。
基于该方案，当第二数据集的数据量少于第一数据集的数据量时，该方法可以减少数据传输量，进而减轻数据传输压力。
在一种可能的实现方法中，该数据分析网元向网元存储功能网元发送第四请求，该第四请求用于请求该第二设备的地址信息；该数据分析网元从该网元存储功能网元接收该第二设备的地址信息。
在一种可能的实现方法中，该第二设备为支持数据湖功能的数据分析网元，或者为支持数据收集协调功能的数据分析网元，或者为支持数据收集功能的数据分析网元。
在一种可能的实现方法中，该第一设备为接入网设备或者业务设备。
在一种可能的实现方法中，该第一关联信息包括以下信息：该第一设备的标识信息、该第一设备为终端设备分配的标识信息、时间戳。
在一种可能的实现方法中，该第一设备为该终端设备分配的标识信息是该第一设备为该终端设备在第一接口上分配的标识信息，该第一接口是该第一设备与该核心网网元之间的接口。
在一种可能的实现方法中，该数据分析网元根据该第四数据集采用纵向联邦学习的方式确定该业务的业务体验模型。
第二方面，本申请实施例提供一种通信装置，该装置可以是数据分析网元，还可以是 用于数据分析网元的芯片。该装置具有实现上述第一方面或基于第一方面的各可能的实现方法的功能。该功能可以通过硬件实现，也可以通过硬件执行相应的软件实现。该硬件或软件包括一个或多个与上述功能相对应的模块。
第三方面，本申请实施例提供一种通信装置，包括处理器，处理器与存储器耦合，存储器用于存储程序或指令，当程序或指令被处理器执行时，使得该装置实现上述第一方面或基于第一方面的各可能的实现方法。该存储器可以位于该装置之内，也可以位于该装置之外。且该处理器包括一个或多个。
第四方面，本申请实施例提供一种通信装置，包括用于执行上述第一方面或基于第一方面的各可能的实现方法的各个步骤的单元或手段(means)。
第五方面，本申请实施例提供一种通信装置，包括处理器和接口电路，所述处理器用于控制接口电路与其它装置通信，并执行上述第一方面或基于第一方面的各可能的实现方法。该处理器包括一个或多个。
第六方面，本申请实施例还提供一种计算机可读存储介质，包括指令，当其在计算机上运行时，使得计算机执行上述第一方面或基于第一方面的各可能的实现方法。
第七方面，本申请实施例还提供一种计算机程序产品，当其在计算机上运行时，使得计算机执行上述第一方面或基于第一方面的各可能的实现方法。
第八方面，本申请实施例还提供一种芯片系统，包括处理器，处理器与存储器耦合，存储器用于存储程序或指令，当程序或指令被处理器执行时，使得该芯片系统实现上述第一方面或基于第一方面的各可能的实现方法。该存储器可以位于该芯片系统之内，也可以位于该芯片系统之外。且该处理器包括一个或多个。
附图说明
图1为本申请实施例所适用的5G网络架构示意图；
图2为基于NWDAF的业务体验评估流程示意图；
图3为纵向联邦学习训练的过程示意图；
图4为数据集划分示意图；
图5为NWDAF功能分解架构示意图；
图6为NWDAF通过两两关联得到UE的数据示意图；
图7为本申请实施例提供一种业务体验模型的确定方法；
图8为本申请实施例提供一种业务体验模型的确定方法；
图9为本申请实施例提供一种业务体验模型的确定方法；
图10为本申请实施例提供一种业务体验模型的确定方法；
图11为本申请实施例提供一种通信装置；
图12为本申请实施例提供另一种通信装置。
具体实施方式
参考图1，为本申请实施例所适用的5G网络架构示意图，图1所示的5G网络架构包括三部分，分别是终端设备部分、数据网络(data network，DN)部分和运营商网络部分。 下面对其中的部分网元的功能进行简单介绍说明。
其中，运营商网络可包括以下网元中的一个或多个：鉴权服务器功能(Authentication Server Function，AUSF)网元、网络开放功能(network exposure function，NEF)网元、策略控制功能(Policy Control Function，PCF)网元、统一数据管理(unified data management，UDM)、统一数据库(Unified Data Repository，UDR)、网络存储功能(Network Repository Function，NRF)网元、接入与移动性管理功能(Access and Mobility Management Function，AMF)网元、会话管理功能(session management function，SMF)网元、无线接入网(Radio Access Network，RAN)设备用户面功能(user plane function，UPF)网元以及网络数据分析功能(Network Data Analytics Function，NWDAF)网元等。上述运营商网络中，除无线接入网部分之外的部分可以称为核心网络部分。在一种可能的实现方法中，运营商网络中还包括应用功能(Application Function，AF)网元。
在具体实现中，本申请实施例中的终端设备，可以是用于实现无线通信功能的设备。其中，终端设备可以是5G网络或者未来演进的公共陆地移动网络(public land mobile network，PLMN)中的用户设备(user equipment，UE)、接入终端、终端单元、终端站、移动站、移动台、远方站、远程终端、移动设备、无线通信设备、终端代理或终端装置等。接入终端可以是蜂窝电话、无绳电话、会话启动协议(session initiation protocol，SIP)电话、无线本地环路(wireless local loop，WLL)站、个人数字助理(personal digital assistant，PDA)、具有无线通信功能的手持设备、计算设备或连接到无线调制解调器的其它处理设备、车载设备或可穿戴设备，虚拟现实(virtual reality，VR)终端设备、增强现实(augmented reality，AR)终端设备、工业控制(industrial control)中的无线终端、无人驾驶(self driving)中的无线终端、远程医疗(remote medical)中的无线终端、智能电网(smart grid)中的无线终端、运输安全(transportation safety)中的无线终端、智慧城市(smart city)中的无线终端、智慧家庭(smart home)中的无线终端等。终端可以是移动的，也可以是固定的。
上述终端设备可通过运营商网络提供的接口(例如N1等)与运营商网络建立连接，使用运营商网络提供的数据和/或语音等服务。终端设备还可通过运营商网络访问DN，使用DN上部署的运营商业务，和/或第三方提供的业务。其中，上述第三方可为运营商网络和终端设备之外的服务方，可为终端设备提供其他数据和/或语音等服务。其中，上述第三方的具体表现形式，具体可根据实际应用场景确定，在此不做限制。
RAN作为接入网网元是运营商网络的子网络，是运营商网络中业务节点与终端设备之间的实施系统。终端设备要接入运营商网络，首先是经过RAN，进而可通过RAN与运营商网络的业务节点连接。本申请中的RAN设备，是一种为终端设备提供无线通信功能的设备，RAN设备也称为接入网设备。本申请中的RAN设备包括但不限于：5G中的下一代基站(g nodeB，gNB)、演进型节点B(evolved node B，eNB)、无线网络控制器(radio network controller，RNC)、节点B(node B，NB)、基站控制器(base station controller，BSC)、基站收发台(base transceiver station，BTS)、家庭基站(例如，home evolved nodeB，或home node B，HNB)、基带单元(baseBand unit，BBU)、传输点(transmitting and receiving point，TRP)、发射点(transmitting point，TP)、移动交换中心等。
AMF网元，主要进行移动性管理、接入鉴权/授权等功能。此外，还负责在UE与PCF间传递用户策略。
SMF网元，主要进行会话管理、PCF下发控制策略的执行、UPF的选择、UE互联网 协议(internet protocol，IP)地址分配等功能。
UPF网元，作为和数据网络的接口UPF，完成用户面数据转发、基于会话/流级的计费统计，带宽限制等功能。
UDM网元，主要负责管理签约数据、用户接入授权等功能。
UDR，主要负责签约数据、策略数据、应用数据等类型数据的存取功能。
NEF网元，主要用于支持能力和事件的开放。
AF网元，主要传递应用侧对网络侧的需求，例如，服务质量(Quality of Service，QoS)需求或用户状态事件订阅等。AF可以是第三方功能实体，也可以是运营商部署的应用服务，如IP多媒体子系统(IP Multimedia Subsystem，IMS)语音呼叫业务。
PCF网元，主要负责针对会话、业务流级别进行计费、QoS带宽保障及移动性管理、UE策略决策等策略控制功能。
NRF网元，可用于提供网元发现功能，基于其他网元的请求，提供网元类型对应的网元信息。NRF还提供网元管理服务，如网元注册、更新、去注册以及网元状态订阅和推送等。
AUSF网元：主要负责对用户进行鉴权，以确定是否允许用户或设备接入网络。
NWDAF网元，主要用于收集网络数据(包括终端设备数据、RAN设备数据、核心网数据以及第三方应用数据中的一种或者多种)，并提供网络数据分析服务，可以输出数据分析结果，供网络、网管及应用执行策略决策使用。NWDAF可以利用机器学习模型进行数据分析。3GPP Release 17中NWDAF的功能被分解，包括数据收集功能(或者数据收集逻辑功能(data collection function))、模型训练功能(或者机器学习模型训练逻辑功能(machine learning model training logical function))以及模型推理功能(或者分析逻辑功能(analytics logical function))。在数据收集功能、训练功能和推理功能分离的场景下，同一模型的数据收集功能、训练功能和推理功能可以分开部署在不同NWDAF实例中。部署数据收集功能的NWDAF(可以称为数据收集NWDAF或数据湖(data lake)或者数据存储功能网元(data repository function，DRF))可以用于从终端设备、RAN设备、核心网设备以及第三方应用中收集数据，部署训练功能的NWDAF(可以称为训练NWDAF)可以根据收集的数据进行模型训练得到训练后的模型，部署推理功能的NWDAF(简称为推理NWDAF)通过获取训练NWDAF提供的模型进行模型推理，提供数据分析服务。NWDAF可以是一个单独的网元，也可以与其他网元合设，例如：将NWDAF设置到PCF网元中。
DN，是位于运营商网络之外的网络，运营商网络可以接入多个DN，DN上可部署多种业务，可为终端设备提供数据和/或语音等服务。例如，DN是某智能工厂的私有网络，智能工厂安装在车间的传感器可为终端设备，DN中部署了传感器的控制服务器，控制服务器可为传感器提供服务。传感器可与控制服务器通信，获取控制服务器的指令，根据指令将采集的传感器数据传送给控制服务器等。又例如，DN是某公司的内部办公网络，该公司员工的手机或者电脑可为终端设备，员工的手机或者电脑可以访问公司内部办公网络上的信息、数据资源等。
图1中Nnwdaf、Nausf、Nnef、Npcf、Nudm、Naf、Namf、Nsmf、N1、N2、N3、N4，以及N6为接口序列号。这些接口序列号的含义可参见3GPP标准协议中定义的含义，在此不做限制。
