KR20210097504A - Method and Device for Managing Metadata of Base Station - Google Patents

Abstract

Disclosed are a method and apparatus for managing metadata of a base station.
According to an aspect of the present invention, in a method for managing metadata of a base station related to a first AP and a second AP located in a specific area, identification information of a base station included in a communication history of a terminal, identification of the first AP converting information and identification information of the second AP into a base station vector, a first vector, and a second vector, respectively; calculating the similarity between the base station vector and the first vector and the similarity between the base station vector and the second vector, respectively; calculating a standard score for the base station based on each similarity; and if the standard score is less than or equal to a preset value, modifying the meta data of the base station.

Description

Method and Device for Managing Metadata of Base Station

The present invention relates to a method and apparatus for managing meta data of a base station for managing meta data of a base station according to a degree of relevance between APs located in a specific area and a base station.

The content described in this section merely provides background information on the present invention and does not constitute the prior art.

Recently, each telecommunication service provider is expanding new base stations to accommodate the increasing number of subscribers. In addition, as the environment such as service coverage and base station load constantly changes, the operating environment of the base station is frequently changed, such as base station relocation and accommodation change.

There are cases in which an error occurs in metadata management of the base station, such as incorrect location information is input to the base station or the input is omitted. When an error occurs in the location information of the base station, it causes a problem of degrading the quality of the location-based service using the location information of the base station. Conventionally, when a new base station is additionally installed, a telecommunication service provider directly manages the meta data of the new base station, or manages the meta data of the new base station by using information of all base stations that can be reached by radio waves adjacent to the new base station. Despite the management of the telecommunication service provider, only information such as coverage and location of the base station is managed, and metadata management of the base station capable of providing a service to a specific area is insufficient. For example, in the case of a subway station, even though the location of the base station is far from the subway station, there is an error in the meta data of the base station that the subway station is included. In addition, there are cases where the number of base stations including subway stations in the metadata is small or absent.

As such, when the operating environment of a base station changes frequently, such as a base station relocation or accommodation change, telecommunication service providers need to understand the meta data of the base station to provide the optimal service to the terminals near the base station. In particular, in order to provide the best service to a terminal located in a specific area, it is necessary to accurately understand the metadata of the base station related to the specific area. Here, the meta data of the base station means various information about the base station, such as location information and identification information for each base station.

Therefore, there is a need for a method for determining whether or not there is an error in the meta data of the base station and updating the meta data of the base station determined in error to the correct meta data.

In the embodiments of the present invention, in order to provide an optimal service to a terminal located in a specific area, using an AP in a specific area, a base station with high relevance to a specific area and a base station with low relevance are found and metadata of each base station is stored. A main object is to provide a method and apparatus for managing metadata of a base station that can be modified.

According to one aspect of the present invention, in a method of operating a management apparatus for managing metadata of a base station according to a degree of relevance to a first AP and a second AP located within a specific area, identification information of a base station included in a communication history of a terminal , converting the identification information of the first AP and the identification information of the second AP into a base station vector, a first vector, and a second vector, respectively; calculating a total similarity of the base station to the first AP and the second AP based on the similarity between the base station vector and the first vector and the similarity between the base station vector and the second vector; calculating a standard score of the base station using the total similarity of the base station and the total similarity to the first AP and the second AP of other base stations including the specific region in metadata; and modifying the meta data of the base station according to whether the standard score of the base station satisfies a preset condition.

According to another aspect of this embodiment, in the metadata management server of the base station related to the first AP and the second AP located in a specific area, the communication history and commands between the first AP, the second AP, and the base station based on the terminal at least one memory for storing instructions; and at least one processor operable to manage the meta data of the base station by executing at least one instruction stored in the memory, wherein the at least one memory is stored in the communication history of the terminal through the at least one processor. The included base station identification information, the first AP identification information, and the second AP identification information are converted into a base station vector, a first vector, and a second vector, respectively, and the similarity between the base station vector and the first vector and the Another base station that calculates the total similarity of the base station to the first AP and the second AP based on the similarity between the base station vector and the second vector, and includes the total similarity of the base station and the specific area in metadata is set to calculate the standard score of the base station using the total similarity to the first AP and the second AP, and to modify the metadata of the base station according to whether the standard score of the base station satisfies a preset condition Provides metadata management server.

