KR20220150198A - METHOD AND APPARATUS FOR MATCHING MARRY INFORMATION USING PREFERENCE MODEL BASED ON Bi-LSTM - Google Patents

Abstract

Marriage information matching method and apparatus using a Bi-LSTM-based preference prediction model are disclosed. In one embodiment, the marriage information matching method using the Bi-LSTM-based preference prediction model comprises the steps of: receiving profile information and social activity information of a first member from a user terminal or a social activity server; predicting a propensity value of the first member with respect to at least one criterion by inputting the profile information and the social activity information into a propensity prediction model; inputting the profile information and social activity information into a preference prediction model, and predicting a preference value that is a preference range of the first member with respect to the at least one criterion; generating member matching coordinates for the at least one criterion; determining a counterpart member to be matched with the first member using the member matching coordinates; and transmitting the profile of the counterpart member to the user terminal.

Description

Marriage information matching method and apparatus using a Bi-LSTM-based preference prediction model

The following embodiments relate to a method and apparatus for matching marriage information, and more particularly, to a method and apparatus for matching marriage information using a Bi-LSTM-based preference prediction model.

In general, in the marriage information service, the business entity collects a certain amount of membership fee from the subscribing member, and based on the subscribed member DB, the business entity provides information about the introduction between members by matching method for each business operator.

The operating form of the existing marriage information service providers is based on the provision of information about membership management by the consultant in charge based on offline, and at this time, all documents required for identification verification by customers of marriageable age, such as a copy of family register, After visiting a marriage information company with a certificate of employment or graduation, the eligibility for membership is determined by a face-to-face interview with the marriage information company's staff (consultants, couple managers, counselors, cyber consultants, etc.).

Therefore, members receive services from the point of payment of a certain amount of membership fee determined by each business operator. A look at the introduction of information provided by the existing marriage information service providers is as follows.

First, members are classified by region/age, and consultants are designated accordingly, and a certain number of members are introduced within a certain period of time. At this time, if male and female members are matched, the consultant in charge provides the member with a brief profile (name, main building, age, occupation, height, parental status, etc.) and communicate the agreed-upon on time, etc.

Here, the matching of male and female members is made by limited information such as member's profile information and the know-how of the consultant in charge, so there is a limit in reflecting the member's tendency or preference. Therefore, member satisfaction may vary depending on the consultant's ability, and the consultant in charge has no choice but to dedicate a lot of time to matching the members.

Accordingly, there is a need for a solution that can match both male and female members by using a variety of information from members while minimizing the time the consultant in charge is put into work.

Matters described in this background section are prepared to promote understanding of the background of the invention, and may include matters that are not already known to those of ordinary skill in the art to which this technology belongs.

The following embodiments have been devised to solve the above-described problems, and an object of the present invention is to provide a method and apparatus for matching marriage information using a Bi-LSTM-based preference prediction model.

Problems to be solved by one embodiment are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the following description.

An embodiment for achieving the above object provides a marriage information matching method using a Bi-LSTM-based preference prediction model.

One embodiment is a marriage information matching method, comprising: receiving profile information and social activity information of a first member from a user terminal or a social activity server; inputting the profile information and social activity information into a propensity prediction model to predict a propensity value of the first member with respect to at least one criterion; inputting the profile information and social activity information into a preference prediction model to predict a preference value that is a preference preference range of the first member with respect to the at least one criterion; generating member matching coordinates for the at least one criterion; determining a counterpart member to be matched with the first member by using the member matching coordinates; and transmitting the profile of the counterpart member to the user terminal.

In an embodiment, the determining of the counterpart member determines a second member whose preference value is included in the preference value of the first member as the counterpart member.

An embodiment may include generating a second preference value of the first member including a first preference value that is the predicted preference value, wherein the determining of the relative member includes determining that the predicted preference value in N units of time is the second preference value. A third member, which is included in the second preference value of the first member, and whose direction of the tendency vector changed from N-1 unit time to N unit time is toward the first preference value, is determined as the counterpart member.

The social activity information of an embodiment includes at least one of call conversation information, message conversation information, and SNS activity information, and extracting a conversation keyword by using the call conversation information or the message conversation information; and extracting intimacy by using the SNS activity information.

In the propensity value prediction step, at least one of the conversation keyword and the intimacy is input to the propensity prediction model, and in the preference value prediction step, at least one of the conversation keyword and the intimacy is input to the preference prediction model.

An embodiment for achieving the above object provides a non-transitory recording medium in which a program for executing the marriage information matching method is recorded, which can be read by a computer.

An embodiment for achieving the above object provides a computer program recorded in a non-transitory recording medium in order to execute the marriage information matching method in the marriage information providing server.

An embodiment for achieving the above object provides a system for providing marriage information that provides a matching service between members by using the members' tendencies and preferences.

The marriage information providing system may include: a social activity server providing social activity information to a marriage information providing server or a user terminal; and a marriage information providing server that predicts the pre-registered members' tendencies and preferences by using the profile information and the social activity information, and determines matching between members.

The marriage information providing server receives the profile information and the social activity information of the first member from the user terminal or the social activity server, and inputs the profile information and the social activity information into a propensity prediction model to at least one predicting a propensity value of the first member with respect to a criterion, inputting the profile information and social activity information into a preference prediction model, to predict a preference value that is a propensity preference range of the first member with respect to the at least one criterion, , generates member matching coordinates for the at least one criterion, determines a counterpart member to be matched with the first member using the member matching coordinates, and transmits the counterpart member's profile to the user terminal.