需要说明的是，本申请实施例中，数据分析网元可以是图1所示的NWDAF网元，也 可以是未来通信系统中具有本申请中NWDAF网元的功能的其它网元，移动性管理网元可以是图1所示的AMF网元，也可以是未来通信系统中具有本申请中AMF网元的功能的其它网元，策略控制网元可以是图1所示的PCF网元，也可以是未来通信系统中具有本申请中PCF网元的功能的其它网元，用户面网元可以是图1所示的UPF网元，也可以是未来通信系统中具有本申请中UPF网元的功能的其它网元，应用功能网元可以是图1所示的AF网元，也可以是未来通信系统中具有本申请中AF网元的功能的其它网元，接入网设备可以是图1所示的RAN设备，也可以是未来通信系统中具有本申请中RAN设备的功能的其它网元。
为便于说明，本申请实施例中，以数据分析网元为NWDAF网元，移动性管理网元为AMF网元，策略控制网元为PCF网元，用户面网元为PCF网元，应用功能网元为AF网元，接入网设备为RAN设备为例进行说明。并且，将NWDAF网元进一步划分为数据湖(也称为数据收集NWDAF)、训练NWDAF网元和推理NWDAF网元。以及，以终端设备为UE为例进行说明。
为便于理解本申请实施例方案，下面先介绍与本申请实施例相关的技术。
一、跨域数据分析
业务提供方(如业务设备或AF)最关心它们的业务在5G网络中的业务体验，能够准确理解其自身的业务特征，因此业务提供方能够准确测量其业务体验从而有效监控业务质量。但是当前5G网络缺失业务体验评估机制，试图通过固定的QoS参数来保障丰富多变的5G业务，导致业务体验需求和网络资源无法精确匹配。
如图2所示，为基于NWDAF的业务体验评估流程示意图。其中，RAN设备、核心网(CN)以及业务提供方三个域的参数共同影响着业务体验。NWDAF确定业务流的业务体验数据分析结果，将业务体验数据分析结果发送给PCF，PCF根据业务体验数据分析结果以及业务体验要求之间的大小关系，判断业务流的业务体验能否得到满足。如果不能满足，PCF可以重新确定该业务流的QoS参数，基于新的QoS参数，网络可以提高该业务流的业务体验，从而使得NWDAF基于该业务流的新的业务体验得到的业务体验数据分析结果能够达到业务体验要求。
如表1所示，为影响业务体验的参数示例。
表1
Figure PCTCN2020141751-appb-000001
Figure PCTCN2020141751-appb-000002
假设表1中的数据都可以汇总到NWDAF，则NWDAF可以训练得到业务体验与RAN的数据、核心网(CN)的数据、AF的数据之间的关系模型，也即业务体验模型。基于业务体验模型，NWDAF可以确定RAN、核心网以及AF的新数据对应的业务体验数据分析结果。以线性回归为例，业务体验模型如下：
h(x)＝w 0x 0+w 1x 1+w 2x 2+w 3x 3+w 4x 4+w 5x 5...+w Dx D   (公式1)
其中，h(x)表示业务体验，一般地，h(x)的值越大表示业务体验越佳，x 0为1，x i(i＝1,2,...,D)表示RAN、CN以及AF的数据，D为数据的维度，w i(i＝0,1,2,...,D)为每个数据影响业务体验的权重。
事实上，RAN、CN、AF各自的私有数据无法集中到同一个NWDAF进行集中式训练，这是因为RAN、CN及AF的私有数据分别来自不同的域，其中，RAN的私有数据来自接入网域，CN的私有数据来自核心网域，AF的私有数据来自第三方设备域。
为实现对来自不同域的数据进行训练，可以借助于纵向联邦学习(Vertical Federated Learning，VFL)实现模型分布式训练。VFL作为一种新型的机器学习技术，解决各个参与方不愿意共享原始数据的情况下的模型训练与推理，适用于参与者训练样本标识(identification，ID)重叠较多，而参与者的数据特征重叠较少的情况。VFL联合多个参与者的共同样本的不同数据特征进行联邦学习，即各个参与者的训练数据是纵向划分的，因此称为纵向联邦学习。
基于纵向联邦学习的业务体验模型如下：
Figure PCTCN2020141751-appb-000003
其中，x i表示第i个样本数据，其中
Figure PCTCN2020141751-appb-000004
是第i个样本数据中分布在RAN上的私有数据，
Figure PCTCN2020141751-appb-000005
是第i个样本数据中分布在CN上的私有数据，
Figure PCTCN2020141751-appb-000006
是第i个样本数据中分布在AF上的私有数据，
Figure PCTCN2020141751-appb-000007
是第i个样本数据中由RAN、CN以及AF主动上报的公共数据(如表1中的RSRP/RSRQ/SINR、QoS flow Bit Rate/QoS flow Packet Delay/QoS flow Packet Error Rate、Buffer Size)，Θ A、Θ B、Θ C、Θ D分别是
Figure PCTCN2020141751-appb-000008
对应的模型参数。
值得说明的是，
Figure PCTCN2020141751-appb-000009
是由一个或者多个数据组成的数据向量，相应的，Θ A、Θ B、Θ C、Θ D是由一个或者多个模型参数组成的参数向量。
二、纵向联邦学习的训练过程
以线性回归算法为例，纵向联邦学习训练的过程如图3所示。
客户端A(Client A)上拥有数据集
Figure PCTCN2020141751-appb-000010
客户端B(Client B)上拥有数据集
Figure PCTCN2020141751-appb-000011
其中y i是标签数据，那么要训练的模型如下：
Figure PCTCN2020141751-appb-000012
假设线性回归用的目标函数如下，
Figure PCTCN2020141751-appb-000013
其中L为损失函数，具体如下：
Figure PCTCN2020141751-appb-000014
由于Client A上的原始数据D A与Client B上的D B无法汇总在一起，因此无法基于传统的集中式训练方法进行训练，但可以基于纵向联邦的训练方法进行训练，如下：
令
Figure PCTCN2020141751-appb-000015
则L变换如下：
Figure PCTCN2020141751-appb-000016
令
Figure PCTCN2020141751-appb-000017
那么
L＝L A+L B+L AB   (公式7)
令残差
Figure PCTCN2020141751-appb-000018
则L关于Θ A和Θ B的梯度如下：
Figure PCTCN2020141751-appb-000019
Figure PCTCN2020141751-appb-000020
相应地，模型参数更新如下：
Figure PCTCN2020141751-appb-000021
Figure PCTCN2020141751-appb-000022
那么纵向联邦学习的训练过程如下：
步骤1，Client A和Client B分别初始化模型参数Θ A和Θ B；
步骤2，Client A基于Θ A计算
Figure PCTCN2020141751-appb-000023
以及L A，然后将其发送给Client B；
步骤3，Client B基于Θ B计算
Figure PCTCN2020141751-appb-000024
进一步基于
Figure PCTCN2020141751-appb-000025
以及y i计算d i、L AB、L B，并最终基于L A、L AB、L B计算得到L。Client B将d i发送给Client A；
步骤4，Client A以及Client B各自基于d i分别计算
Figure PCTCN2020141751-appb-000026
以及
Figure PCTCN2020141751-appb-000027
然后在分别基于
Figure PCTCN2020141751-appb-000028
以及
Figure PCTCN2020141751-appb-000029
更新模型参数Θ A和Θ B。
其中，步骤2至步骤4循环执行，直到达到模型训练结束条件，比如迭代次数达到设定阈值(比如10000次)或损失函数L的取值小于设定阈值(比如0.001)。
通过上述技术，避免了各个域之间原始数据交互，也能训练得到业务体验模型。训练结束后，针对一组数据
Figure PCTCN2020141751-appb-000030
在推理阶段，Client A以及Client B分别基于训练好的模型参数Θ A和Θ B计算本地推理结果
Figure PCTCN2020141751-appb-000031
以及
Figure PCTCN2020141751-appb-000032
然后Client A将本地推理结果
Figure PCTCN2020141751-appb-000033
发送给Client B，由Client B最终确定推理结果
Figure PCTCN2020141751-appb-000034
三、模型训练时的数据集划分、NWDAF功能分解
参考图4，为数据集划分示意图。为了防止训练过程中发生过拟合问题，实际训练时，需要将数据集切分为：训练集、验证集以及测试集。其中，训练集用于根据不同的算法训练得到各个算法对应的模型，验证集用于验证各个模型的结果，并且在训练过程中可以不断更新算法来调整模型，根据验证结果可以选择出其中最好的模型，最终基于测试集以及最好的模型，确定模型的测试结果。
参考图5，为NWDAF功能分解架构示意图。NWDAF可以分解为训练NWDAF(负责模型训练)、推理NWDAF(负责数据分析结果的推理)以及数据湖(训练数据或者推理数据的收集与管控)。
四、NWDAF通过两两关联实现业务数据分析
通常情况下，NWDAF需要做RAN、CN以及AF之间的端到端的UE粒度(per UE)数据分析，然后就需要考虑将UE分布在RAN、CN以及AF上的数据进行关联。NWDAF可以通过关联信息两两关联确定UE的数据。示例性地，本申请实施例采用如下这一种思路：RAN-AMF-SMF-UPF-AF。
参考图6，为NWDAF通过两两关联得到UE的数据示意图。RAN和AMF在上报数据时，都携带时间戳(Timestamp)、为UE在N2接口上分配的RAN UE NGAP ID以及RAN全球唯一标识(Global RAN Node ID)，也即每条数据都用关联信息进行标识，该关联信息包括Timestamp、RAN UE NGAP ID和Global RAN Node ID。其中，NGAP是下一代应用层协议(Next Generation Application Protocol)的简称。因此，NWDAF通过Timestamp、RAN UE NGAP ID以及Global RAN Node ID来关联UE在RAN和AMF上的数据。
类似地，NWDAF通过Timestamp和AN隧道信息(AN Tunnel Info)关联UE在RAN和UPF上的数据，通过Timestamp和签约永久标识(Subscription Permanent Identifier，SUPI)关联UE在SMF和PCF上的数据，通过Timestamp和SUPI关联UE在AMF和SMF上的数据，通过Timestamp和UE IP关联UE在SMF和UPF上的数据，通过Timestamp和互联网协议五元组(Internet Protocol 5-tuple，IP 5-tuple)关联UE在AF和UPF上的数据。
