Information processing device, system, and method

Abstract

To provide an information processing device capable of efficiently allocating communication resources even when communicating with a plurality of communication nodes at different frequencies, a system, and a method.SOLUTION: In a communication system 12 of a communication network system 100, a collection analysis device 40 comprises: a frequency measurement unit which measures the communication frequency between each of a plurality of communication nodes and a data network and generates observation data consisting of a binary time series representing the measured frequency; a clustering unit which classifies the observation data into one of a plurality of clusters on the basis of a mixture probability distribution with the maximum likelihood parameters estimated by machine learning; and a policy assignment unit which assigns the a policy rules associated with the classified cluster to each communication node.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to an information processing device, a system, and a method for allocating communication resources to a terminal device. [Background technology]

[0002] In recent years, communication technologies using machine learning have been proposed to improve communication quality in communication systems such as fixed communication systems and mobile communication systems (for example, fifth generation mobile communication systems). This type of communication technology is disclosed, for example, in Patent Document 1 (JP Patent Publication No. 2023-123991). Patent Document 1 discloses a technology that uses a machine learning model such as a neural network to allocate resources such as frequency bands and time for wireless communication in a mobile communication system. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2023-123991 A Summary of the Invention [Problem to be solved by the invention]

[0004] In a communication system that dynamically allocates communication resources according to the communication status with a large number of communication nodes, it is preferable to efficiently allocate communication resources to each communication node. In particular, when there is a possibility that communication frequency with an unspecified number of communication nodes (e.g., mobile terminals using a 5th generation mobile communication system) is high, efficient allocation of communication resources to each communication node becomes important.

[0005] However, when a communication system communicates using common communication resources, allocating the same amount of communication resources to a communication node that communicates infrequently and a communication node that communicates frequently at the same time is inefficient and results in a lot of waste from the perspective of effective use of communication resources.

[0006] In view of the above, an object of the present disclosure is to provide an information processing device, system, and method that enable efficient allocation of communication resources even when communicating with multiple communication nodes at different frequencies. [Means for solving the problem]

[0007] According to an embodiment of the present disclosure, an information processing device includes: a frequency measurement unit that measures a frequency of communication between each of a plurality of communication nodes and a data network, and generates observation data for each of the communication nodes, the observation data being a binary time series representing the measured frequency; a clustering unit that classifies the observation data into one of a plurality of clusters based on a mixture probability distribution having a maximum likelihood parameter estimated by machine learning; The system further comprises a policy assignment unit that assigns policy rules associated with the classified clusters to each of the communication nodes.

[0008] According to another embodiment of the present disclosure, there is provided a processor-implemented method, comprising: measuring a frequency of communication between each of a plurality of communication nodes and a data network, and generating observation data for each of the communication nodes, the observation data being a binary time series representing the measured frequency; classifying the observed data into one of a plurality of clusters based on a mixture probability distribution having a maximum likelihood parameter estimated by machine learning; and assigning a policy rule associated with the classified cluster to each of the communication nodes. [Brief description of the drawings]

[0009] [Figure 1] 1 is a block diagram illustrating a schematic configuration example of a communication network system according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a schematic diagram of an example hardware configuration for implementing components of a communication system according to an embodiment of the present disclosure. [Diagram 3] 11 is a flowchart illustrating an example of a procedure for a communication resource allocation process according to an embodiment of the present disclosure. [Figure 4] FIG. 2 is a schematic diagram for explaining frequency of user communication according to one embodiment of the present disclosure. [Diagram 5] 2 is a diagram illustrating observation data representing measured frequencies of packet communication signals according to one embodiment of the present disclosure. [Figure 6] 11 is a flowchart illustrating an example of a procedure for a parameter generation process according to an embodiment of the present disclosure. [Figure 7] 1 illustrates an example of a policy table according to one embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] [Description of the embodiments of the present disclosure] First, the contents of the embodiment of the present disclosure will be listed and described. One embodiment of the present disclosure has the following configuration.