One embodiment includes at least one processor; and a memory (memory) for storing instructions instructing the at least one processor to perform at least one operation (operation).

The at least one operation may include: receiving profile information and social activity information of the first member from a user terminal or a social activity server; predicting the propensity value of the first member with respect to at least one criterion by inputting the profile information and social activity information into a propensity prediction model; inputting the profile information and social activity information into a preference prediction model to predict a preference value that is a preference preference range of the first member with respect to the at least one criterion; generating member matching coordinates for the at least one criterion; determining a counterpart member to be matched with the first member by using the member matching coordinates; and transmitting the profile of the counterpart member to the user terminal.

According to the exemplary embodiments as described above, it is possible to match an optimal counterpart member suitable for a member based on his/her inclination and preference predicted using the member's profile information and social activity information.

And, according to an embodiment, by reflecting the member's tendency and preference, which varies for each unit time, the member's desired counterpart member may be matched in real time.

In addition, according to an embodiment, it is possible to provide various marriage information efficiently and conveniently by minimizing the work input time of the consultant in charge.

Effects of one embodiment are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is a diagram illustrating a system for providing marriage information according to an embodiment.
2 is a diagram illustrating a configuration of a marriage information providing server according to an embodiment.
3 is a diagram for explaining a propensity prediction model according to an embodiment.
4 is a diagram for describing a preference prediction model according to an embodiment.
5 is a conceptual diagram of a method of converting voice conversation information into text information according to an embodiment.
6 is a flowchart of a method for extracting conversational keywords according to an embodiment.
7 is a conceptual diagram of a language model according to an embodiment.
8 is a diagram illustrating social network information in which intimacy is reflected according to an embodiment.
9 is a diagram illustrating member matching coordinates according to an embodiment.
10 is a diagram for explaining a member matching method according to an embodiment.
11 is a diagram illustrating tendencies and preferences updated for each unit time according to an embodiment.
12 is a conceptual diagram illustrating a second preference according to an embodiment.
13 is a diagram for explaining a member matching method according to another embodiment.
14 is a diagram illustrating a configuration of a user terminal according to an embodiment.
15 is a diagram illustrating a wireless communication system that can be applied in a communication process according to an embodiment of the present invention.
FIG. 16 is a diagram illustrating a base station in the wireless communication system according to FIG. 15 .
17 is a diagram illustrating a terminal in the wireless communication system according to FIG. 15 .
18 is a diagram illustrating a communication interface in the wireless communication system according to FIG. 15 .

Advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only these embodiments allow the disclosure of the present invention to be complete, and those of ordinary skill in the art to which the present invention pertains. It is provided to fully inform those skilled in the art of the scope of the present invention, and the present invention is only defined by the scope of the claims.

The terminology used herein is for the purpose of describing the embodiments and is not intended to limit the present invention. In this specification, the singular also includes the plural, unless specifically stated otherwise in the phrase. As used herein, “comprises” and/or “comprising” does not exclude the presence or addition of one or more other components in addition to the stated components. Like reference numerals refer to like elements throughout, and "and/or" includes each and every combination of one or more of the recited elements. Although "first", "second", etc. are used to describe various elements, these elements are not limited by these terms, of course. These terms are only used to distinguish one component from another. Therefore, it goes without saying that the first component mentioned below may be the second component within the spirit of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used herein will have the meaning commonly understood by those of ordinary skill in the art to which this invention belongs. In addition, terms defined in a commonly used dictionary are not to be interpreted ideally or excessively unless clearly specifically defined.

Hereinafter, an embodiment will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a system 10 for providing marriage information according to an embodiment.

Referring to FIG. 1 , a marriage information providing system 10 according to an embodiment may include a marriage information providing server 100 , a user terminal 200 , and a social activity server 300 .

Communication between various entities included in the marriage information providing system 10 may be performed through a wired/wireless network. Wired/wireless networks may use standard communication technologies and/or protocols.

The marriage information providing server 100, the user terminal 200 and the social activity server 300 in the marriage information providing system 10 are, for example, a computer, an Ultra Mobile PC (UMPC), a workstation, a net-book (net-book). ), PDA (Personal Digital Assistants), portable computer, web tablet, wireless phone, mobile phone, smart phone, PMP (portable multimedia player), etc. As one of the electronic devices, it may include all electronic devices capable of installing and executing an application related to an embodiment. The electronic device may perform overall service operations such as, for example, configuration of a service screen, data input, data transmission/reception, data storage, etc. under the control of the application.

The marriage information providing server 100 may receive and store profile information from the user terminal 200 used by a member of the marriage information service. According to an embodiment, the profile information may include a member profile and feedback information.

A member profile is a variety of data that expresses a member, and may include, for example, the member's name, age, occupation, height, body type, living area, school, drinking status, parental status, ideal type information, etc., and the member is a user terminal ( 200), you can enter your own profile.

The feedback information is feedback data on the profile of the counterpart member, and the member may evaluate the profile of the counterpart member provided through the user terminal 200 and input the evaluation information into the user terminal 200 as feedback information.

The marriage information providing server 100 may receive and store the member's social activity information from the user terminal 200 or the social activity server 300 . The social activity information according to an embodiment is various data generated while the member uses the social service, and may include voice conversation information between members, message conversation information, and SNS activity information.

Here, the voice conversation information is data generated while making a call with the other member (here, the call includes both a voice call and a video call), and may be composed of a voice file including a call with the other member.