According to another aspect of this embodiment, in the training method of a language model for converting the identification information of the AP or the base station communicated by the terminal into a vector, the step of selecting a plurality of identification information listed in chronological order from the communication history of the terminal ; and adjusting a parameter such that when one piece of identification information from among the plurality of identification information is input to the language model, a probability of outputting another piece of identification information from among the plurality of identification information is increased. provides

As described above, according to an embodiment of the present invention, by using an AP in a specific area to find a base station having a high degree of relevance to a specific area and a base station having a low relevance to a specific area, and modifying the metadata of each base station, the metadata of the base station can be accurately managed and the optimal service can be provided to terminals located in a specific area.

1A and 1B are diagrams illustrating information used in a method for managing metadata of a base station using a terminal according to an embodiment of the present invention.
2 is a diagram illustrating a process of calculating a similarity of a base station according to an embodiment of the present invention.
3 is a flowchart illustrating a method for managing metadata of a base station according to an embodiment of the present invention.
4 is a diagram illustrating a result of managing metadata of a base station according to an embodiment of the present invention.

Hereinafter, some embodiments of the present invention will be described in detail with reference to exemplary drawings. In adding reference numerals to the components of each drawing, it should be noted that the same components are given the same reference numerals as much as possible even though they are indicated on different drawings. In addition, in describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

In addition, in describing the components of the present invention, terms such as first, second, A, B, (a), (b), etc. may be used. These terms are only for distinguishing the components from other components, and the essence, order, or order of the components are not limited by the terms. Throughout the specification, when a part 'includes' or 'includes' a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated. . In addition, terms such as '~ unit' and 'module' described in the specification mean a unit that processes at least one function or operation, which may be implemented as hardware or software or a combination of hardware and software.

1A and 1B are diagrams illustrating information used in a method for managing metadata of a base station using a terminal according to an embodiment of the present invention.

Referring to FIG. 1A , a method for managing metadata of a base station according to an embodiment of the present invention is provided for a specific area 100 , a terminal 110 , another base station 120 , a first base station 130 , and an access point (AP). : Access Point, 140), the second base station 150 and the server 160 can be used.

The terminal 110 provides a communication service by communicating with other base stations 120 , the first base station 130 , the AP 140 , and the second base station 150 while moving from the specific area 100 into the specific area 100 . can receive

The other base station 120 is a base station located outside the specific area 100 , and although it can communicate with the terminal 110 , it is an inappropriate base station to provide a communication service to the specific area 100 .

The first base station 130 is located outside the specific area 100 but is highly related to the specific area 100 , and provides a communication service to the terminal 110 even if the terminal 110 is located in the specific area 100 . It is an appropriate base station for

The AP 140 is a repeater located within a specific area 100 , and accurate location information is stored in the server 160 , and a specific area among other base stations 120 , the first base station 130 , and the second base station 150 . It is used to find a base station with a high degree of relevance to (100) and a base station with a low degree of relevance. According to an embodiment of the present invention, the AP 140 is not limited to a repeater, but may be replaced by a base station or a gateway.

The second base station 150 is a base station located within a specific area 100 , and is a base station connected after the terminal 110 communicates with the AP 140 , and provides a communication service to the terminal 110 located within the specific area 100 . It is an appropriate base station to provide. The second base station 150 according to an embodiment of the present invention is a base station located within a specific area 100 , and may be a base station to which the terminal 110 is connected before communicating with the AP 140 .

The server 160 includes a communication history of the terminal 110 and metadata about other base stations 120 , the first base station 130 , the AP 140 , and the second base station 150 . In this case, the meta data for the base station may be modified according to an embodiment of the present invention, and it is assumed that the meta data for the AP is correct. The meta data of the base station means various information about the base station, such as location information and identification information for each base station.

Table 1 is a table illustrating a meta data table of a base station when a specific area 100 is limited to a subway station.

Base Station Identification Information location line station name A (x1, y1) Line 2 Gangnam B (x2, y2) Line 2 Gangnam C (x3, y3) Line 2 Yeoksam

Referring to Table 1, metadata for base station A, base station B, and base station C are shown.

Table 2 is a table showing the meta data table of the AP in the specific area 100 when the specific area 100 is limited to a subway station.

AP
identification information location line station name go (x4, y4) Line 2 Gangnam me (x5, y5) Line 2 Gangnam all (x6, y6) Line 2 Yeoksam

Referring back to FIG. 1A , the terminal 110 receives a communication service from the first base station 130 outside the specific area 100 , and the communication history between the terminal 110 and the first base station 130 is the server 160 . is stored in The terminal 110 moves into a specific area 100 while communicating with the first base station 130 .