示例性地，将RAN与AMF之间的关联信息用a表示，AMF与SMF之间的关联信息用b表示，SMF与UPF之间的关联信息用c表示，UPF与AF之间的关联信息用d表示，则：
RAN收集的UE的样本数据的格式为(a，UE的数据)；
AMF收集的UE的样本数据的格式为(a，b,UE的数据)，然后AMF向NWDAF上报的是(a，b,UE的数据)；
SMF收集的UE的样本数据的格式为(b，c,UE的数据)，然后SMF向NWDAF上报的是(b，c,UE的数据)；
UPF收集的UE的样本数据的格式为(c，d,UE的数据)，然后UPF向NWDAF上报的是(c，d,UE的数据)；
AF收集的UE的样本数据的格式为(d，UE的数据)，然后AF向NWDAF上报的是(d，UE的数据)。
NWDAF收到AMF、SMF、UPF、AF上报的样本数据为：(a，UE的数据)，(a，b， UE的数据)，(b，c，UE的数据)，(c，d，UE的数据)，(d，UE的数据)，然后NWDAF可以根据a与b之间的对应关系，b与c之间的对应关系，c与d之间的对应关系，将样本数据的格式都转成(a，UE的数据)。
本申请实施例提供一种业务体验模型的确定方法，该方法可以由核心网内的NWDAF或用于NWDAF的芯片执行。
本申请实施例中，第一数据集包括终端设备的业务在核心网网元上的数据，也即第一数据集用于表示终端设备的业务在核心网网元上的数据的集合。示例性地，这里的核心网网元可以是UPF、SMF、AMF、PCF等中的一个或多个。获取的核心网网元的数据可以包括表1所示的UPF的数据、AMF的数据等。第一关联信息用于将第一数据集与第一设备(可以是接入网设备或业务设备(也称为AF))上的第三数据集进行关联。第一关联信息还可以用于标识第一数据集的数据。第一关联信息对应第一数据集。示例性地，第一关联信息可以包括以下信息：第一设备的标识信息、第一设备为终端设备分配的标识信息、时间戳(Timestamp)。第一设备为终端设备分配的标识信息可以是第一设备为终端设备在第一接口上分配的标识信息，第一接口可以是第一设备与核心网网元之间的接口。这里的第一设备可以是接入网设备或AF。以第一设备为接入网设备为例，则第一关联信息可以包括：Timestamp、RAN UE NGAP ID和Global RAN Node ID。其中，Timestamp是时间戳，RAN UE NGAP ID是接入网设备为终端设备分配的标识信息，Global RAN Node ID是接入网设备的标识信息。
本申请实施例中，第三数据集包括终端设备的业务在第一设备(可以是接入网设备或AF)上的数据，也即第三数据集用于表示终端设备的业务在第一设备上的数据的集合。示例性地，获取的第一设备上的数据可以包括表1所示的RAN的数据或AF的数据。其中，第三数据集与第一数据集对应相同的业务。第三关联信息用于将第三数据集与NWDAF上的第一数据集进行关联。第三关联信息还可以用于标识第三数据集的数据。第三关联信息对应第三数据集。示例性地，第三关联信息可以包括以下信息：第一设备的标识信息、第一设备为终端设备分配的标识信息、时间戳(Timestamp)。第一设备为终端设备分配的标识信息可以是第一设备为终端设备在第一接口上分配的标识信息，第一接口可以是第一设备与核心网网元之间的接口。以第一设备为接入网设备为例，则第三关联信息可以包括：Timestamp、RAN UE NGAP ID和Global RAN Node ID。其中，Timestamp是时间戳，RAN UE NGAP ID是接入网设备为终端设备分配的标识信息，Global RAN Node ID是接入网设备的标识信息。
本申请实施例中，第二数据集包括第二关联信息在第一数据集中对应的数据，第二关联信息是第一关联信息和第三关联信息的交集。因此，第二数据集是第一数据集的真子集或者第二数据集与第一数据集相同。
示例性地，第一数据集包括10000份数据，对应的第一关联信息分别表示为ID-1至ID-10000。第三数据集包括9000份数据，对应的第三关联信息分别表示为ID-5001至ID-14000。因此，第二数据集包括5000份数据，对应的第二关联信息分别表示为ID-5001至ID-10000。
本申请实施例中，第四数据集是第二数据集的子集或者第四数据集与第二数据集相同。第四数据集包括训练集，训练集对应第四关联信息。可选的，第四数据集还包括验证集、 测试集。验证集对应第五关联信息，测试集对应第六关联信息。
如图7所示，本申请实施例提供一种业务体验模型的确定方法，该方法包括以下步骤：
步骤701，NWDAF获取第一数据集对应的第一关联信息。
步骤702，NWDAF根据第一关联信息获取第二数据集。
步骤703，NWDAF根据第一信息以及第二数据集确定第四数据集，第一信息包括第一设备的能力信息和/或NWDAF的能力信息。
这里的能力信息包括中央处理器(central processing unit，CPU)的算力、图形处理器(Graphics Processing Unit，GPU)的算力资源、内存资源、硬盘资源、NWDAF与第一设备(RAN设备或者AF)之间传输带宽资源、或者时延中的一项或多项。
步骤704，NWDAF根据第四数据集，确定业务的业务体验模型。
比如，NWDAF可以根据第四数据集进行纵向联邦学习，得到业务的业务体验模型。比如，参考图3，NWDAF可以是图3中的客户端A或客户端B，具体模型训练方法可以参考前述描述。
基于上述方案，NWDAF获取终端设备的业务在核心网网元上的数据，并与第一设备进行数据对齐，然后基于对齐后的数据进行模型训练得到准确的业务体验模型，NWDAF无需获取终端设备的业务在第一设备上的数据。
作为一种实现方法，在上述步骤702之前，NWDAF还可以从第二设备获取第一数据集。比如，NWDAF向第二设备发送第一请求，该第一请求携带第一设备的标识信息，第一请求用于请求第一数据集，或者理解为，第一请求用于请求获取核心网网元上的与第一设备相关的终端设备的业务数据。然后第二设备向NWDAF发送第一数据集。
在NWDAF获取到第一数据集的情况下，上述步骤702中，NWDAF获取第二数据集的方法比如可以是：NWDAF向第一设备发送第一关联信息，第一设备根据第三关联信息和第一关联信息确定第二关联信息，也即将第三关联信息和第一关联信息的交集确定为第二关联信息，然后第一设备向NWDAF发送第二关联信息，接着NWDAF根据第二关联信息以及第一数据集获取第二数据集，也即NWDAF获取第一数据集中对应于第二关联信息的数据，构成第二数据集。
在NWDAF获取到第一数据集的情况下，上述步骤702中，NWDAF获取第二数据集的方法比如还可以是：第一设备向NWDAF发送第三关联信息，NWDAF根据第三关联信息和第一关联信息确定第二关联信息，也即NWDAF将第三关联信息和第一关联信息的交集确定为第二关联信息，然后NWDAF根据第二关联信息以及第一数据集获取第二数据集，也即NWDAF获取第一数据集中对应于第二关联信息的数据，构成第二数据集。
作为另一种实现方法，NWDAF可以不需要从第二设备获取第一数据集，而是直接从第二设备获取第二数据集。当第二数据集的数据量少于第一数据集的数据量时，该方法可以减少数据传输量，进而减轻数据传输压力。基于该方法，则上述步骤701可以是：NWDAF向第二设备发送第二请求，第二请求携带第一设备的标识信息，第二请求用于请求第一数据集对应的第一关联信息，或者理解为，第二请求用于请求获取核心网网元上的与第一设备相关的终端设备的业务数据对应的关联信息，然后第二设备向NWDAF发送第一关联信 息。在NWDAF从第二设备获取到第一关联信息之后，NWDAF与第一设备进行关联信息对齐，也即NWDAF根据第一关联信息确定第二关联信息，具体方法可以参考前述描述。然后，NWDAF向第二设备发送第三请求，该第三请求携带第二关联信息，该第三请求用于请求第二关联信息对应的第二数据集，第二设备根据第二关联信息获取第二数据集并将第二数据集发送给NWDAF。也即，上述步骤702中，NWDAF是根据第一关联信息确定第二关联信息，然后根据第二关联信息从第二设备获取第二数据集。
作为一种实现方法，基于上述实现方法，在NWDAF从第二设备获取第一关联信息、第一数据集或第二数据集之前，NWDAF还向网元存储功能网元(如NRF)发送第四请求，第四请求用于请求第二设备的地址信息，然后网元存储功能网元向NWDAF发送第二设备的地址信息。第二设备可以是支持数据湖功能的NWDAF(也可以称为数据湖)，或者为支持数据收集协调功能的NWDAF，或者为支持数据收集功能的NWDAF。
作为一种实现方法，上述第四数据集包括训练集，该训练集对应第四关联信息，也即训练集的数据由第四关联信息。该训练集是第四数据集的真子集或者训练集与第四数据集相同。训练集用于根据不同的算法训练得到各个算法对应的模型。因此上述步骤704可以是：NWDAF根据第四关联信息以及第四数据集获取训练集，NWDAF根据训练集确定至少一个候选业务体验模型，NWDAF从至少一个候选业务体验模型中确定业务的业务体验模型。示例性地，NWDAF基于算法1和训练集进行模型训练得到候选业务体验模型1，基于算法2和训练集进行模型训练得到候选业务体验模型2，基于算法3和训练集进行模型训练得到候选业务体验模型3。然后NWDAF从候选业务体验模型1、候选业务体验模型2和候选业务体验模型3中确定业务的业务体验模型。
作为一种实现方法，上述第四数据集还包括验证集，该验证集对应第五关联信息，也即验证集的数据由第五关联信息。该验证集是第四数据集的真子集或者验证集与第四数据集相同。验证集用于验证各个模型的结果，并且在训练过程中可以不断更新算法来调整模型，根据验证结果可以选择出其中最好的模型。NWDAF从至少一个候选业务体验模型中确定业务的业务体验模型，比如可以是：NWDAF根据第五关联信息以及第四数据集获取验证集，NWDAF根据验证集确定至少一个候选业务体验模型分别对应的验证结果，NWDAF根据至少一个候选业务体验模型分别对应的验证结果确定业务的业务体验模型。比如，将对应最好验证结果的候选业务体验模型确定为业务体验模型。再比如，将验证结果排序，取排名靠前的N(N为大于1的整数)个验证结果，从这N个验证结果中随机选择一个验证结果，将选择的验证结果对应的候选业务体验模型确定为业务体验模型。
作为一种实现方法，上述第四数据集还包括测试集，该测试集对应第六关联信息，也即测试集的数据由第六关联信息。该测试集是第四数据集的真子集或者测试集与第四数据集相同。测试集用于确定业务体验模型的测试结果。比如，NWDAF可以根据第六关联信息以及第四数据集获取测试集，然后NWDAF根据测试集确定业务的业务体验模型的测试结果。
作为一种实现方法，NWDAF可以将第四数据集划分为训练集、验证集和测试集，训练集、验证集和测试集相互之间没有交集。可选的，训练集、验证集和测试集的并集等于第四数据集。
作为另一种实现方法，NWDAF还可以将第四数据集划分为训练集、验证集和测试集，且训练集、验证集和测试集相互之间可以有交集。可选的，训练集、验证集和测试集的并集等于第四数据集。
作为一种实现方法，在上述步骤704之后，NWDAF还可以向第一设备发送训练集对应的第四关联信息，则第一设备可以根据第四关联信息和第三数据集确定第一设备上的训练集(也可以称为第五数据集)，并基于该第五数据集进行模型训练，得到第一设备上的对应上述业务的至少一个候选业务体验模型。