In addition, the message conversation information is data generated while having a conversation with a counterpart member through an instant messager, and may be composed of a text file including a message conversation with the counterpart member.

And the SNS activity information is data about the member's activity history within the SNS service, and the member's SNS friend list, the number of posts or comments made by the member, the number of writings, the number of comments, the number of feedback such as 'likes', and the SNS activity time and the like.

The marriage information providing server 100 may predict a member's disposition and preference through machine learning by using profile information and social activity information.

First, the marriage information providing server 100 may learn the propensity prediction model using the propensity learning data input from the manager, and may learn the preference prediction model using the preference learning data.

And the marriage information providing server 100 can predict the predicted tendency of the member by inputting the profile information and social activity information into the learned tendency prediction model, and input the profile information and social activity information into the learned preference prediction model of the member preferences can be predicted.

The marriage information providing server 100 may use a conversation keyword extracted from voice conversation information and message conversation information and intimacy extracted from SNS activity information as social activity information input to the tendency prediction model and the preference prediction model.

To this end, the marriage information providing server 100 may extract conversation keywords by inputting voice conversation information and message conversation information into a language model, and extract intimacy using the contact frequency and emotional state calculated from the SNS activity information.

The predicted propensity of one embodiment is propensity data representing members with respect to a criterion. Here, one criterion may be set by various criteria. For example, it may be a personality-related criterion such as introvert/extrovert, it may be an information collection-related criterion such as sensory/intuitive, and it may be a decision-making-related criterion such as thinking/emotion, It may be a criterion related to job preference, which is a preference for a stable job.

The marriage information providing server 100 may predict the member tendency based on a plurality of criteria. Hereinafter, in the present specification, the first disposition will be described as the disposition of the personality-related criteria, and the second disposition will be described as the disposition of the decision-making criterion. However, it is self-evident that the marriage information providing server 100 can additionally predict the Nth tendency, such as the third tendency and the fourth tendency, of the membership tendency.

In one embodiment, the predicted preference is data of the disposition range of the other member preferred by the member with respect to one criterion. Here, the preference may be predicted in a certain range from one criterion.

The marriage information providing server 100 may quantify the predicted member's disposition and preference. Marriage information providing server 100, for example, the minimum value of -100, the maximum value of 100 can be predicted by quantifying what kind of value the member's propensity has, and quantifying and predicting what range value the member's preference has. can

The marriage information providing server 100 may generate member matching coordinates by using the quantified member's disposition and preference. In one embodiment, the member matching coordinates are coordinates that can show the member's disposition and preference based on at least one disposition, and in the member matching coordinates, the member's disposition is shown as a point, and the member's disposition is shown as a range. can be

The marriage information providing server 100 may determine an opponent member to match with the target member by using the member's disposition and preference shown in the member matching coordinates. The marriage information providing server 100 may extract the other member whose disposition is included in the target member's preference, and match the target member with the other member.

The marriage information providing server 100 may predict the propensity and preference of the member for each unit time, and may update the member's propensity and preference predicted for each unit time to the member matching coordinates. The marriage information providing server 100 may use the updated member matching coordinates to extract the other member whose disposition is included in the target member's preference to match the target member with the other member.

The marriage information providing server 100 may transmit the matched partner member's profile to the target member's user terminal 200 .

The detailed configuration and functions of the marriage information providing server 100 will be described in detail in FIG. 2 below.

The user terminal 200 may receive profile information from a member and transmit the inputted profile information to the marriage information providing server 100 . As described above, the profile information of an embodiment may include a member profile and feedback information.

The user terminal 200 may output the profile of the other member received from the marriage information providing server 100 to the user.

The user terminal 200 may receive social activity information from the social activity server 300 and transmit it to the marriage information providing server 100 .

A detailed configuration and function of the user terminal 200 will be described in detail below with reference to FIG. 16 .

The social activity server 300 is a server that provides a social service to members, and the social activity server 300 may provide the member's social activity information to the marriage information providing server 100 or the user terminal 200 .

The social activity server 300 may include a call server 310 , a messenger server 320 , and an SNS server 330 .

Here, the call server 310 is a server configured to provide a voice/video call service between members, and the call server 310 provides voice information for conversation between members to the marriage information providing server 100 or the user terminal 200 . This may be, for example, a server of a telecommunication company (SKT, KT, LGT).

In addition, the messenger server 320 is a server configured to provide an instant message service between members, and the messenger server 320 may provide information about a conversation message between members to the marriage information providing server 100 or the user terminal 200, , for example, KakaoTalk, Line, and Whatsapp service operation server may correspond to this.

And, the SNS server 330 is a server configured to provide SNS services to members, and may provide SNS activity information of members to the SNS server 330, the marriage information providing server 100 or the user terminal 200, for example. As such, Instagram, Facebook, and KakaoStory service operating servers may be included.

2 is a diagram illustrating the configuration of the marriage information providing server 100 according to an embodiment.

As shown in FIG. 2 , the marriage information providing server 100 includes a communication unit 110 , an input unit 120 , an output unit 130 , a memory 140 , a power supply unit 150 and a control unit 160 . can

The configurations shown in FIG. 2 are exemplary diagrams for implementing embodiments of the present invention, and appropriate hardware/software configurations at a level obvious to those skilled in the art may be additionally included in the marriage information providing server 100 .

The communication unit 110 may communicate with an external device through various communication methods. For example, the communication unit 110 may transmit and receive various data by performing communication with the external user terminal 200 and the social server 300 .