[0011] [Configuration 1] A characteristic configuration of an information processing device for achieving the above-mentioned object is that it includes a frequency measurement unit (42) that measures the frequency of communication between each of a plurality of communication nodes and a data network, and generates observation data for each of the communication nodes, the observation data consisting of a binary time series representing the measured frequency, a clustering unit (43) that classifies the observation data into one of a plurality of clusters based on a mixed probability distribution having a maximum likelihood parameter estimated by machine learning, and a policy assignment unit (44) that assigns a policy rule corresponding to the classified cluster to each of the communication nodes.

[0012] [Configuration 2] Another characteristic configuration of the information processing device according to the present disclosure is that the mixed probability distribution is a mixed Bernoulli distribution.

[0013] [Configuration 3] Another characteristic configuration of the information processing device according to the present disclosure is that the machine learning is maximum likelihood estimation using an EM (Expectation Maximization) algorithm.

[0014] [Configuration 4] Another characteristic configuration of the information processing device according to the present disclosure is that it further includes a data storage unit (50) that stores a lookup table (53) that defines a correspondence between the clusters and the policy rules, and the policy assignment unit determines a policy rule to be assigned to each of the communication nodes by referring to the lookup table.

[0015] [Configuration 5] Another characteristic configuration of the information processing device according to the present disclosure is that the policy rule includes a communication speed to be assigned to each of the communication nodes.

[0016] [Configuration 6] Another characteristic configuration of the information processing device according to the present disclosure is that it further includes a machine learning unit (48) that performs machine learning based on a mixture probability distribution using a training data set consisting of binary time series to estimate maximum likelihood parameters of the mixture probability distribution.

[0017] [Configuration 7] Another characteristic configuration of the information processing device according to the present disclosure is that a policy rule is previously associated with each training data of the training data set.

[0018] [Configuration 8] Another characteristic configuration of the system according to the present disclosure is that it further comprises the information processing device according to claim 1 and a user plane function node that applies the assigned policy rule to each of the communication nodes.

[0019] [Configuration 9] A characteristic configuration of a method for achieving the above-mentioned object is that the method is executed by a processor and includes the steps of: measuring the frequency of communication between each of a plurality of communication nodes and a data network; and generating observation data for each of the communication nodes, the observation data consisting of a binary time series representing the measured frequency (S304); classifying the observation data into one of a plurality of clusters based on a mixed probability distribution having a maximum likelihood parameter estimated by machine learning (S306); and assigning a policy rule corresponding to the classified cluster to each of the communication nodes (S308).

[0020] [Configuration 10] Another characteristic configuration of the method according to the present disclosure is that it further comprises a step of determining a policy rule to be assigned to each of the communication nodes by referring to a lookup table (53) that defines a correspondence between the clusters and policies.

[0021] [Configuration 11] Another characteristic configuration of the method according to the present disclosure is that it further includes a step of performing machine learning based on a mixture probability distribution using a training dataset consisting of binary time series to estimate maximum likelihood parameters of the mixture probability distribution.

[0022] [Details of the embodiment of the present disclosure] Hereinafter, an embodiment of the present disclosure will be described with reference to the drawings. Note that the following description is merely an example, and is not intended to limit the technical scope of the present invention to the following embodiment. In the drawings, identical or similar elements are given identical or similar reference symbols, and duplicate descriptions of identical or similar elements may be omitted in the description of each embodiment. In addition, the features shown in each embodiment can be applied to other embodiments as long as they are not mutually inconsistent. However, the embodiment of the present disclosure is not necessarily limited to such an aspect. It will be clear to those skilled in the art that the embodiment of the present disclosure can take various aspects included in the scope defined in the claims.

[0023] 1 is a block diagram illustrating a schematic configuration example of a communication network system 100 according to an embodiment of the present disclosure. As illustrated in FIG. 1, the communication network system 100 includes terminal devices (UE: User Equipment) 111 to 113 that are communication nodes. N and base stations 201 to 20 capable of wireless communication with N These base stations 201 to 20 N and a communication system 21 connected to the UEs 111 to 111. N is an integer equal to or greater than 2. N is a mobile terminal used by a subscriber of the service of the communication network system 100. n ~11 N Examples of such devices include, but are not limited to, smartphones, mobile phones, tablet devices, wearable devices, and laptop computers.