作为一种实现方法，在上述步骤704之后，NWDAF还可以向第一设备发送验证集对应的第五关联信息，则第一设备可以根据第五关联信息和第三数据集确定第一设备上的验证集(也可以称为第六数据集)，并基于该第六数据集对上述至少一个候选业务体验模型进行验证，并基于验证结果从至少一个候选业务体验模型中选择一个业务体验模型，作为第一设备上的对应上述业务的业务体验模型。
作为一种实现方法，在上述步骤704之后，NWDAF还可以向第一设备发送测试集对应的第六关联信息，则第一设备可以根据第六关联信息和第三数据集确定第一设备上的测试集(也可以称为第七数据集)，并基于该第七数据集对上述业务体验模型进行测试，得到第一设备上的对应上述业务的业务体验模型的测试结果。
其中，第一设备确定的业务体验模型可以称为第一设备的业务体验子模型，NWDAF确定的业务体验模型可以称为核心网的业务体验子模型。
基于上述方案，第一设备获取终端设备的业务在第一设备上的数据，并与NWDAF进行数据对齐，然后基于对齐后的数据进行模型训练得到准确的业务体验模型，第一设备无需获取终端设备的业务在核心网网元上的数据。
下面结合图8至图10对应的具体实施例，对上述图7所示的方法进行说明。
如图8所示，为本申请实施例提供的一种业务体验模型的确定方法示意图。基于该方法，在核心网侧存在数据湖，NWDAF根据RAN设备的地址信息从数据湖中取与该RAN设备对应的核心网内的UE的数据，这部分UE的数据可以与RAN设备中的UE的数据实现纵向联邦学习。
该方法包括以下步骤：
步骤801a，RAN设备收集并存储接入网内UE的数据。
RAN设备收集数据包括第三数据集。在一种可能的实现方式中，该第三数据集可以包括多个UE的数据，每个UE的数据对应一个关联信息(也称为第三关联信息)。
该第三关联信息为RAN与AMF之间的关联信息，该第三关联信息包含Global RAN Node ID、RAN UE NGAP ID和Timestamp。其中，Global RAN Node ID是RAN的标识，RAN UE NGAP ID是RAN设备为UE在N2接口上分配的标识，Timestamp为RAN设备收集UE的数据对应的时间戳。
步骤801b，数据湖收集并存储核心网内UE的数据。
比如，数据湖收集UPF、AMF、SMF、PCF等核心网网元上UE的数据。在一种可能的实现方式中，当核心网网元为UPF和/或AMF时，数据湖收集的UE的数据可以参考表1。
下面以核心网网元为AMF进行举例说明。为了确保接入网内的UE的数据与核心网内的UE的数据可以关联，数据湖在通过Namf_EventExposure_Subscribe服务操作向AMF订阅UE在AMF上的数据时，可以在服务操作中携带指示信息，该指示信息用于指示AMF上报的UE的数据包含的关联信息的类型。在一种可能的实现方式中，该关联信息包含Global RAN Node ID、RAN UE NGAP ID和Timestamp。在另一种可能的实现方式中，该关联信息包含Global RAN Node ID，AMF UE NGAP ID，Timestamp。
数据湖可以根据前述介绍的两两关联的方法，得到核心网上的UE的数据。最终，核心网内的UE的数据对应的关联信息均包含Global RAN Node ID、RAN UE NGAP ID和Timestamp。
步骤802，NWDAF向数据湖发送请求消息。相应地，数据湖收到该请求消息。
该请求消息携带Global RAN node ID、特定时间段，该请求消息用于请求数据湖反馈特定时间段的与该RAN设备相关的核心网上的UE的数据。其中，Global RAN node ID为RAN设备的标识信息。可选的，该请求消息中还包括第一指示信息，第一指示信息包括事件标识(Event ID)和网络功能类型(network function type，NF type)，该事件标识＝Correlated CN Dataset，网络功能类型包括AMF、UPF、PCF等。可选的，该请求消息中还包括第二指示信息，第二指示信息用于指示数据湖反馈特定时间段的与该RAN设备相关的核心网上的关联后的UE的数据，其中，数据湖关联UE在核心网域内各个网元上的数据的方法可以参考前述介绍的两两关联的方法。
作为一种实现方法，NWDAF根据本地运营商策略确定需要针对RAN设备训练业务体验模型，则NWDAF向数据湖发送上述请求消息。示例性的，业务提供方(即AF)向运营商反映，在某个特定区域内的某个特定时间段的业务的平均业务体验很差，NWDAF作为运营商的数据分析网元，决定基于纵向联邦学习方法，联合RAN设备和核心网，或者联合核心网和AF，或者联合核心网、RAN设备和AF，训练得到一个业务体验模型。
其中，NWDAF获取RAN设备的标识的方法比如可以是：NWDAF向运维操作维护(Operation,Administration and Management，OAM)设备查询该特定区域内部署的RAN设备的标识。由于NWDAF是从业务提供方获取的业务体验差信息，因此NWDAF可以获知AF的地址信息。
步骤803，数据湖向NWDAF发送响应消息，相应地，NWDAF收到该响应消息。
该响应消息携带与RAN设备相关的核心网内的UE的数据以及该UE的数据对应的关联信息，该关联信息包含Global RAN Node ID、RAN UE NGAP ID和Timestamp。
NWDAF从数据湖获取到的UE的数据称为第一数据集，NWDAF从数据湖获取到的UE的数据对应的关联信息称为第一关联信息。在一种可能的实现方式中，该第一数据集可以包括多个UE的数据，每个UE的数据对应一个关联信息(也称为第一关联信息)。
步骤804，NWDAF与RAN设备进行数据对齐，得到第二数据集。
作为一种实现方法，NWDAF可以将获取的第一关联信息发送至RAN设备，RAN设备根据第一关联信息以及第三关联信息，确定第二关联信息，第二关联信息是第一关联信息与第三关联信息的交集。然后RAN设备将第二关联信息发送给NWDAF。从而，RAN 设备可以根据第二关联信息以及第三数据集确定接入网内的候选数据集，NWDAF可以根据第二关联信息以及第一数据集确定核心网内的候选数据集(也称为第二数据集)，该第二数据集对应第二关联信息。
作为另一种实现方法，RAN设备可以将获取的第三关联信息发送至NWDAF，NWDAF根据第一关联信息以及第三关联信息，确定第二关联信息，第二关联信息是第一关联信息与第三关联信息的交集，然后将第二关联信息发送给RAN设备。从而，RAN设备可以根据第二关联信息确定接入网内的候选数据集，NWDAF可以根据第二关联信息以及第一数据集确定核心网内的候选数据集(也称为第二数据集)。
比如，RAN侧的候选数据集包括1000万条UE的数据如下所示：
UE1的数据a1：对应关联信息1；
UE2的数据a2：对应关联信息2；
UE3的数据a3：对应关联信息3；
UE4的数据a4：对应关联信息4；
……
NWDAF侧的候选数据集(即第二数据集)包括1000万条UE的数据如下所示：
UE1的数据b1：对应关联信息1；
UE2的数据b2：对应关联信息2；
UE3的数据b3：对应关联信息3；
UE4的数据b4：对应关联信息4；
……
步骤805a，NWDAF向RAN设备发送请求消息。相应地，RAN设备收到该请求消息。
该请求消息用于请求获取RAN设备在特定时间段内的可参与纵向联邦训练的能力信息。比如，该请求消息携带特定时间段信息。可选的，该请求消息携带指示信息，该指示信息用于指示获取RAN设备在特定时间段内的可参与纵向联邦训练的能力信息。
步骤805b，RAN设备向NWDAF发送RAN设备在特定时间段内的可参与纵向联邦训练的能力信息。相应地，NWDAF收到RAN设备在特定时间段内的可参与纵向联邦训练的能力信息。
步骤806，NWDAF根据RAN的能力信息、NWDAF的能力信息以及的NWDAF侧的候选数据集(即第二数据集)的大小，确定第四数据集。
NWDAF根据RAN的能力信息、NWDAF的能力信息以及第二数据集的大小，确定第二数据集是否可以全部参与训练，如果不是，则从第二数据集中选出部分UE的数据构成第四数据集。
示例性的，假设经过步骤804确定可以参与纵向联邦学习的第二数据集包含1000万个UE的数据，RAN设备上每10％的能力可以容纳100万条UE的数据参与训练，假设NWDAF确定RAN侧剩余80％的能力可以用于训练，则NWDAF可以从1000万条UE的数据中随机抽取800万条UE的数据参与联邦训练，也即选定的第四数据集包含该800万条UE的数据。
步骤807，NWDAF将第四数据集进行切分，得到训练数据集、验证数据集和测试数据集。
比如，在上述示例中，针对选定的800万条UE的数据，NWDAF可以从中随机切分成700万条-50万条-50万条，分别作为训练数据集、验证数据集和测试数据集。其中，训练数据集包含一个或多个数据，该训练数据集对应第四关联信息。验证数据集包含一个或多个数据，该验证数据集对应第五关联信息。测试数据集包含一个或多个数据，该测试数据集对应第六关联信息。
步骤808，NWDAF向RAN发送配置信息。相应地，RAN收到该配置信息。
该配置信息用于指示RAN侧用于纵向联邦学习的训练数据集、验证数据集和测试数据集分别对应的关联信息，即上述第四关联信息、第五关联信息和第六关联信息。从而，RAN可以根据该配置信息，从RAN侧的候选数据集选出部分UE的数据构成RAN侧的训练数据集、选出部分UE的数据构成RAN侧的验证数据集以及选出部分UE的数据构成RAN侧的测试数据集。
可选的，配置信息还包含纵向联邦学习所使用的算法类型、每个算法内部所设置的参数信息(如神经网络算法的层数、每一层所使用的激活函数、损失函数类型等)。
由于NWDFA与RAN都确定了纵向联邦学习的训练数据集、验证数据集和测试数据集，后续二者可以分别基于纵向联邦学习方法进行模型训练，分别得到业务体验模型。比如可以参考前述描述的方法，将RAN作为Client A，将NWDAF作为Client B，二者分别基于各自的训练数据集、验证数据集和测试数据集进行模型训练，分别得到业务体验模型。
基于上述方案，NWDAF获取终端设备的业务在核心网网元上的数据，并与RAN设备进行数据对齐，然后基于对齐后的数据进行模型训练得到准确的业务体验模型，NWDAF无需获取终端设备的业务在RAN设备上的数据。
需要说明的是，NWDAF还可以分不同阶段(如训练阶段、验证阶段以及测试阶段)分别向RAN设备发送训练数据集、验证数据集和测试数据集对应的关联信息，用于指示RAN设备进行模型训练或者模型验证或者模型测试。
需要说明的是，上述NWDAF可以是训练NWDAF，训练NWDAF在根据纵向联邦方法进行模型训练后得到业务体验模型，然后将业务体验模型传递给推理NWDAF，用于产生核心网侧的数据分析结果。推理NWDAF主要是基于训练好的业务体验模型以及新的数据确定数据分析结果。
如图9所示，为本申请实施例提供的一种业务体验模型的确定方法示意图。基于该方法，在核心网侧存在数据湖，NWDAF根据RAN设备的地址信息从数据湖中取与该RAN设备对应的核心网内的UE的数据，这部分UE的数据可以与RAN设备中的UE的数据实现纵向联邦学习。
该方法包括以下步骤：
步骤901a至步骤901b，同上述步骤801a至步骤801b，可参考前述描述。
步骤902，NWDAF向数据湖发送请求消息。相应地，数据湖收到该请求消息。