The input unit 120 may receive various inputs from an administrator who is a user of the marriage information providing server 100 and transmit it to the control unit 160 . In particular, the input unit 120 may include a touch sensor, a (digital) pen sensor, a pressure sensor, a key, or a microphone. The touch sensor may use, for example, at least one of a capacitive type, a pressure sensitive type, an infrared type, and an ultrasonic type. The (digital) pen sensor may be, for example, a part of a touch panel or may include a separate recognition sheet. The key may include, for example, a physical button, an optical key, or a keypad. The microphone is a configuration for receiving the manager's voice, and may be provided inside the marriage information providing server 100, but this is only an embodiment, and is provided outside the marriage information providing server 100 to provide a marriage information server ( 100) may be electrically connected to.

The output unit 130 may provide various screens. For example, the output unit 130 may output member-related information to the manager.

The memory 140 may store commands or data related to at least one other component of the marriage information providing server 100 . For example, the memory 140 may store instructions instructing the processor (or the controller 160 ) to perform at least one operation. Various operations of the server 100 to be described later may be included in the at least one operation. In particular, the memory 140 may be implemented as a non-volatile memory, a volatile memory, a flash-memory, a hard disk drive (HDD), or a solid state drive (SSD). The memory 140 is accessed by the controller 160 , and reading/writing/modification/deletion/update of data by the controller 160 may be performed. In the present invention, the term "memory" refers to a memory 140, a ROM (not shown) in the control unit 160, a RAM (not shown), or a memory card (not shown) mounted in the marriage information providing server 100 (eg, micro SD card, memory stick). In addition, programs and data for configuring various screens to be displayed on the display area of the output unit 130 may be stored in the memory 140 .

The power supply unit 150 receives external power and internal power under the control of the control unit 160 to supply power required for operation of each component.

The control unit 160 is electrically connected to the communication unit 110 , the input unit 120 , the output unit 130 , the memory 140 and the power supply unit 150 , and controls the overall operation and function of the marriage information providing server 100 . can be controlled For example, the controller 160 may be referred to as a processor. In particular, the controller 160 may provide a marriage information providing service using various modules stored in the memory 140 .

It will be apparent that various operations on the marriage information providing server 100 described below may be performed under the control of the controller 160 .

The controller 160 may control the communication unit 110 to receive profile information from the user terminal 200 used by a member of the marriage information service and store the profile information in the memory 140 .

As described above, the profile information according to an embodiment may include a member profile and feedback information. A member profile is a variety of data that expresses a member, and may include, for example, the member's name, age, occupation, height, body type, living area, school, drinking status, parental status, ideal type information, etc., and the member is a user terminal ( 200), you can enter your own profile. In addition, the feedback information is feedback data on the profile of the counterpart member, the member may evaluate the profile of the counterpart member provided through the user terminal 200 , and input the evaluation information into the user terminal 200 as feedback information.

The controller 160 may control the communication unit 110 to receive the member's social activity information from the user terminal 200 or the social activity server 300 and store it in the memory 140 .

As described above, the social activity information according to an embodiment is various data generated while the member uses the social service, and may include voice conversation information, message conversation information, and SNS activity information between members. Here, the voice conversation information is voice data generated while making an audio/video call with the other member, the message conversation information is text data generated while having a conversation with the other member through an instant messager, and the SNS activity information is the activity history in the SNS service. is data about

The controller 160 may predict the member's disposition and preference through machine learning by using the profile information and social activity information.

First, the control unit 160 may learn the propensity prediction model by using the propensity learning data input from the manager, and input the member's profile information and social activity information into the learned propensity prediction model to predict the member's propensity. .

3 is a diagram for explaining a tendency prediction model 162 according to an embodiment.

Referring to FIG. 3 , the control unit 160 may include a tendency prediction model learning unit 161 and a tendency prediction model 162 . The propensity prediction model learning unit 161 and the propensity prediction model 162 are configured according to the functions of the control unit 160 , and it is obvious that the control unit 160 can implement all the corresponding functions.

An artificial neural network may be used as the propensity prediction model 162 according to an embodiment. An artificial neural network is a predictive model implemented in software or hardware that mimics the computational power of a biological system using a large number of artificial neurons (or nodes).

The propensity prediction model 162 may be supervised by the propensity prediction model learning unit 161 using profile information, social activity information, and propensity information. In this case, supervised learning means learning to find an output value according to a given input value by using data having an input value and an output value according to it as learning data, and means learning performed in a state where the correct answer is known. The set of input and output values given to supervised learning is called training data. That is, the above-described profile information and social activity information are input values, and the propensity information is an output value, and may be used as training data for supervised learning of the propensity prediction model 162 .

For example, the propensity prediction model learning unit 161 converts the member's profile information or social activity information into a unique first one-hot vector to generate an input value, and obtains the member's propensity information. After generating an output value by transforming it into a unique second one-hot vector, the propensity prediction model 162 may be supervised by using the generated input and output values. Here, the first one-hot vector and the second one-hot vector may be vectors in which one of component values constituting the vector is '1' and the remaining component values are composed of '0'.

In one example, the propensity prediction model 162 receives an input value and multiplies the connection strength (or weight) for each of the output values of the input layer and the input layer having nodes corresponding to the number of components of the first one-hot vector, and , one or more hidden layers outputting by adding a bias; and an output layer for multiplying each output value of the hidden layer by a connection strength (or a weight), and outputting the result using an activation function. Here, the activation function may be a LeRU function or a Softmax function, but is not limited thereto. Connection strength and bias can be continuously updated by supervised learning.