Referring to FIG. 1B , the terminal 110 connected to the first base station 130 is connected to the AP 140 in the specific area 100 after moving into the specific area 100 . Similarly, the communication history between the terminal 110 and the AP 140 is stored in the server 160 .

The terminal 110 connected to the AP 140 may be directly connected to the second base station 150 within a specific area 100 . In addition, the terminal 110 may go out of coverage of the AP 140 while moving within the specific area 100 and thus be connected to the second base station 150 . The communication history between the terminal 110 and the second base station 150 is stored in the server 160 .

The server 160 may store all communication histories of the terminal 110 .

Table 3 is a table illustrating the communication history of the terminal (X) stored in the server 160 through a series of processes.

terminal communication point base station AP X 2019-09-01 13:44:38 A X 2019-09-01 13:44:55 A X 2019-09-01 13:45:05 go X 2019-09-01 13:45:38 B

Table 4 is a table illustrating communication histories for a plurality of terminals stored in the server 160 through a series of processes.

communication point terminal X terminal Y terminal Z n base station A base station A base station B n+1 base station B AP is AP is n+2 AP is base station B base station B n+3 base station C AP I base station C

Referring to Table 4, it is assumed that an AP and an AP are located in different regions, and Table 4 illustrates identification information of a base station and an AP through which a plurality of terminals X, Y, and Z communicated according to a communication time point. Terminal X communicated with base station A, base station B, AP, and base station C in that order, terminal Y communicated with base station A, AP, base station B, and AP B in that order, and terminal Z communicated with base station B, AP, and base station B and base station C in the order of communication.

In the method for managing metadata of a base station according to an embodiment of the present invention, communication between the terminal 110 and the AP 140 is based on the time when the terminal 110 communicates with the AP 140 located in the specific area 100 . It is to find a base station connected to the terminal 110 before and after, and to determine it as a base station having a high degree of relevance to the specific area 100 . In Table 4, since terminal X, terminal Y, and terminal Z are connected to base station B before or after communication with the AP, base station B is determined as a base station with high relevance to a specific area where the AP is located.

2 is a diagram illustrating a process of calculating a similarity of a base station according to an embodiment of the present invention.

Referring to FIG. 2 , another base station 120 , a first base station 130 , a first AP 140 , a second base station 150 , and a second AP 200 having a communication history with the terminal 110 are is shown. The server 160 communicates identification information of the other base station 120 , the first base station 130 , the first AP 140 , the second base station 150 , and the second AP 200 with which the terminal 110 communicated. It is stored according to the point in time.

The server 160 converts the identification information of the other base station 120 , the first base station 130 , the first AP 140 , the second base station 150 , and the second AP 200 into vectors, respectively. At this time, the server 160 inputs identification information of the other base station 120 , the first base station 130 , the first AP 140 , the second base station 150 , and the second AP 200 to the trained language model. to create each vector. The server 160 converts the identification information of the first base station 130 into a first base station vector, and converts the identification information of the first AP 140 and the second AP 200 into a first vector and a second vector, respectively. The identification information of the second base station 150 is converted into a second base station vector, and the identification information of the other base station 120 is converted into another base station vector.

The trained language model converts each piece of identification information into a vector, but converts it into a vector in consideration of the order in which the terminal 110 communicated. For example, when the terminal 110 communicates with the first base station 130 and the second base station 150 before and after communicating with the first AP 140 , the language model is a first base station vector and a second base station vector and transform each identification information so that the first vectors have high similarity values.

The process of training the language model is to select a plurality of identification information listed in chronological order from the communication history of the terminal, select one identification information from among the plurality of identification information, and input one identification information from among the plurality of identification information. This is performed by adjusting the parameters of the model to increase the probability of outputting the identification information of the other one of the identification information. That is, identification information between the AP or the base station communicating with the terminal at a time similar in time in the communication history of the terminal is converted into a vector having a high similarity value. Conversely, by adjusting the parameters of the language model to reduce the probability that the language model outputs other identification information not included in the plurality of identification information when inputting one identification information among the plurality of identification information into the language model, the language model is are trained That is, identification information between an AP or a base station communicating with the terminal at a point in time that is not temporally adjacent in the communication history of the terminal is converted into a vector having a low similarity value.