该请求消息携带Global RAN node ID、特定时间段，该请求消息用于请求数据湖反馈特定时间段的与该RAN设备相关的核心网内的UE的数据对应的关联信息。其中，Global RAN node ID为RAN设备的标识信息。可选的，该请求消息中还包括第一指示信息，第一指示信息包括事件标识(Event ID)和网络功能类型(network function type，NF type)，该 事件标识＝Sample ID list，网络功能类型包括AMF、UPF、PCF等。可选的，该请求消息中还包括第二指示信息，第二指示信息用于指示数据湖反馈特定时间段的与该RAN设备相关的核心网内的关联后的UE的数据对应的关联信息，其中，数据湖关联UE在核心网域内各个网元上的数据的方法可以参考前述介绍的两两关联的方法。
作为一种实现方法，NWDAF根据本地运营商策略确定需要针对RAN设备训练业务体验模型，则NWDAF向数据湖发送上述请求消息。示例性的，业务提供方(即AF)向运营商反映，在某个特定区域内的某个特定时间段的业务的平均业务体验很差，NWDAF作为运营商的数据分析网元，决定基于纵向联邦学习方法，联合RAN设备和核心网，或者联合核心网和AF，或者联合核心网、RAN设备和AF，训练得到一个业务体验模型。
其中，NWDAF获取RAN设备的标识的方法比如可以是：NWDAF向OAM设备查询该特定区域内部署的RAN设备的标识。由于NWDAF是从业务提供方获取的业务体验差信息，因此NWDAF可以获知AF的地址信息。
步骤903，数据湖向NWDAF发送响应消息，相应地，NWDAF收到该响应消息。
该响应消息携带与RAN设备相关的核心网内的UE的数据对应的关联信息，该关联信息包含Global RAN Node ID、RAN UE NGAP ID和Timestamp。可选的，该响应消息还携带与RAN设备相关的核心网内的UE的数据的总大小，或者携带与RAN设备相关的核心网内的数据中的每个UE对应的数据的大小。
NWDAF从数据湖获取到的关联信息称为第一关联信息。
步骤904，NWDAF与RAN设备进行数据对齐，得到第二关联信息。
作为一种实现方法，NWDAF可以将获取的第一关联信息发送至RAN设备，RAN设备根据第一关联信息以及第三关联信息，确定第二关联信息，第二关联信息是第一关联信息与第三关联信息的交集。然后RAN将第二关联信息发送给NWDAF。从而，RAN设备可以根据第二关联信息以及第三数据集确定接入网内的候选数据集。
作为另一种实现方法，RAN可以将获取的第三关联信息发送至NWDAF，NWDAF根据第一关联信息以及第三关联信息，确定第二关联信息，第二关联信息是第一关联信息与第三关联信息的交集，然后将第二关联信息发送给RAN设备。从而，RAN设备可以根据第二关联信息以及第三数据集确定接入网内的候选数据集。
步骤905，NWDAF向数据湖发送请求消息。相应地，数据湖收到该请求消息。
该请求消息携带第二关联信息，该请求消息用于请求获取第二关联信息对应的UE的数据。
步骤906，数据湖向NWDAF发送响应消息，相应地，NWDAF收到该响应消息。
该响应消息携带第二关联信息对应的与RAN设备相关的核心网内的UE的数据，该第二关联信息包含Global RAN Node ID、RAN UE NGAP ID和Timestamp。其中，第二关联信息对应的与RAN设备相关的核心网内的UE的数据构成的集合称为第二数据集。
因此，通过上述步骤902至步骤903，NWDAF可以从数据湖获取到核心网内的UE的数据对应的关联信息(即第一关联信息)，通过上述步骤905至步骤906，NWDAF可以获取到对齐后的第二关联信息对应的UE的数据(即第二数据集)。
步骤907a至步骤910，同上述步骤805a至步骤808，可参考前述描述。
由于NWDFA与RAN都确定了纵向联邦学习的训练数据集、验证数据集和测试数据集，后续二者可以分别基于纵向联邦学习方法进行模型训练，分别得到业务体验模型。比 如可以参考前述描述的方法，将RAN作为Client A，将NWDAF作为Client B，二者分别基于各自的训练数据集、验证数据集和测试数据集进行模型训练，然后Client A和Client B通过交互中间结果
Figure PCTCN2020141751-appb-000035
d i等信息，两个Client联合得到业务体验模型。其中，业务体验模型对应的参数Θ A保留在Client A上，参数Θ B保留在Client B上，分别用于本地推理。
基于上述方案，NWDAF获取终端设备的业务在核心网网元上的数据，并与RAN设备进行数据对齐，然后基于对齐后的数据进行模型训练得到准确的业务体验模型，NWDAF无需获取终端设备的业务在RAN设备上的数据。并且，该方案NWDAF没有从数据湖获取核心网内全部的UE的数据，还是获取对齐后的第二关联信息对应的UE的数据，因而可以减轻NWDAF与数据湖之间的传输压力。
需要说明的是，上述NWDAF可以是训练NWDAF，训练NWDAF在根据纵向联邦方法进行模型训练后得到业务体验模型，然后将业务体验模型传递给推理NWDAF，用于产生核心网侧的数据分析结果。推理NWDAF主要是基于训练好的业务体验模型以及新的数据确定数据分析结果。
如图10所示，为本申请实施例提供的一种业务体验模型的确定方法示意图。该方案中没有数据湖的参与，核心网侧的UE的数据分布在不同的NWDAF(具有数据收集功能)中，每个NWDAF在收集完核心网的UE的数据后，可以将核心网的UE的数据对应的RAN设备的信息注册在NRF上，辅助支持联邦学习训练的NWDAF通过NRF寻址到这些NWDAF，然后向这些NWDAF请求与某个RAN设备相关的核心网的UE的数据。
该方法包括以下步骤：
步骤1001，同上述步骤801a，可参考前述描述。
步骤1002，核心网侧的不同NWDAF收集并存储核心网内与同一个RAN设备相关的UE的数据。
其中，不同NWDAF可以收集相同或不同核心网网元上的与同一个RAN设备相关的UE的数据，比如，NWDAF1至NWDAF3收集AMF上的与RAN设备相关的UE的数据，NWDAF4至NWDAF7收集SMF上的与RAN设备相关的UE的数据，NWDAF8至NWDAF10收集UPF上的与RAN设备相关的UE的数据，等等。
步骤1003，核心网侧的不同NWDAF向NRF发送注册请求，相应地，NRF收到注册请求。
该注册请求用于请求NWDAF的标识信息注册至NRF，该NWDAF是存储有与关联信息对应的UE的数据的NWDAF。
作为一种实现方法，该注册请求携带Global RAN Node ID、Timestamp和NWDAF ID。
作为另一种实现方法，该注册请求携带Global RAN Node ID、RAN UE NGAP ID、Timestamp和NWDAF ID。
步骤1004，训练NWDAF向NRF发送请求消息。相应地，NRF收到请求消息。
该请求消息携带Global RAN node ID、特定时间段，该请求消息用于请求与该Global RAN node ID和该特定时间段对应的NWDAF ID，也即该NWDAF ID指示的NWDAF中存储有在该特定时间段内的、与该Global RAN node ID对应的RAN设备相关的UE的数据。
步骤1005，NRF向训练NWDAF发送响应消息。相应地，训练NWDAF收到响应消 息。
该响应消息中携带一个或多个NWDAF ID。
步骤1006，训练NWDAF向不同的NWDAF发送请求消息。相应地，不同的NWDAF收到请求消息。
该请求消息携带Global RAN node ID和特定时间段，该请求消息用于请求反馈特定时间段的与该RAN设备相关的UE的数据。可选的，该请求消息携带指示信息，该指示信息用于指示反馈特定时间段的与该RAN设备相关的UE的数据。
步骤1007，不同的NWDAF向训练NWDAF发送响应消息，相应地，训练NWDAF收到该响应消息。
该响应消息携带与RAN设备相关的UE的数据以及每个UE的数据对应的关联信息。其中，关联信息比如包含Global RAN Node ID、RAN UE NGAP ID和Timestamp，或者包含Global RAN Node ID、AN Tunnel Info和Timestamp，或者包含Global RAN Node ID、SUPI和Timestamp等等。
步骤1008，训练NWDAF处理来自不同的NWDAF的UE的数据。
比如，训练NWDAF将来自不同的NWDAF的UE的数据对应的关联信息统一为Global RAN Node ID、RAN UE NGAP ID和Timestamp，以及对UE的数据进行去重处理。
其中，训练NWDAF处理之后得到的UE的数据的集合称为第一数据集合，该第一数据集合对应的关联信息称为第一关联信息。
步骤1009至步骤1013，同上述步骤804至步骤808，可参考前述描述。
基于上述方案，NWDAF获取终端设备的业务在核心网网元上的数据，并与RAN设备进行数据对齐，然后基于对齐后的数据进行模型训练得到准确的业务体验模型，NWDAF无需获取终端设备的业务在RAN设备上的数据。并且，上述方案是针对不存在数据湖的场景，与某个RAN设备相关联的UE的数据可能存在于不同的NWDAF中，这样支持联邦学习的训练NWDAF可以根据RAN设备的标识分别向这些NWDAF请求UE的数据。
参考图11，为本申请实施例提供的一种通信装置示意图。该通信装置用于实现上述各实施例中对应数据分析网元的各个步骤，如图11所示，该通信装置1100包括关联信息获取单元1110、数据集获取单元1120、数据集确定单元1130和业务体验模型确定单元1140。可选的，还包括收发单元1150。
关联信息获取单元1110，用于获取第一数据集对应的第一关联信息，所述第一数据集包括业务在核心网网元上的数据；数据集获取单元1120，用于根据所述第一关联信息获取第二数据集，所述第二数据集包括第二关联信息在所述第一数据集中对应的数据，所述第二关联信息是所述第一关联信息和第三关联信息的交集，所述第三关联信息对应第三数据集，所述第三数据集包括所述业务在所述第一设备上的数据；数据集确定单元1130，用于根据第一信息以及所述第二数据集，确定第四数据集，所述第四数据集为所述第二数据集的子集或者全部，所述第一信息包括所述第一设备的能力信息和/或所述数据分析网元的能力信息；业务体验模型确定单元1140，用于根据所述第四数据集，确定所述业务的业务体验模型。
在一种可能的实现方法中，所述第四数据集包括训练集，所述训练集对应第四关联信息，所述业务体验模型确定单元1140，具体用于：根据所述第四关联信息以及所述第四数 据集，获取所述训练集；根据所述训练集，确定至少一个候选业务体验模型；从所述至少一个候选业务体验模型中确定所述业务的业务体验模型。
在一种可能的实现方法中，所述第四数据集还包括验证集，所述验证集对应第五关联信息，所述业务体验模型确定单元1140，具体用于：根据所述第五关联信息以及所述第四数据集，获取所述验证集；根据所述验证集，确定所述至少一个候选业务体验模型分别对应的验证结果；根据所述至少一个候选业务体验模型分别对应的验证结果，确定所述业务的业务体验模型。
在一种可能的实现方法中，所述第四数据集还包括测试集，所述测试集对应第六关联信息，所述业务体验模型确定单元1140，还用于：根据所述第六关联信息以及所述第四数据集，获取所述测试集；根据所述测试集，确定所述业务的业务体验模型的测试结果。
在一种可能的实现方法中，收发单元1150，用于向所述第一设备发送所述训练集对应的所述第四关联信息。