Specifically, the propensity prediction model 162 may be supervised so that an output value of a loss function according to a given input value (a first one-hot vector) and an output value (a second one-hot vector) is minimized. . For example, the loss function H(Y, Y`) may be defined as in Equation 1 below.

In Equation 1, Ym may be the m-th component of the second one-hot vector, and Y'm may be the m-th component of the output vector output by receiving the first one-hot vector from the propensity prediction model 162 .

As the amount of training data increases, more supervised learning is performed on the propensity prediction model 162 to increase the accuracy of the propensity prediction model 162 .

An artificial neural network may be used for the propensity prediction model 162 . For example, the propensity prediction model 162 may be implemented as a Bidirectional LSTM (Bi-LSTM) or a Convolutional Neural Network (CNN).

And the marriage information providing server 100 can learn the preference prediction model using the preference learning data input from the manager, and input the member's profile information and social activity information to the learned preference prediction model, and the member's preference predictable.

4 is a diagram for explaining a preference prediction model 164 according to an embodiment.

Referring to FIG. 4 , the controller 160 may include a preference prediction model learner 163 and a preference prediction model 164 . The preference prediction model learning unit 163 and the preference prediction model 164 are configured according to the functions of the control unit 160 , and it is obvious that the control unit 160 can implement all the corresponding functions.

The preference prediction model learning unit 163 may learn the preference prediction model 164 by using the received preference learning data. Here, the preference learning data may be training data having profile information and social activity information of a member as input values and preference information corresponding thereto as output values.

For example, the preference prediction model learning unit 163 converts the member's profile information and social activity information into a unique first one-hot vector to generate an input value, and converts the corresponding preference information into a unique second one-hot vector. After generating an output value by converting it into a vector, the generated input value and output value may be supervised by the propensity prediction model 164 . Here, the first one-hot vector and the second one-hot vector may be vectors in which one of component values constituting the vector is '1' and the remaining component values are composed of '0'.

Like the tendency prediction model 162 , an artificial neural network may be used in the preference prediction model 164 according to an embodiment. For example, the preference prediction model 164 may be implemented as a Bidirectional LSTM (Bi-LSTM) or a Convolutional Neural Network (CNN).

The tendency prediction model 164 may be supervised by the preference prediction model learning unit 163 using profile information and preference information corresponding to social activity information.

In detail, the propensity prediction model 164, like the propensity prediction model 162, has an output value of a loss function according to a given input value (a first one-hot vector) and an output value (a second one-hot vector). It can be supervised so that it is minimized. Here, the loss function may be the same as Equation (1).

After the supervised learning is completed, the control unit 160 inputs the member's profile information and social activity information to the learned preference prediction model 164 as input values, and as the output value of the tendency prediction model 164, the member's preference can be printed out.

Here, the controller 160 may use the social activity information input to the propensity prediction model 162 and the preference prediction model 164, and use the conversation keywords extracted from the voice conversation information and the message conversation information and the intimacy extracted from the SNS activity information. have.

To this end, the controller 160 may extract conversation keywords from voice conversation information and message conversation information, and extract intimacy using the contact frequency and emotional state calculated from the SNS activity information.

Hereinafter, a method of extracting conversational keywords and intimacy will be described.

First, the controller 160 may convert voice conversation information composed of a voice file into text information. 5 is a conceptual diagram of a method of converting voice conversation information into text information according to an embodiment.

The control unit 160 may include a voice recognition module 165, and the voice recognition module 165, when receiving voice inputs of a plurality of speakers from voice conversation information, recognizes the input voice and transcribed text information into text. can create

Furthermore, if the voice recognition module 165 identifies a speaker during voice recognition, as shown in FIG. 5 , the voice transcribed into the text information may be displayed separately for each speaker. This is because, in the case of voice conversations, it can be important to distinguish the speaker.

The controller 160 may convert the message conversation information to fit the format of the text information extracted from the voice recognition module 165 . That is, the controller 160 may convert the voice conversation information and the message conversation information into text information of the same format.

The controller 160 may extract a conversation keyword between members by using text information converted from voice conversation information and message conversation information.

6 is a flowchart of a method for extracting conversational keywords according to an embodiment.

Referring to FIG. 6 , the dialog keyword extraction method according to an embodiment includes a morpheme analysis step (S100), a candidate keyword extraction step (S110), a filtering performing step (S120), a rank calculation step (S130), and a dialog keyword extraction step ( S140) may be included.

First, in the morpheme analysis step ( S100 ), the control unit 103 analyzes the morpheme of text information, which is a target for extracting dialogue keywords. Here, the text information may consist of a plurality of words. Hereinafter, text information, which is an extraction target of a conversational keyword, will be simply referred to as an 'extraction target document'.

Then, in the candidate keyword extraction step ( S110 ), the controller 103 extracts the word(s) corresponding to the predetermined morpheme pattern as keyword candidates. For example, word(s) corresponding to the morpheme pattern 'NNG XSN NNG' may be extracted as a keyword candidate.