In order to manage the meta data of the first base station 130 , the server 160 calculates a similarity between the first base station vector and the first vector, and calculates a similarity between the first base station vector and the second vector. The server 160 calculates the total similarity of the first base station 130 based on each similarity. The total similarity is an average value of similarities for APs included in a specific area of the base station.

The server 160 calculates a standard score of each base station based on at least one of the total similarity of the first base station 130 , the total similarity of the second base station 150 , and the total similarity of other base stations 120 . Here, the standard score is divided into a first standard score calculated using the total similarity of base stations included in the specific area 100 and a second standard score calculated using the total similarity of all base stations. Referring to FIG. 2 , the first standard score of the first base station 130 is calculated based on the total similarity of the first base station 130 and the total similarity of the second base station 150 . On the other hand, the second standard score of the first base station 130 is calculated based on the total similarity of the first base station 130 , the total similarity of the second base station 150 , and the total similarity of other base stations 120 .

The server 160 combines the conditions for the first standard score and the second standard score, determines that each base station is a base station related to a specific area 100 when the preset conditions are satisfied, and does not satisfy the preset conditions In this case, it is determined that the base station is not related to the specific area 100, and the metadata of each base station is modified accordingly. The preset condition may be variously combined using the first standard score and the second standard score. The preset condition according to an embodiment of the present invention may be a condition for any one of the first standard score and the second standard score.

If the first standard score and the second standard score of the first base station 130 do not satisfy a preset condition, the server 160 determines that the first base station 130 has low relevance to the specific area 100 . and, when the metadata of the first base station 130 includes the specific area 100 , the metadata of the first base station 130 is modified. When the specific area 100 is a subway station, the subway line and station name of the metadata of the first base station 130 is changed to a null value or modified to a different subway line and station name. On the other hand, when the first standard score and the second standard score of the first base station 130 satisfy a preset condition, the server 160 may maintain the metadata without modifying it.

On the other hand, when the first standard score and the second standard score of the other base station 120 satisfy a preset condition, the server 160 determines that the other base station 120 is highly related to the specific area 100 and , when the metadata of the other base station 120 does not include the specific region 100 , the metadata of the other base station 120 may be modified. That is, the specific area 100 information may be included in the metadata of the other base station 120 .

3 is a flowchart illustrating a method for managing metadata of a base station according to an embodiment of the present invention.

Referring to FIG. 3 , the metadata management server inquires a communication history between a base station and a first AP and a second AP included in a specific area based on the terminal ( S300 ).

The server converts the identification information of the base station communicating with the terminal, the identification information of the first AP, and the identification information of the second AP into a base station vector, a first vector, and a second vector, respectively, using the trained language model ( S302 ). In the training process of the language model, an embedding technique such as a word to vector among machine learning techniques may be used.

The server calculates the similarity between the base station vector and the first vector and the similarity between the base station vector and the second vector, respectively, and calculates a standard score for the base station (S304). Here, the standard score is an index indicating how related the base station is to a specific area.

Specifically, cosine similarity between vectors of natural language processing (NLP) may be used to calculate the similarity between vectors. However, the present invention is not limited thereto, and the similarity between vectors may mean a Euclidean distance between vectors. When using the cosine similarity between vectors according to an embodiment of the present invention, a non-negative sum of square may be used to calculate a high degree of similarity between vectors of identification information having a similar communication time with the terminal.

Equation 1 is an equation using a non-negative sum of square to calculate the total similarity for APs included in a specific area of the base station. This may be a criterion for determining whether the base station is a base station with high relevance to a specific area or a low base station.

In Equation 1, n is the number of APs included in a specific area, and cossim is the cosine similarity between one base station vector and the kth AP vector among n. In order to prevent the sum of the cosine similarities from being calculated as 0 and to give weight to the similarity, the sum of the squares of the cosine similarities of the base station for n APs is summed, and then the value divided by n is calculated as the total similarity for a specific area of the base station. do.

On the other hand, when cosine similarity is used to calculate the similarity between the base station vector and the first vector and between the base station vector and the second vector, the similarity range is -1 to 1, but between the base station vector and the first vector or the second vector. When a non-negative sum of square is used to calculate the similarity, the values of x range from 0 (x is a positive real number). In this case, there may be a large deviation between the total similarities of base stations in different specific regions. Therefore, when the total similarity to one base station is calculated, a standard score (Z-score) is calculated through normalization.