在一种可能的实现方法中，收发单元1150，用于：向第二设备发送第一请求，所述第一请求携带所述第一设备的标识信息，所述第一请求用于请求所述第一数据集；从所述第二设备接收所述第一数据集。
在一种可能的实现方法中，所述数据集获取单元1120，具体用于：通过所述收发单元1150向所述第一设备发送所述第一关联信息；通过所述收发单元1150接收来自所述第一设备的所述第二关联信息；根据所述第二关联信息以及所述第一数据集，获取所述第二数据集。
在一种可能的实现方法中，所述数据集获取单元1120，具体用于：通过所述收发单元1150从所述第一设备接收所述第三关联信息；根据所述第三关联信息和所述第一关联信息，确定所述第二关联信息；根据所述第二关联信息以及所述第一数据集，获取所述第二数据集。
在一种可能的实现方法中，所述关联信息获取单元1110，具体用于：通过所述收发单元1150向第二设备发送第二请求，所述第二请求携带所述第一设备的标识信息，所述第二请求用于请求所述第一数据集对应的第一关联信息；通过所述收发单元1150从所述第二设备接收所述第一关联信息。
在一种可能的实现方法中，所述数据集获取单元1120，具体用于：根据所述第一关联信息，确定所述第二关联信息；通过所述收发单元1150向所述第二设备发送第三请求，所述第三请求携带所述第二关联信息，所述第三请求用于请求所述第二数据集；通过所述收发单元1150从所述第二设备接收所述第二数据集。
在一种可能的实现方法中，所述收发单元1150，还用于：向网元存储功能网元发送第四请求，所述第四请求用于请求所述第二设备的地址信息；从所述网元存储功能网元接收所述第二设备的地址信息。
在一种可能的实现方法中，所述第二设备为支持数据湖功能的数据分析网元，或者为支持数据收集协调功能的数据分析网元，或者为支持数据收集功能的数据分析网元。
在一种可能的实现方法中，所述第一设备为接入网设备或者业务设备。
在一种可能的实现方法中，所述第一关联信息包括以下信息：所述第一设备的标识信息、所述第一设备为终端设备分配的标识信息、时间戳。
在一种可能的实现方法中，所述第一设备为所述终端设备分配的标识信息是所述第一 设备为所述终端设备在第一接口上分配的标识信息，所述第一接口是所述第一设备与所述核心网网元之间的接口。
可选地，上述通信装置还可以包括存储单元，该存储单元用于存储数据或者指令(也可以称为代码或者程序)，上述各个单元可以和存储单元交互或者耦合，以实现对应的方法或者功能。例如，处理单元1120可以读取存储单元中的数据或者指令，使得通信装置实现上述实施例中的方法。
应理解以上通信装置中单元的划分仅仅是一种逻辑功能的划分，实际实现时可以全部或部分集成到一个物理实体上，也可以物理上分开。且通信装置中的单元可以全部以软件通过处理元件调用的形式实现；也可以全部以硬件的形式实现；还可以部分单元以软件通过处理元件调用的形式实现，部分单元以硬件的形式实现。例如，各个单元可以为单独设立的处理元件，也可以集成在通信装置的某一个芯片中实现，此外，也可以以程序的形式存储于存储器中，由通信装置的某一个处理元件调用并执行该单元的功能。此外这些单元全部或部分可以集成在一起，也可以独立实现。这里所述的处理元件又可以成为处理器，可以是一种具有信号的处理能力的集成电路。在实现过程中，上述方法的各步骤或以上各个单元可以通过处理器元件中的硬件的集成逻辑电路实现或者以软件通过处理元件调用的形式实现。
在一个例子中，以上任一通信装置中的单元可以是被配置成实施以上方法的一个或多个集成电路，例如：一个或多个特定集成电路(application specific integrated circuit，ASIC)，或，一个或多个微处理器(digital singnal processor，DSP)，或，一个或者多个现场可编程门阵列(field programmable gate array，FPGA)，或这些集成电路形式中至少两种的组合。再如，当通信装置中的单元可以通过处理元件调度程序的形式实现时，该处理元件可以是通用处理器，例如中央处理器(central processing unit，CPU)或其它可以调用程序的处理器。再如，这些单元可以集成在一起，以片上系统(system-on-a-chip，SOC)的形式实现。
参考图12，为本申请实施例提供的一种通信装置示意图，用于实现以上实施例中数据分析网元的操作。如图12所示，该通信装置包括：处理器1210和接口1230，可选地，该通信装置还包括存储器1220。接口1230用于实现与其他设备进行通信。
以上实施例中数据分析网元执行的方法可以通过处理器1210调用存储器(可以是数据分析网元中的存储器1220，也可以是外部存储器)中存储的程序来实现。即，数据分析网元可以包括处理器1210，该处理器1210通过调用存储器中的程序，以执行以上方法实施例中数据分析网元执行的方法。这里的处理器可以是一种具有信号的处理能力的集成电路，例如CPU。数据分析网元可以通过配置成实施以上方法的一个或多个集成电路来实现。例如：一个或多个ASIC，或，一个或多个微处理器DSP，或，一个或者多个FPGA等，或这些集成电路形式中至少两种的组合。或者，可以结合以上实现方式。
具体的，图11中的关联信息获取单元1110、数据集获取单元1120、数据集确定单元1130、业务体验模型确定单元1140和收发单元1150的功能/实现过程可以通过图12所示的通信装置1200中的处理器1210调用存储器1220中存储的计算机可执行指令来实现。或者，图11中的关联信息获取单元1110、数据集获取单元1120、数据集确定单元1130和业务体验模型确定单元1140的功能/实现过程可以通过图12所示的通信装置1200中的处理器1210调用存储器1220中存储的计算机执行指令来实现，图11中的收发单元1150的 功能/实现过程可以通过图12中所示的通信装置1200中的接口1230来实现，示例性的，收发单元1150的功能/实现过程可以通过处理器调用存储器中的程序指令以驱动接口1230来实现。
本领域普通技术人员可以理解：本申请中涉及的第一、第二等各种数字编号仅为描述方便进行的区分，并不用来限制本申请实施例的范围，也表示先后顺序。“和/或”，描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。字符“/”一般表示前后关联对象是一种“或”的关系。“至少一个”是指一个或者多个。至少两个是指两个或者多个。“至少一个”、“任意一个”或其类似表达，是指的这些项中的任意组合，包括单项(个)或复数项(个)的任意组合。例如，a，b，或c中的至少一项(个、种)，可以表示：a，b，c，a-b，a-c，b-c，或a-b-c，其中a，b，c可以是单个，也可以是多个。“多个”是指两个或两个以上，其它量词与之类似。
应理解，在本申请的各种实施例中，上述各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本发明实施例的实施过程构成任何限定。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包括一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘(solid state disk，SSD))等。
本申请实施例中所描述的各种说明性的逻辑单元和电路可以通过通用处理器，数字信号处理器，专用集成电路(ASIC)，现场可编程门阵列(FPGA)或其它可编程逻辑装置，离散门或晶体管逻辑，离散硬件部件，或上述任何组合的设计来实现或操作所描述的功能。通用处理器可以为微处理器，可选地，该通用处理器也可以为任何传统的处理器、控制器、微控制器或状态机。处理器也可以通过计算装置的组合来实现，例如数字信号处理器和微处理器，多个微处理器，一个或多个微处理器联合一个数字信号处理器核，或任何其它类似的配置来实现。
本申请实施例中所描述的方法或算法的步骤可以直接嵌入硬件、处理器执行的软件单元、或者这两者的结合。软件单元可以存储于随机存取存储器(Random Access Memory，RAM)、闪存、只读存储器(Read-Only Memory，ROM)、EPROM存储器、EEPROM存 储器、寄存器、硬盘、可移动磁盘、CD-ROM或本领域中其它任意形式的存储媒介中。示例性地，存储媒介可以与处理器连接，以使得处理器可以从存储媒介中读取信息，并可以向存储媒介存写信息。可选地，存储媒介还可以集成到处理器中。处理器和存储媒介可以设置于ASIC中。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个或多个示例性的设计中，本申请所描述的上述功能可以在硬件、软件、固件或这三者的任意组合来实现。如果在软件中实现，这些功能可以存储与电脑可读的媒介上，或以一个或多个指令或代码形式传输于电脑可读的媒介上。电脑可读媒介包括电脑存储媒介和便于使得让电脑程序从一个地方转移到其它地方的通信媒介。存储媒介可以是任何通用或特殊电脑可以接入访问的可用媒体。例如，这样的电脑可读媒体可以包括但不限于RAM、ROM、EEPROM、CD-ROM或其它光盘存储、磁盘存储或其它磁性存储装置，或其它任何可以用于承载或存储以指令或数据结构和其它可被通用或特殊电脑、或通用或特殊处理器读取形式的程序代码的媒介。此外，任何连接都可以被适当地定义为电脑可读媒介，例如，如果软件是从一个网站站点、服务器或其它远程资源通过一个同轴电缆、光纤电脑、双绞线、数字用户线(DSL)或以例如红外、无线和微波等无线方式传输的也被包含在所定义的电脑可读媒介中。所述的碟片(disk)和磁盘(disc)包括压缩磁盘、镭射盘、光盘、数字通用光盘(英文：Digital Versatile Disc，简称：DVD)、软盘和蓝光光盘，磁盘通常以磁性复制数据，而碟片通常以激光进行光学复制数据。上述的组合也可以包含在电脑可读媒介中。
本领域技术人员应该可以意识到，在上述一个或多个示例中，本申请所描述的功能可以用硬件、软件、固件或它们的任意组合来实现。当使用软件实现时，可以将这些功能存储在计算机可读介质中或者作为计算机可读介质上的一个或多个指令或代码进行传输。计算机可读介质包括计算机存储介质和通信介质，其中通信介质包括便于从一个地方向另一个地方传送计算机程序的任何介质。存储介质可以是通用或专用计算机能够存取的任何可用介质。
以上所述的具体实施方式，对本申请的目的、技术方案和有益效果进行了进一步详细说明，所应理解的是，以上所述仅为本申请的具体实施方式而已，并不用于限定本申请的保护范围，凡在本申请的技术方案的基础之上，所做的任何修改、等同替换、改进等，均应包括在本申请的保护范围之内。本申请说明书的上述描述可以使得本领域技术任何可以利用或实现本申请的内容，任何基于所公开内容的修改都应该被认为是本领域显而易见的，本申请所描述的基本原则可以应用到其它变形中而不偏离本申请的发明本质和范围。因此，本申请所公开的内容不仅仅局限于所描述的实施例和设计，还可以扩展到与本申请原则和所公开的新特征一致的最大范围。
尽管结合具体特征及其实施例对本申请进行了描述，显而易见的，在不脱离本申请的精神和范围的情况下，可对其进行各种修改和组合。相应地，本说明书和附图仅仅是所附权利要求所界定的本申请的示例性说明，且视为已覆盖本申请范围内的任意和所有修改、变化、组合或等同物。显然，本领域的技术人员可以对本申请进行各种改动和变型而不脱 离本申请的范围。这样，倘若本申请的这些修改和变型属于本申请权利要求及其等同技术的范围之内，则本申请也意图包括这些改动和变型在内。