[0024] Each entity shown in the communication network system 100 includes a network element that implements the main functions provided in the communication network system 100, such as a mobility management function, a session management function, a policy control function, a user data management function, a subscriber profile management function, and a user plane function. In addition to the illustrated network elements, other network elements may be used to implement some or all of the main functions of the communication network system 100. In addition to the illustrated network elements, other network elements may be included.

[0025] In this embodiment, the communication network system 100 is configured as a system that complies with the 5th Generation (5G) standard, but is not limited to this example. The communication network system 100 can be compliant with other communication standards such as the 4th Generation (4G) and the 6th Generation (6G). In the following, the communication network system 100 will be described as being compliant with the 5G standard.

[0026] The following describes each entity constituting the communication network system 100 shown in Fig. 1. In 5GC (5G core network), various network functions are defined as NFs (Network Functions) for each role.

[0027] 1, a communication system 12 of a communication network system 100 includes NFs (Network Functions) such as an AMF (Access and Mobility Management Function) 22, an SMF (Session Management Function) 24, a PCF (Policy Control Function) 26, a UDM (Unified Data Management) 28, a UDR (Unified Data Repository) 30, and a UPF (User Plane Function) 60. These NFs are connected to a logical communication bus, and the UPF 60 and the UDR 30 are further connected to a collection and analysis device 40.

[0028] In this example, the AMF 22 may be referred to as a mobility management node. The AMF 22 has a registration management function, a connection management function, and a function for managing the terminal devices 111 to 111. N The AMF 122 may have a plurality of AMFs 122.

[0029] In this example, the SMF 24 may be referred to as a session management node. The SMF 24 manages and controls user data sessions. For example, it establishes, maintains, and releases data sessions. The SMF 24 can be connected to the PCF 26, one or more AMFs 22 (1 in the illustrated example), and one or more UPFs 60 (1 in the illustrated example).

[0030] In this example, the PCF 26 may be referred to as a policy control node. The PCF 26 provides a policy control function, and manages various policies on the network, such as allocating network resources, controlling quality, and applying security policies.

[0031] In this embodiment, the UDM 28 may be referred to as a user data management node. The UDM 28 manages user context data and identity information, and provides security and access control to the service providers of the network.

[0032] In this embodiment, the UDR 30 may be referred to as a subscriber data management node. The UDR 30 registers the database and status of all subscribers who have contracts with the communication carrier.

[0033] In this example, the UPF 60 may be referred to as a user plane function node. The UPF 60 is a component that transmits and receives user data. N It handles the transfer of data sent or received by the UE111~11 and supports low latency and high bandwidth communications. N For example, the UPF 60 can receive data transmitted from the UEs 111 to 111 and transmit the data to an appropriate destination. N The UPF 60 can enable a connection between the UEs 111 to 111 and a data network (DN) 70 such as the Internet. N The packet communication signal PS between the DN 70 and the DN 70 is observed, and the observed packet communication signal PS is transmitted to the collection and analysis device 40.

[0034] Furthermore, the UPF 60 can apply the policy rule transmitted from the collection and analysis device 40 to each subscriber identification number. The policy rule is, for example, a PCC (Policy and Charging Control) rule. The policy rule includes restrictions on data usage, communication speed, setting of priority, application of security policy, etc. When the policy rule is set in the UPF 60, the UEs 111 to 11 N It is used to control the quality and communication speed of the communication services provided to the

[0035] The collection and analysis device 40 includes an allocation unit 41, a parameter generation unit 47, and a data storage unit 50.

[0036] The allocation unit 41 includes a frequency measurement unit 42, a clustering unit 43, and a policy allocation unit 44. The configuration of the allocation unit 41 will be described in detail later.

[0037] The parameter generating unit 47 includes a machine learning unit 48 that estimates maximum likelihood parameters of a mixture probability distribution by performing machine learning based on the mixture probability distribution using the training data set 51. The configuration of the parameter generating unit 47 will be described in detail later.