Table 1 below is a part of the morpheme code. That is, in one embodiment, pattern(s) of morphemes generally used as keywords are pre-stored, and a keyword having the corresponding morpheme pattern is selected as a candidate.

morpheme code Explanation morpheme code Explanation NNG common noun IC interjection NNP proper noun JKS nominative investigation NNB dependent noun JKC complementary investigation NR Investigation JKG tubular case study NP pronoun JKO purposeful investigation VV verb JKB sub-firing investigation VA adjective JKV scrutiny VX auxiliary verb JKQ Quotation Investigation VCP affirmative adjective JX assistant VCN negative designator JC connection investigation MM detective EP fresh fish mother MAG common adverb EF terminating suffix MAJ conjunctive adverb SN number ...

In addition, in the filtering operation step S120 , the control unit 160 may perform filtering on the extracted keyword candidate in consideration of at least one of grammaticality and frequency of appearance. Filtering in consideration of grammar means removing word(s) that do not fit the grammar from among keyword candidates. The frequency of appearance may mean a frequency of appearance in the 'extraction target document'. Word(s) that do not appear frequently may be filtered to be removed from keyword candidates by determining that their importance is not high.

The aforementioned filtering may serve to roughly filter out keywords that are not suitable as keyword candidates in a simple way.

In particular, according to an embodiment, the filtering step may use a Language Model (LM). The language model means a probability distribution for at least two consecutive word sequences (word sequences). That is, it means a model that assigns probabilities to word sequences to model the phenomenon of language. The language model calculates the probability that a particular sequence of words will be used.

7 is a conceptual diagram of a language model according to an embodiment.

In applying the language model according to an embodiment, two types of language models may be combined and utilized. One of the two types is a 'grammar language model' that learns from a reliable source in order to focus on grammatical characteristics considering grammar, and the other is a 'grammar language model' for reflecting the content in the 'extraction target document'. informational language model.

Since the content of the 'extraction target document' itself may be a transcription of voice input into text, grammatical reliability may not be high. Therefore, in one embodiment, it is proposed to use a 'grammar language model' in order to analyze grammatical properties of a word sequence.

To this end, the controller 160 may extract ngrams (unigram, bigram, etc.) from the 'extraction target document' and learn the 'grammar language model' based on the extracted contents.

The 'information language model 604' is a model that reflects the contents in the 'extraction target document'. This is to calculate the probability of how often a specific word sequence is used within the corresponding 'extraction target document'. In other words, the analysis source of the 'information language model' will be the 'extraction target document'.

To this end, the controller 160 may extract ngrams from the 'extraction target document' and learn the 'information language model' from this. Simply put, the 'information language model' means a model trained based on the 'extraction target document'.

The controller 160 may perform filtering in consideration of both the 'grammar language model' and the 'information language model'. If both types of models are considered together, filtering that considers both grammatical and appearance frequency may be performed. In one embodiment, in order to use two types of models together, it is proposed to use 'pointwise KL (kullback) divergence'. However, utilizing 'pointwise KL (kullback) divergence' is only an example, and the present invention is not limited to this method.

In addition, in the rank calculation step S130 , the controller 160 may calculate a rank for each of the filtered keyword candidates. Here, the rank is determined based on the weight of the corresponding keyword on the 'extraction target document', and may mean the importance of the keyword on the 'extraction target document'. That is, the weight means an importance score for determining a relative ranking among keyword candidates. The controller 160 may obtain a comprehensive weight by combining the weights calculated by various approaches.

In addition, in the dialog keyword extraction step ( S140 ), the controller 160 may extract a conversation keyword by excluding a keyword candidate having a low importance based on the calculated rank. When the number of keywords to be finally selected is N, the controller 160 may leave a keyword of N*(N-1) and exclude the rest. This is because semantic duplication needs to be removed for the remaining keyword candidates, because there is a large amount of computation to consider semantic duplication. That is, in order to efficiently manage the amount of computation, keyword candidates with low importance are removed.

Subsequently, the controller 160 may remove the duplicate keywords in consideration of the semantic overlap between the remaining keyword candidates. That is, for duplicate keywords, only conversational keywords are left and removed.

Accordingly, the controller 160 may extract a conversation keyword from the text information.

In addition, the controller 160 may extract intimacy between members by using the social activity information. To this end, the controller 160 may collect information and posts of the member and at least one acquaintance from the SNS server 330 , and analyze the contact frequency and emotional state based on the collected posts to extract intimacy.

The control unit 160 may extract information unique to the user terminal 200 of the member as member information on the SNS server 330 and extract acquaintance information forming an acquaintance relationship with the member. Here, member information and acquaintance information may include at least one or more of members and acquaintances' photos, e-mail addresses, social network service IDs, gender, hobbies, age, occupation, work, education, phone number, date of birth, and address. not limited

The controller 160 may collect posts registered between members and acquaintances within a specific period. Specifically, the controller 160 may collect posts between the member and the acquaintance based on the collected ID information of the member and the acquaintance.

The controller 160 may count the number of posts registered between the member and the acquaintance within a specific period, and calculate the contact frequency between the member and the acquaintance according to the counted number. For example, the controller 160 may count the number of posts between a member and an acquaintance by date, accumulate the number of posts counted by date for a specific period, and calculate a numerical value of the cumulative contact frequency between the member and the acquaintance. In this case, the higher the cumulative contact frequency, the higher the intimacy between the member and the acquaintance may be set.

The controller 160 may extract morphemes by parsing the content of the post, and calculate the emotional state value between the member and the acquaintance based on the extracted morpheme according to a preset standard.