Claims (34)

  1. 一种业务体验模型的确定方法，其特征在于，包括：
    数据分析网元获取第一数据集对应的第一关联信息，所述第一数据集包括业务在核心网网元上的数据；
    所述数据分析网元根据所述第一关联信息获取第二数据集，所述第二数据集包括第二关联信息在所述第一数据集中对应的数据，所述第二关联信息是所述第一关联信息和第三关联信息的交集，所述第三关联信息对应第三数据集，所述第三数据集包括所述业务在第一设备上的数据；
    所述数据分析网元根据第一信息以及所述第二数据集，确定第四数据集，所述第四数据集为所述第二数据集的子集或者全部，所述第一信息包括所述第一设备的能力信息和/或所述数据分析网元的能力信息；
    所述数据分析网元根据所述第四数据集，确定所述业务的业务体验模型。
  2. 根据权利要求1所述的方法，其特征在于，所述第四数据集包括训练集，所述训练集对应第四关联信息，所述数据分析网元根据所述第四数据集确定所述业务的业务体验模型，包括：
    所述数据分析网元根据所述第四关联信息以及所述第四数据集，获取所述训练集；
    所述数据分析网元根据所述训练集，确定至少一个候选业务体验模型；
    所述数据分析网元从所述至少一个候选业务体验模型中确定所述业务的业务体验模型。
  3. 根据权利要求2所述的方法，其特征在于，所述第四数据集还包括验证集，所述验证集对应第五关联信息，所述数据分析网元从所述至少一个候选业务体验模型中确定所述业务的业务体验模型，包括：
    所述数据分析网元根据所述第五关联信息以及所述第四数据集，获取所述验证集；
    所述数据分析网元根据所述验证集，确定所述至少一个候选业务体验模型分别对应的验证结果；
    所述数据分析网元根据所述至少一个候选业务体验模型分别对应的验证结果，确定所述业务的业务体验模型。
  4. 根据权利要求3所述的方法，其特征在于，所述第四数据集还包括测试集，所述测试集对应第六关联信息，所述方法还包括：
    所述数据分析网元根据所述第六关联信息以及所述第四数据集，获取所述测试集；
    所述数据分析网元根据所述测试集，确定所述业务的业务体验模型的测试结果。
  5. 根据权利要求2至4任一项所述的方法，其特征在于，所述方法还包括：
    所述数据分析网元向所述第一设备发送所述训练集对应的所述第四关联信息。
  6. 根据权利要求1至5中任一项所述的方法，其特征在于，所述方法还包括：
    所述数据分析网元向第二设备发送第一请求，所述第一请求携带所述第一设备的标识信息，所述第一请求用于请求所述第一数据集；
    所述数据分析网元从所述第二设备接收所述第一数据集。
  7. 根据权利要求1至6中任一项所述的方法，其特征在于，所述数据分析网元根据所述第一关联信息获取第二数据集，包括：
    所述数据分析网元向所述第一设备发送所述第一关联信息；
    所述数据分析网元接收来自所述第一设备的所述第二关联信息；
    所述数据分析网元根据所述第二关联信息以及所述第一数据集，获取所述第二数据集。
  8. 根据权利要求1至6中任一项所述的方法，其特征在于，所述数据分析网元根据所述第一关联信息获取第二数据集，包括：
    所述数据分析网元从所述第一设备接收所述第三关联信息；
    所述数据分析网元根据所述第三关联信息和所述第一关联信息，确定所述第二关联信息；
    所述数据分析网元根据所述第二关联信息以及所述第一数据集，获取所述第二数据集。
  9. 根据权利要求1至5中任一项所述的方法，其特征在于，所述数据分析网元获取第一数据集对应的第一关联信息，包括：
    所述数据分析网元向第二设备发送第二请求，所述第二请求携带所述第一设备的标识信息，所述第二请求用于请求所述第一数据集对应的第一关联信息；
    所述数据分析网元从所述第二设备接收所述第一关联信息。
  10. 根据权利要求9所述的方法，其特征在于，所述数据分析网元根据所述第一关联信息获取第二数据集，包括：
    所述数据分析网元根据所述第一关联信息，确定所述第二关联信息；
    所述数据分析网元向所述第二设备发送第三请求，所述第三请求携带所述第二关联信息，所述第三请求用于请求所述第二数据集；
    所述数据分析网元从所述第二设备接收所述第二数据集。
  11. 根据权利要求6、9或10中任一项所述的方法，其特征在于，所述方法还包括：
    所述数据分析网元向网元存储功能网元发送第四请求，所述第四请求用于请求所述第二设备的地址信息；
    所述数据分析网元从所述网元存储功能网元接收所述第二设备的地址信息。
  12. 根据权利要求6、9至11中任一项所述的方法，其特征在于，所述第二设备为支持数据湖功能的数据分析网元，或者为支持数据收集协调功能的数据分析网元，或者为支持数据收集功能的数据分析网元。
  13. 根据权利要求1至12中任一项所述的方法，其特征在于，所述第一设备为接入网设备或者业务设备。
  14. 根据权利要求1至13中任一项所述的方法，其特征在于，所述第一关联信息包括以下信息：
    所述第一设备的标识信息、所述第一设备为终端设备分配的标识信息、时间戳。
  15. 根据权利要求14所述的方法，其特征在于，所述第一设备为所述终端设备分配的标识信息是所述第一设备为所述终端设备在第一接口上分配的标识信息，所述第一接口是所述第一设备与所述核心网网元之间的接口。
  16. 一种通信装置，其特征在于，包括：
    关联信息获取单元，用于获取第一数据集对应的第一关联信息，所述第一数据集包括业务在核心网网元上的数据；
    数据集获取单元，用于根据所述第一关联信息获取第二数据集，所述第二数据集包括第二关联信息在所述第一数据集中对应的数据，所述第二关联信息是所述第一关联信息和 第三关联信息的交集，所述第三关联信息对应第三数据集，所述第三数据集包括所述业务在第一设备上的数据；
    数据集确定单元，用于根据第一信息以及所述第二数据集，确定第四数据集，所述第四数据集为所述第二数据集的子集或者全部，所述第一信息包括所述第一设备的能力信息和/或所述数据分析网元的能力信息；
    业务体验模型确定单元，用于根据所述第四数据集，确定所述业务的业务体验模型。
  17. 根据权利要求16所述的装置，其特征在于，所述第四数据集包括训练集，所述训练集对应第四关联信息，所述业务体验模型确定单元，具体用于：
    根据所述第四关联信息以及所述第四数据集，获取所述训练集；
    根据所述训练集，确定至少一个候选业务体验模型；
    从所述至少一个候选业务体验模型中确定所述业务的业务体验模型。
  18. 根据权利要求17所述的装置，其特征在于，所述第四数据集还包括验证集，所述验证集对应第五关联信息，所述业务体验模型确定单元，具体用于：
    根据所述第五关联信息以及所述第四数据集，获取所述验证集；
    根据所述验证集，确定所述至少一个候选业务体验模型分别对应的验证结果；
    根据所述至少一个候选业务体验模型分别对应的验证结果，确定所述业务的业务体验模型。
  19. 根据权利要求18所述的装置，其特征在于，所述第四数据集还包括测试集，所述测试集对应第六关联信息，所述业务体验模型确定单元，还用于：
    根据所述第六关联信息以及所述第四数据集，获取所述测试集；
    根据所述测试集，确定所述业务的业务体验模型的测试结果。
  20. 根据权利要求17至19任一项所述的装置，其特征在于，所述装置还包括收发单元，用于向所述第一设备发送所述训练集对应的所述第四关联信息。
  21. 根据权利要求16至20中任一项所述的装置，其特征在于，所述装置还包括收发单元，用于：
    向第二设备发送第一请求，所述第一请求携带所述第一设备的标识信息，所述第一请求用于请求所述第一数据集；
    从所述第二设备接收所述第一数据集。
  22. 根据权利要求16至21中任一项所述的装置，其特征在于，所述装置还包括收发单元；
    所述数据集获取单元，具体用于：
    通过所述收发单元向所述第一设备发送所述第一关联信息；
    通过所述收发单元接收来自所述第一设备的所述第二关联信息；
    根据所述第二关联信息以及所述第一数据集，获取所述第二数据集。
  23. 根据权利要求16至21中任一项所述的装置，其特征在于，所述装置还包括收发单元；
    所述数据集获取单元，具体用于：
    通过所述收发单元从所述第一设备接收所述第三关联信息；
    根据所述第三关联信息和所述第一关联信息，确定所述第二关联信息；
    根据所述第二关联信息以及所述第一数据集，获取所述第二数据集。
  24. 根据权利要求16至20中任一项所述的装置，其特征在于，所述装置还包括收发单元；
    所述关联信息获取单元，具体用于：
    通过所述收发单元向第二设备发送第二请求，所述第二请求携带所述第一设备的标识信息，所述第二请求用于请求所述第一数据集对应的第一关联信息；
    通过所述收发单元从所述第二设备接收所述第一关联信息。
  25. 根据权利要求24所述的装置，其特征在于，所述数据集获取单元，具体用于：
    根据所述第一关联信息，确定所述第二关联信息；
    通过所述收发单元向所述第二设备发送第三请求，所述第三请求携带所述第二关联信息，所述第三请求用于请求所述第二数据集；
    通过所述收发单元从所述第二设备接收所述第二数据集。
  26. 根据权利要求21、24或25中任一项所述的装置，其特征在于，所述收发单元，还用于：
    向网元存储功能网元发送第四请求，所述第四请求用于请求所述第二设备的地址信息；
    从所述网元存储功能网元接收所述第二设备的地址信息。
  27. 根据权利要求21、24至26中任一项所述的装置，其特征在于，所述第二设备为支持数据湖功能的数据分析网元，或者为支持数据收集协调功能的数据分析网元，或者为支持数据收集功能的数据分析网元。
  28. 根据权利要求16至27中任一项所述的装置，其特征在于，所述第一设备为接入网设备或者业务设备。
  29. 根据权利要求16至28中任一项所述的装置，其特征在于，所述第一关联信息包括以下信息：
    所述第一设备的标识信息、所述第一设备为终端设备分配的标识信息、时间戳。
  30. 根据权利要求29所述的装置，其特征在于，所述第一设备为所述终端设备分配的标识信息是所述第一设备为所述终端设备在第一接口上分配的标识信息，所述第一接口是所述第一设备与所述核心网网元之间的接口。
  31. 一种通信装置，其特征在于，包括：处理器，所述处理器与存储器耦合，所述存储器用于存储程序或指令，当所述程序或指令被所述处理器执行时，使得所述装置执行如权利要求1至15任一项所述的方法。
  32. 一种芯片系统，其特征在于，包括：所述芯片系统包括至少一个处理器，和接口电路，所述接口电路和所述至少一个处理器耦合，所述处理器通过运行指令，以执行权利要求1至15任一项所述的方法。
  33. 一种计算机可读存储介质，其特征在于，包括指令，当其在计算机上运行时，使得计算机执行如权利要求1至15任一项所述的方法。
  34. 一种通信系统，其特征在于，包括:
    数据分析网元，用于执行如权利要求1至15任一项所述的方法；以及
    用于与所述数据分析网元进行通信的核心网网元。
PCT/CN2020/141751 2020-12-30 2020-12-30 业务体验模型的确定方法及通信装置 WO2022141279A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
CN202080103972.7A CN116097734A (zh) 2020-12-30 2020-12-30 业务体验模型的确定方法及通信装置
EP20967619.6A EP4262268A4 (en) 2020-12-30 2020-12-30 METHOD FOR DETERMINING A SERVICE EXPERIENCE MODEL AND COMMUNICATION DEVICE
PCT/CN2020/141751 WO2022141279A1 (zh) 2020-12-30 2020-12-30 业务体验模型的确定方法及通信装置
US18/344,628 US20230353465A1 (en) 2020-12-30 2023-06-29 Method for determining service experience model and communication apparatus