There are two methods for calculating the standard score of the base station. The first is to calculate the first standard score by using the total similarities to the specific area of the base stations including the specific area in the metadata. The second is to calculate the second standard score by using all of the above similarity and the total similarity of the base stations that do not include the specific area in the metadata to the specific area. According to an embodiment of the present invention, in the case of the first standard score, it can be used to remove a base station with low relevance among base stations that have been managed as being related to a specific area, and in the case of the second standard score, it is associated with a specific area It may be used to include a specific region in metadata of a base station with high relevance among base stations that have been managed as unrelated. However, the present invention is not limited thereto, and a standard score criterion serving as a meta data management criterion may be set in various ways.

As a result, the server obtains a standard score of the base station indicating how high the relevance between the base station and the specific area including the first AP and the second AP is.

When the standard score of the base station is less than or equal to a preset value, the server modifies the meta data of the base station (S306). When a specific area is included in the metadata of the base station, the specific area of the metadata may be stored as a null value or changed to another area. Meanwhile, when the standard score for a specific area of another base station satisfies a preset condition, the other base station may be stored as a base station related to the specific area by changing the area name of the metadata of the other base station to the specific area.

4A and 4B are diagrams for explaining a result of meta data management of a base station according to an embodiment of the present invention.

4A and 4B , as a result of metadata management of a base station according to an embodiment of the present invention, a first standard score (Z-score) and a second standard score of a plurality of base stations for Sadang Station, Gangnam Station, and Yeoksam Station are shown respectively.

In the case of sadang station, the number of base stations including sadang station in the existing metadata is equal to the total number of points in the remaining area, the missing area, and the additional area. After performing the method for managing base station metadata according to an embodiment of the present invention, as shown in FIG. 4 , base stations having a high standard score with Sadang Station remain or are added, and base stations having a low standard score with Sadang Station are dropped. Here, the residual means the state of the base station to maintain the metadata, and the dropout means the state of the base station that needs to modify the meta data.

In this way, only base stations with high relevance to Sadang Station, Gangnam Station, and Yeoksam Station are extracted separately, and when providing a communication service to terminals located in each subway station, when providing a service to a terminal using only the separately extracted base station, the terminal is the best Communication services can be provided.

Although it is described in FIG. 3 that steps S300 to S306 are sequentially executed, this is merely illustrative of the technical idea of an embodiment of the present invention. In other words, one of ordinary skill in the art to which an embodiment of the present invention pertains may change the order described in FIG. 6 or perform one of steps S300 to S306 within a range that does not depart from the essential characteristics of an embodiment of the present invention. Since the above process may be variously modified and modified by executing the above process in parallel, FIG. 3 is not limited to a time-series order.

Meanwhile, the processes illustrated in FIG. 3 can be implemented as computer-readable codes on a computer-readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data readable by a computer system is stored. That is, the computer-readable recording medium may be a non-transitory medium such as ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical data storage device, and also carrier wave (for example, , transmission over the Internet) and may further include a transitory medium such as a data transmission medium. In addition, the computer-readable recording medium is distributed in a network-connected computer system so that the computer-readable code can be stored and executed in a distributed manner.

The above description is merely illustrative of the technical idea of this embodiment, and various modifications and variations will be possible by those skilled in the art to which this embodiment belongs without departing from the essential characteristics of the present embodiment. Accordingly, the present embodiments are intended to explain rather than limit the technical spirit of the present embodiment, and the scope of the technical spirit of the present embodiment is not limited by these embodiments. The protection scope of this embodiment should be interpreted by the following claims, and all technical ideas within the equivalent range should be interpreted as being included in the scope of the present embodiment.

100: specific area 110: terminal
120: other base station 130: first base station
140: AP 150: second base station

Claims (12)