[0038] The data storage unit 50 stores a training data set 51, parameter data 52 indicating a parameter group of a mixture probability distribution having a maximum likelihood parameter estimated by machine learning using the training data set 51, a lookup table 53, and a policy table 54 that defines the correspondence between policy rules and subscriber identification numbers. The training data set 51 is a data set used for generating the parameter data 52 for the parameter generation unit 47.

[0039] The parameter generating section 47 stores the parameter data 52 at an appropriate timing in the data storage section 50. The allocation section 41 can read out the parameter data 52 and use it.

[0040] All or part of the components of the communication system 21 described above (AMF 22, SMF 24, PCF 26, UDM 28, UDR 30, collection and analysis device 40, and UPF 60) may be realized by one computer including one or more processors, or by multiple computers interconnected via communication paths. All or part of the components of the communication system 21 can be realized by one or more processors including one or more processing units that execute processing according to computer program codes (instructions) read from a non-volatile memory (computer-readable recording medium). For example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or an NPU (Neural Network Processing Unit) can be used as the processing unit.

[0041] 2 is a schematic diagram of an information processing device (computer) 200, which is an example of a hardware configuration for implementing the components of the communication system 12. The information processing device 200 includes a processor 201 including a plurality of processor cores μC, ..., μC, a random access memory (RAM) 202, a nonvolatile memory 203, a large capacity memory 204, an input / output interface 205, and a signal path 206. The signal path 206 is a bus for mutually connecting the processor 201, the RAM 202, the nonvolatile memory 203, the large capacity memory 204, and the input / output interface 205. The RAM 202 is a data storage area used when the processor 201 executes digital signal processing. The nonvolatile memory 203 has a data storage area in which the code (group of instructions) of a computer program executed by the processor 201 is stored.

[0042] UE111~11 N The UEs 111 to 111 are mobile terminals used by users who are subscribers of a mobile communication service provider, and can accept various input operations by the users. Nis equipped with a SIM (Subscriber Identity Module) and is realized, for example, as a mobile terminal such as a smartphone, a PDA (Personal Digital Assistant), a wearable terminal, a tablet computer, or a laptop computer (a so-called notebook computer).

[0043] Each UE111~11 N The SIM installed in each UE 111 stores a contract profile of a user who has a contract with a mobile communication carrier. The SIM's contract profile stores user subscriber information, and each UE 111 to 11 N This includes individual identifier information such as the subscriber identification number (IMSI: International Mobile Subscriber Identity) assigned to the line contract, the subscriber's telephone number (MSISDN: Mobile Subscriber International Subscriber Directory Number), and SIM card number (ICCID: Integrated Circuit Card Identifier).

[0044] Next, the allocation unit 41 of the collection and analysis device 40 will be described in detail with reference to Fig. 3 to Fig. 5. Fig. 3 is a flowchart showing an example of a procedure of a communication resource allocation process performed by the allocation unit 41.

[0045] In step S302, the frequency measurement unit 42 (FIG. 1) measures the frequency of each of the UEs 111 to 111 which are communication nodes monitored by the UPF 60. N The communication signal of user communication (eg, packet communication) of DN (data network) 70 is collected, and the frequency of the user communication over a predetermined period (eg, one month) is measured.

[0046] In step S302, the frequency measurement unit 42 generates observation data x n Generate.

[0047] FIG. 4 is a schematic diagram for explaining the frequency of user communications. N Packet communication signals PS1 to PS N 4, the vertical axis represents the UEs 111 to 111 corresponding to the subscriber identification numbers of all users who have a communication contract with a communication carrier. N The horizontal axis indicates time.