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/141751 WO2022141279A1 (zh) 2020-12-30 2020-12-30 业务体验模型的确定方法及通信装置

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/344,628 Continuation US20230353465A1 (en) 2020-12-30 2023-06-29 Method for determining service experience model and communication apparatus

Publications (1)

Publication Number Publication Date
WO2022141279A1 true WO2022141279A1 (zh) 2022-07-07

Family

ID=82260047

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/141751 WO2022141279A1 (zh) 2020-12-30 2020-12-30 业务体验模型的确定方法及通信装置

Country Status (4)

Country Link
US (1) US20230353465A1 (zh)
EP (1) EP4262268A4 (zh)
CN (1) CN116097734A (zh)
WO (1) WO2022141279A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019242664A1 (zh) * 2018-06-20 2019-12-26 华为技术有限公司 一种资源管理方法及装置
CN110972200A (zh) * 2018-09-30 2020-04-07 华为技术有限公司 通信方法和相关设备
CN110972193A (zh) * 2018-09-28 2020-04-07 华为技术有限公司 一种切片信息处理方法及装置
US20200358670A1 (en) * 2019-05-07 2020-11-12 Electronics And Telecommunications Research Institute Method and system for providing service experience analysis based on network data analysis

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019242664A1 (zh) * 2018-06-20 2019-12-26 华为技术有限公司 一种资源管理方法及装置
CN110972193A (zh) * 2018-09-28 2020-04-07 华为技术有限公司 一种切片信息处理方法及装置
CN110972200A (zh) * 2018-09-30 2020-04-07 华为技术有限公司 通信方法和相关设备
US20200358670A1 (en) * 2019-05-07 2020-11-12 Electronics And Telecommunications Research Institute Method and system for providing service experience analysis based on network data analysis

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
CHINA MOBILE: "KI #1, New Sol: NWDAF functionality Split", 3GPP DRAFT; S2-2004058, 3RD GENERATION PARTNERSHIP PROJECT (3GPP), MOBILE COMPETENCE CENTRE ; 650, ROUTE DES LUCIOLES ; F-06921 SOPHIA-ANTIPOLIS CEDEX ; FRANCE, vol. SA WG2, no. Elbonia; 20200601 - 20200612, 22 May 2020 (2020-05-22), Mobile Competence Centre ; 650, route des Lucioles ; F-06921 Sophia-Antipolis Cedex ; France , XP051890065 *
See also references of EP4262268A4 *

Also Published As

Publication number Publication date
CN116097734A8 (zh) 2024-05-21
EP4262268A4 (en) 2024-01-17
CN116097734A (zh) 2023-05-09
US20230353465A1 (en) 2023-11-02
EP4262268A1 (en) 2023-10-18

Similar Documents

Publication Publication Date Title
WO2020224463A1 (zh) 一种数据分析方法及装置
WO2020169086A1 (zh) 一种用户面路径的选择方法及装置
EP4132066A1 (en) Communication method, apparatus and system
US20130166759A1 (en) Apparatus, systems, and methods of ip address discovery for tunneled direct link setup
WO2021017550A1 (zh) 一种事件报告的发送方法、装置及系统
US20230070712A1 (en) Communication method, apparatus, and system
JP7360479B2 (ja) ロジカルｔｓｎブリッジの方法および装置
CN114667746A (zh) 无线通信系统中用于psa-upf重定位的装置和方法
WO2022061784A1 (zh) 通信方法、装置及系统
US20220263879A1 (en) Multicast session establishment method and network device
WO2022141295A1 (zh) 一种通信方法和装置
US20230289615A1 (en) Training a machine learning model
US20240048986A1 (en) Communication method and apparatus
WO2022141279A1 (zh) 业务体验模型的确定方法及通信装置
US20220215038A1 (en) Distributed storage of blocks in blockchains
WO2023060409A1 (zh) 感知控制方法、装置、设备、系统及存储介质
Wirtz et al. Interest-based cloud-facilitated opportunistic networking
Koné et al. Network resource management and core network slice implementation: A testbed for rural connectivity
WO2023159363A1 (zh) Ai业务数据的传输方法、装置、设备及存储介质
WO2024099016A1 (zh) 一种通信方法及装置
US20230370897A1 (en) Service data flow transmission method, communication apparatus, and communication system
WO2023213226A1 (zh) 一种授权的方法和装置
WO2023051631A1 (zh) 数据调用的方法和装置
WO2021159523A1 (zh) 通信方法、装置及系统
US20230239728A1 (en) Service data flow transmission method, communication apparatus, and communication system

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20967619

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2020967619

Country of ref document: EP

Effective date: 20230714

NENP Non-entry into the national phase

Ref country code: DE