The controller 160 determines whether the parsed morpheme matches the emotional state database previously stored in the memory 140 . Here, the emotional state database may be a list in which words, emotional state numerical values, and emotional state items are matched with each other based on a predetermined criterion. For example, the emotional state database may match the word 'happiness', the emotional state value of '6.08', and 'hui (喜)' among the four emotional state items of joy, sorrow, and joy (喜努哀樂). That is, the controller 160 checks whether the parsed morpheme is a word in the emotional state database, searches for an emotional state value matching the word if it is a word in the emotional state database, and accumulates the searched numerical value in the emotional state item. have.

The controller 160 may extract the intimacy between the member and the acquaintance according to the calculated contact frequency and emotional state. The controller 160 may calculate the intimacy as the sum of the contact frequency and the emotional state, and may calculate the intimacy by giving weight to the contact frequency or emotional state.

The controller 160 may generate a network map between the member and the acquaintance by using the extracted intimacy. The higher the intimacy, the shorter the length of the line connecting the member and the acquaintance, and the lower the intimacy, the longer the length of the line connecting the member and the acquaintance. That is, the controller 160 may locate the acquaintance having a relatively high calculated intimacy at a location closer to the user than other acquaintances.

8 is a diagram illustrating social network information in which intimacy is reflected according to an embodiment.

As noted above, the predicted propensity of one embodiment is trend data representing members with respect to a criterion. Here, the work standard may be set in various standards. For example, it may be a personality-related standard such as introvert/extrovert, it may be an information collection-related standard such as sensory/intuitive, it may be a decision-making-related standard such as thinking/emotion, and stable It may be a criterion related to job preference that is a preference for a job.

The controller 160 may predict the membership tendency based on a plurality of criteria. For example, the control unit 160 may predict the first tendency as the tendency of the personality-related criterion and the second tendency as the tendency of the decision-making criterion, and additionally predict the N-th tendency such as the third tendency and the fourth tendency of the member. can

In addition, the predicted preference according to an embodiment is trend data representing the member's preference for the other member with respect to one criterion. Here, the preference may be predicted in a certain range from one criterion.

The control unit 160 may quantify the predicted member's disposition and preference. For example, the controller 160 may quantify and predict what kind of value a member's preference has within a minimum value of -100 and a maximum value of 100, and may quantify and predict what range value the member's preference has.

The control unit 160 may generate member matching coordinates by using the digitized member's disposition and preference. According to an exemplary embodiment, the member matching coordinates may be coordinates showing the member's disposition and preference based on at least one disposition, the member disposition may be shown as one point, and the member disposition may be shown as one range.

9 is a diagram illustrating member matching coordinates 500 according to an embodiment.

The member matching coordinates 500 shown in FIG. 9 are 2 generated based on the first tendency (personality criterion) and the second tendency (decision making criterion), the first tendency as the X axis, and the second tendency as the Y axis. dimensional coordinates.

The controller 160 may show the predicted member's disposition and preference in the generated member matching coordinates 500 .

As shown in the drawing, when the member matching coordinates 500 are two-dimensional coordinates, the member's tendency is a two-dimensional point, and the member's preference is a two-dimensional circle having a center and a radius. can be shown in In the drawings, preference according to an embodiment is shown as a circle, but this is only an example and may be illustrated in various forms depending on the predicted value of the preference.

The control unit 160 may generate the member matching coordinates 500 as three-dimensional coordinates according to three propensities rather than two-dimensional coordinates as shown in the drawing, and as N-dimensional coordinates according to N kinds of propensities, and the propensity to be a reference of the coordinates The type of membership may be decided by the administrator or member.

The controller 160 may use the member matching coordinates 500 to determine the counterpart member to be matched with the target member. Specifically, the control unit 160 may extract a member whose disposition is included in the preference of the target member, and determine the member as a counterpart member of the target member. Accordingly, the controller 160 may match the target member with the counterpart member.

10 is a diagram for explaining a member matching method according to an embodiment.

The controller 160 may show the propensity of all members in the member matching coordinates 500 . And in order to match the target member, the controller 160 may show the preference of the target member in the member matching coordinates 500 .

The controller 160 may determine a member whose disposition is included in the preference of the target member as the counterpart member, and may match the target member with the counterpart member. In the drawing, the preference of member 1, which is a target member, is shown in the member matching coordinates 500, and members 4 and 6, which are included in the preference range of the target member, are determined as counterpart members. The control unit 160 may match member 1 with member 4 and member 6.

The controller 160 may set a higher priority for the counterpart member having a similar propensity vector distance (d, a distance between propensity points) between the target member and the target member even if the propensity is equally included in the preference of the target member.

Referring to the drawing, the control unit 160 may set a higher priority than member 6 for member 4, which is close to member 1, which is a target member, and the propensity vector distance d.

The controller 160 may predict the member's propensity and preference for each unit time, and may update the member's propensity and preference predicted for each unit time in the member matching coordinates 500 . The marriage information providing server 100 may use the updated member matching coordinates 500 to determine the other member whose disposition is included in the target member's preference to match the member with the other member.

11 is a diagram illustrating tendencies and preferences updated for each unit time according to an embodiment.

The control unit 160 receives the social activity information of the member for each unit time, inputs the received social activity information into the propensity prediction model 162 and the preference prediction model 164, and calculates the propensity and preference updated for each unit time. can

The controller 160 may update the member's tendency and preference calculated for each unit time in the member matching coordinates 500 .

The controller 160 may set the unit time to various times (eg, 1 hour, 1 day, 1 week, etc.), and may set the unit time by receiving input from an administrator or a member.

In the drawing, it can be seen that the tendency of member 1 is changed to the +X axis and the -Y axis, and the center of preference is changed to the -X axis and the -Y axis, and the radius is changed to become smaller.