[0048] FIG. 4 shows a packet communication signal PS a ,…,PS b ,…,PS c ,…,PS d ,PS e ,…,PS f 4 is a time chart showing an example of packet communication signals observed at t=1, t=2, t=3, t=4, t=5, t=6, t=7, t=8, t=9, t=10, t=11, t=12, t=13, t=14, t=15, t=16, t=17, t=18, t=19, t=20, t=21, t=22, t=23, t=24, t=25, t=26, t=27, t=28, t=29, t=30, t=31, t=32, t=33, t=34, t=35, t=36, t=37, t=38, t=40, t=41, t=42, t=43, t=44, t=45, t=46, t=47, t=48, t=50, t=51, t=52, t=19, t=29, t=29, t=34, t=49, t=51, t=29, t=35, t=49, t=29, t=36, t=49, t=19, t=21, t=22, t=37, t=23, t=38, t=39, t=44, t=39, t=45, t=29, t=39, t=49, t=39, t=49, t=29, t=39, t=49, t=39, t=49, t=39, t=49,

[0049] FIG. 5 shows the packet communication signal PS f Observational data x representing the measured frequency of n In the example of FIG. 5, the packet communication signal PS f When the packet communication signal PS f When the observed data x is not observed, the observed data value is "0". n is expressed as a binary vector whose elements are binary variables that take the value of "0" or "1." In the example of Figure 5, the observed data x n is the vector of bits (1, 1, 0, 1, 1, 0.1.1, 1, 1, 0, 1, 1, 1, 1, 0, 1).

[0050] As described above, the parameter generating unit 47 stores the parameter data 52 indicating a parameter group of a mixture probability distribution having a maximum likelihood parameter estimated by machine learning in the data storage unit 50. The assigning unit 41 can read out the parameter data 52 from the data storage unit 50 and use the parameter data 52.

[0051] In step S306, the clustering unit 43 (FIG. 1) uses the parameter data 52 read from the data storage unit 50 to perform clustering (cluster classification) using a parameter set of the mixed probability distribution indicated by the parameter data 52. That is, the clustering unit 36 ​​classifies the observed data x n can be classified into one of multiple clusters.

[0052] In this embodiment, the observation data x n Since each of the binary variables that are elements of can be considered to follow a Bernoulli distribution, a mixed Bernoulli distribution can be used as the mixed probability distribution represented by the parameter data 52. As machine learning, a likelihood estimation method using the EM (Expectation-Maximization) algorithm can be used.

[0053] According to the maximum likelihood estimation method using the EM algorithm of the mixed Bernoulli distribution, the observed data x n The proportion of people who belong to the kth cluster is the burden rate γ nk It can be calculated as follows: nk is expressed by the following equation (1).

[0054]

number

[0055] Here, π k ,μ k are the maximum likelihood parameters estimated by the maximum likelihood estimation method using the EM algorithm.

[0056] Next, in step S306, the clustering unit 43 n The burden rate γ n1 ,…,γ nK The observed data x can be classified into the cluster that corresponds to the highest burden rate. n are grouped into clusters that share common characteristics.

[0057] Next, in step S308, the policy assignment unit 44 (FIG. 1) refers to the lookup table 53 and assigns the observed data x n The policy rule corresponding to the cluster into which the UEs 111 to 111 are classified is determined. The policy rule is, for example, a PCC rule, and specifies policy management, traffic management, session management, and the like in the mobile communication network. A correspondence relationship between the cluster and the policy rule is previously defined in a lookup table 53 in the data storage unit 50. As the policy rule, for example, N The communication speed to be assigned to the

[0058] Next, in step S310, the policy assignment unit 44 sets the policy rule (e.g., communication speed) determined for each cluster for each subscriber identification number. The policy assignment unit 44 also generates a policy table 54 that defines the correspondence between the set policy rule and the subscriber identification number, and transmits it to the UPF 60 via the UDR 30, UDM 28, PCR 26, and SMF 24. The policy assignment unit 44 instructs the UPF 60 of the policy rule determined for each cluster.

[0059] 7 illustrates an example of the policy table 54 according to an embodiment of the present disclosure. In the policy table 54, a correspondence relationship between a policy rule for each cluster and a subscriber identification number is defined. For example, the policy rule is NIn the case where the communication speeds are specified as above, policy 1 (communication speed 10M) corresponds to subscriber identification numbers x to xxxx, policy 2 (communication speed 15M) corresponds to subscriber identification numbers y to yyyy, and policy N (communication speed 50M) corresponds to subscriber identification numbers z to zzzz.