A method of operating a management device for managing metadata of a base station according to a degree of relevance to a first AP and a second AP located in a specific area, the method comprising:
converting the identification information of the base station, the identification information of the first AP, and the identification information of the second AP included in the communication history of the terminal into a base station vector, a first vector, and a second vector, respectively;
calculating a total similarity of the base station to the first AP and the second AP based on the similarity between the base station vector and the first vector and the similarity between the base station vector and the second vector;
calculating a standard score of the base station using the total similarity of the base station and the total similarity to the first AP and the second AP of other base stations including the specific region in metadata; and
modifying the metadata of the base station according to whether the standard score of the base station satisfies a preset condition;
A method of operating an apparatus for managing metadata of a base station, comprising: According to claim 1,
When the base station includes the specific area in the metadata, the standard score is the first standard score of the base station,
The method of operation is
calculating a total similarity of an external base station that does not include the specific region in the metadata to the first AP and the second AP;
calculating a second standard score of the base station using the total similarity of the base station, the total similarity of other base stations, and the total similarity of the external base station; and
When the first standard score and the second standard score of the base station do not satisfy a preset condition, the method further comprising removing the specific region included in the meta data of the base station. . 3. The method of claim 2,
When the first standard score and the second standard score of the external base station satisfy a preset condition, including the specific region in the metadata of the external base station. According to claim 1,
The converting step is
generating the base station vector, the first vector, and the second vector by inputting the identification information of the base station, the identification information of the first AP, and the identification information of the second AP into a trained language model;
The training method of the language model is,
selecting a plurality of pieces of identification information listed in chronological order from the communication history of the terminal; and
adjusting a parameter to increase a probability of outputting another piece of identification information from among the plurality of pieces of identification information when inputting one piece of identification information from among the plurality of pieces of identification information into the language model;
A method of operating an apparatus for managing metadata of a base station, comprising: 5. The method of claim 4,
The training method of the language model,
adjusting a parameter to reduce a probability of outputting other identification information not included in the plurality of identification information when one identification information among the plurality of identification information is input to the language model;
Method of operating the apparatus for managing metadata of a base station further comprising a. 6. The method of claim 5,
The plurality of identification information and the other identification information,
A method of operating an apparatus for managing metadata of a base station including at least one of identification information of an AP with which the terminal communicated and identification information of a base station. In the meta data management server of the base station related to the first AP and the second AP located in a specific area,
at least one memory for storing communication history and instructions between the first AP, the second AP, and a base station based on a terminal; and
at least one processor operative to manage metadata of the base station by executing at least one instruction stored in the memory;
The at least one memory through the at least one processor,
Converting the identification information of the base station, the identification information of the first AP, and the identification information of the second AP included in the communication history of the terminal into a base station vector, a first vector, and a second vector, respectively,
calculating the total similarity of the base station to the first AP and the second AP based on the similarity between the base station vector and the first vector and the similarity between the base station vector and the second vector;
calculating the standard score of the base station using the total similarity of the base station and the total similarity to the first AP and the second AP of another base station including the specific region in metadata;
A metadata management server configured to modify metadata of the base station according to whether the standard score of the base station satisfies a preset condition. 8. The method of claim 7,
When the base station includes the specific area in the metadata, the standard score is the first standard score of the base station,
The at least one memory is the at least one processor,
Calculating the total similarity to the first AP and the second AP of an external base station that does not include the specific region in the metadata;
calculating a second standard score of the base station using the total similarity of the base station, the total similarity of other base stations, and the total similarity of the external base station;
When the first standard score and the second standard score of the base station do not satisfy a preset condition, the metadata management server stores commands set to remove the specific region included in the meta data of the base station. 9. The method of claim 8,
The at least one memory is the at least one processor,
When the first standard score and the second standard score of the external base station satisfy a preset condition, the metadata management server stores commands set to include the specific region in the metadata of the external base station. 8. The method of claim 7,
The at least one memory is the at least one processor,
Store the instructions set to generate the base station vector, the first vector, and the second vector by inputting the identification information of the base station, the identification information of the first AP, and the identification information of the second AP into a trained language model; ,
The training method of the language model is,
selecting a plurality of pieces of identification information listed in chronological order from the communication history of the terminal; and
adjusting a parameter to increase a probability of outputting another piece of identification information from among the plurality of pieces of identification information when inputting one piece of identification information from among the plurality of pieces of identification information into the language model;
A metadata management server that includes. In the training method of the language model for converting the identification information of the AP or the base station communicated by the terminal into a vector,
selecting a plurality of pieces of identification information listed in chronological order from the communication history of the terminal; and
adjusting a parameter to increase a probability of outputting another piece of identification information from among the plurality of pieces of identification information when inputting one piece of identification information from among the plurality of pieces of identification information into the language model;
A training method of a language model comprising 12. The method of claim 11,
adjusting a parameter to decrease a probability of outputting other identification information not included in the plurality of identification information when one identification information among the plurality of identification information is input to the language model;
A training method of a language model further comprising a.

Images