[0060] When the UPF 60 acquires the policy rule set for each subscriber identification number, the UPF 60 applies the determined policy rule to each of the UEs 111 to 11 N can be applied to.

[0061] As described above, the allocation unit 41 of the UPF 60 allocates the UEs 111 to 11 N Observation data x consisting of a binary time series representing the frequency of measurements for each communication node n Using the machine learning, we perform clustering (cluster classification) based on a mixture probability distribution with maximum likelihood parameters estimated by machine learning, and the observed data x n The policy rule corresponding to the cluster into which the UEs 111 to 11 are classified can be determined. N In the case of a communication speed to be assigned to a terminal device classified into a cluster with a high communication frequency, the communication speed can be set slower than that of a terminal device classified into a cluster with an average communication frequency. This makes it possible to suppress the occurrence of burst traffic by slowing down the communication speed of heavy users with a high communication frequency. N Even if each of the devices communicates (transmits and receives communication signals) at different frequencies, communication resources can be allocated efficiently.

[0062] Next, the parameter generating unit 47 (FIG. 1) will be described in detail below with reference to Fig. 6. Fig. 6 is a flowchart showing an example of a procedure for parameter generation processing by the parameter generating unit 47.

[0063] The training data set 51 (Fig. 1) is the training data x nThe training data set 51 may be generated by an algorithm that generates a random bit string of 0 and 1. The training data set 51 may be created from a packet communication signal observed by the UPF 60. The training data set 51 may be prepared for multiple subscribers, for example, a number that is a fraction of the number corresponding to all subscriber identification numbers.

[0064] Training data x n Each of the parameter data 51 is associated with an allocation content associated with each policy rule in advance. A method for generating the parameter data 52 based on the training data set 51 will be described below.

[0065] In step S602, the machine learning unit 48 (FIG. 1) reads out the training data set 51 from the data storage unit 50.

[0066] Next, in step S604, the machine learning unit 48 uses the read training data set 51 to execute machine learning based on a mixture probability distribution, and estimates maximum likelihood parameters of the mixture probability distribution.

[0067] Next, in step S606, the machine learning unit 48 stores a group of parameters of a mixture probability distribution having the estimated maximum likelihood parameters as parameter data 52 in the data storage unit 50.

[0068] Here, the procedure of the maximum likelihood estimation method using the EM algorithm of the mixed Bernoulli distribution executed by the machine learning unit 48 in step S604 will be described below.

[0069] N binary vectors x1, x2, …, x N A set of binary data sets X is denoted by K parameter vectors μ1, μ2, …, μ K The set of binary data sets X and parameter data sets M are expressed as follows (T is the transpose symbol):

[0070]

number

[0071] Here, the n-th binary vector x n and the kth parameter vector μ k Each of these is represented as a vector with D elements as follows:

[0072]

number

[0073] Binary vector x n Element x of n Each of (i) (i=1,2,…,D) is a binary variable that takes the value 0 or 1. Each binary vector x n are independently generated by the following mixed Bernoulli distribution:

[0074]

number

[0075] Here, π is a parameter data set consisting of K mixture ratios, which can be expressed as follows:

[0076]

number

[0077] The likelihood function P(X|M,π) is expressed by the following equation.

[0078]

number

[0079] The log-likelihood function P(X|M,π) is expressed by the following equation.

[0080]

number

[0081] The EM algorithm is a method for finding the maximum likelihood solution of the parameter data set M,π of a mixed Bernoulli distribution so that the likelihood (the likelihood function (X|M,π) or the log-likelihood function P(X|M,π)) is maximized.

[0082] First, the machine learning unit 48 executes an initial step. That is, the parameter data M={μ1, μ2, ..., μ K},π={π1,π2,…,π K} and initialize the likelihood.

[0083] Next, the E-step is executed. That is, a value called the Responsibility γ is calculated using the current parameter set M,π according to the following formula: nk Calculate the burden rate γ nk is a binary vector x n represents the proportion of clusters belonging to the kth cluster.

[0084]

number

[0085] Next, the machine learning unit 48 executes the M step. That is, the current burden rate γ nk Update the parameter set,M,π,using,.

[0086]

number

[0087] Next, the machine learning unit 48 judges whether a predetermined convergence condition is satisfied. Specifically, if the likelihood converges within a predetermined numerical range, if the parameter set M,π converges, or if both the likelihood and the parameter set M,π converge, it may be judged that the convergence condition is satisfied. If it is judged that the convergence condition is not satisfied, the process returns to step E and repeats the calculation. If the predetermined convergence condition is not obtained even after repeating the calculation a predetermined number of times or for a certain period of time, it may be possible to change any or all of the parameter set M,π and the number of clusters K, and execute the processing from the initial step.

[0088] Although the embodiment of the present invention has been described above, the above-mentioned embodiment of the invention is intended to facilitate understanding of the present invention and does not limit the present invention. The present invention may be modified or improved without departing from the spirit thereof, and the present invention naturally includes equivalents thereof. Furthermore, within the scope of solving at least a part of the above-mentioned problems or achieving at least a part of the effects, any combination of the embodiments and modifications is possible, and any combination or omission of each component described in the claims and specification is possible. [Explanation of symbols]

[0089] 100…System 111~11 N …User Equipment (UE) 201~20 N …Base station 21:Communication Systems 22...AMF 24...SMF 26…PCF 28…UDM 30…UDR 40...Collection and analysis equipment 41…Allocation section 42…Frequency measurement section 43…Clustering Department 44…Policy assignment section 47...Parameter generation section 48…Machine Learning Department 50…Data storage section 51…Training data 52...Parameter data 53...Lookup table 54...Policy table 60...UPF 70…Data network

Claims

1. a frequency measurement unit that measures a frequency of communication between each of a plurality of communication nodes and a data network, and generates observation data for each of the communication nodes, the observation data being a binary time series representing the measured frequency; a clustering unit that classifies the observation data into one of a plurality of clusters based on a mixture probability distribution having a maximum likelihood parameter estimated by machine learning; a policy assignment unit that assigns a policy rule associated with the classified cluster to each of the communication nodes; An information processing device comprising:

2. The information processing device according to claim 1 , wherein the mixed probability distribution is a mixed Bernoulli distribution.

3. The information processing device according to claim 1 , wherein the machine learning is maximum likelihood estimation using an Expectation Maximization (EM) algorithm.

4. a data storage unit that stores a lookup table that defines a correspondence relationship between the clusters and the policy rules; The information processing device according to claim 1 , wherein the policy assignment unit determines a policy rule to be assigned to each of the communication nodes by referring to the lookup table.

5. The information processing device according to claim 1 , wherein the policy rule includes a communication speed to be assigned to each of the communication nodes.

6. The information processing device according to claim 1 , further comprising a machine learning unit that performs machine learning based on a mixture probability distribution using a training data set consisting of a binary time series to estimate a maximum likelihood parameter of the mixture probability distribution.

7. The information processing apparatus according to claim 6 , wherein each training data of the training data set is previously associated with a policy rule.

8. The information processing device according to claim 1 ; The system further comprises a user plane function node that applies the assigned policy rules to each of the communication nodes.

9. 1. A processor-implemented method, comprising: measuring a frequency of communication between each of a plurality of communication nodes and a data network, and generating observation data for each of the communication nodes, the observation data being a binary time series representing the measured frequency; classifying the observed data into one of a plurality of clusters based on a mixture probability distribution having a maximum likelihood parameter estimated by machine learning; assigning a policy rule associated with the classified cluster to each of the communication nodes; A method comprising:

10. The method according to claim 9 , further comprising the step of determining a policy rule to be assigned to each of the communication nodes by referring to a lookup table that defines a correspondence between the clusters and policies.

11. The method of claim 9 or 10, further comprising performing machine learning based on a mixture probability distribution using a training dataset of binary time series to estimate maximum likelihood parameters of the mixture probability distribution.

Images