The controller 160 may determine a counterpart member according to the changed disposition and preference of the target member, and match the determined counterpart member to the target member.

The preference according to an embodiment may be composed of a first preference and a second preference. Here, the first preference is a range value of the preference tendency derived as an output value from the above-described preference prediction model 164 , and the second preference is a range value of the preference tendency generated from the first preference. The range of the second preference may include the range of the first preference.

The controller 160 may set the preference predicted by the preference prediction model 164 as the first preference, and generate the second preference using the first preference.

The controller 160 may generate the second preference from the first preference using various methods. For example, the control unit 160 may generate a second preference having the same center as the first preference and the same radius as the center of the first preference, and the center may be generated from the center of the first preference to the member's tendency direction. Move n%, and the radius may create a second preference that is k% (n<k) greater than the radius of the first preference.

12 is a conceptual diagram illustrating a second preference according to an embodiment.

Referring to FIG. 12 , it can be seen that member 1's second preference has the same center as the first preference and has a larger radius than the first preference. However, since the controller 160 may generate the second preference using the first preference in various ways, the present invention is not limited to the drawings, and various types of second preference including all the ranges of the first preference can create

In another embodiment, the controller 160 may perform member matching using the first preference and the second preference.

13 is a diagram for explaining a member matching method according to another embodiment.

Referring to FIG. 13 , member 1's disposition, first preference, and second preference are shown in member matching coordinates 500 . In addition, the member matching coordinates 500 show the propensities of other members (member A, member B, member C) that change for each unit time.

When the disposition changed in N unit time is included in the second preference of the target member, and the disposition change direction (propensity vector direction) changed from N-1 unit time to N unit time is directed toward the first preference , the corresponding member can be set as the other member.

In the drawing, member A and member B are included in the second preference of member 1 whose disposition after N unit time is the target member, and the direction of disposition movement from N-1 unit time to N unit time indicates the first preference of the target member. Therefore, member A and member B can be set as counterpart members of member 1.

In the drawing, member C is included in the second preference of member 1 whose disposition after N unit time is the target member, but the direction of movement from N-1 unit time to N unit time is not toward the first preference of member 1, Member C may not be set as the counterpart member of Member 1.

The controller 160 may control the communication unit 110 to transmit the matched profile of the counterpart member to the user terminal 200 of the target member.

14 is a diagram illustrating a configuration of a user terminal 200 according to an embodiment. Hereinafter, components constituting the user terminal 200 shown in FIG. 14 will be described in turn.

The wireless communication unit 210 may include one or more components that perform wireless communication between the user terminal 200 and the wireless communication system or between the user terminal 200 and the network in which the user terminal 200 is located. . For example, the wireless communication unit 210 may include a broadcast reception module 211 , a mobile communication module 212 , a wireless Internet module 213 , a short-range communication module 214 , and a location information module 215 . .

The broadcast reception module 211 receives a broadcast signal and/or broadcast related information from an external broadcast management server through a broadcast channel. Here, the broadcast channel may include a satellite channel and a terrestrial channel. Meanwhile, broadcast-related information may be provided through a mobile communication network, and in this case, may be received by the mobile communication module 312 .

In addition, the mobile communication module 212 transmits/receives a radio signal to and from at least one of a base station, an external terminal, and a server on a mobile communication network. Here, the wireless signal may include various types of data according to transmission/reception of a voice call signal, a video call signal, or a text/multimedia message.

The wireless Internet module 213 refers to a module for wireless Internet access, and may be built-in or external to the user terminal 200 .

The short-range communication module 214 refers to a module for short-range communication. As the short-distance communication technology, Bluetooth, Radio Frequency Identification (RFID), Infrared Data Association (IrDA), Ultra Wideband (UWB), ZigBee, etc. may be used.

Also, the location information module 215 is a module for confirming or obtaining the location of the user terminal 200 . An example is a Global Position System (GPS) module. The GPS module receives location information from a plurality of satellites. Here, the location information may include coordinate information indicated by latitude and longitude.

Meanwhile, the A/V (Audio/Video) input unit 220 is for inputting an audio signal or a video signal, and may include a camera 221 and a microphone 222 . The camera 221 processes an image frame such as a still image or a moving image obtained by an image sensor in a video call mode or a shooting mode. Then, the processed image frame may be displayed on the display unit 251 .

The image frame processed by the camera 221 may be stored in the memory 360 or transmitted to the outside through the wireless communication unit 210 . Two or more cameras 221 may be provided according to the configuration of the user terminal 200 .

The microphone 222 receives an external sound signal by a microphone in a call mode, a recording mode, a voice recognition mode, and the like, and processes it as electrical voice data. In addition, the processed voice data may be converted into a form that can be transmitted to the mobile communication base station through the mobile communication module 212 in the call mode and output. The microphone 222 may implement various noise removal algorithms for removing noise generated in the process of receiving an external sound signal.

The user input unit 230 receives an input operation from the user and generates input data for controlling the operation of the user terminal 200 .

The sensing unit 240 detects the current state of the user terminal 200 , such as the location of the user terminal 200 , the presence of user contact, the orientation of the user terminal 200 , acceleration/deceleration of the user terminal 200 , and the user A sensing signal for controlling the operation of the terminal 200 is generated.

The interface unit 270 serves as an interface with all external devices connected to the user terminal 200 . For example, wired/wireless headset ports, external charger ports, wired/wireless data ports, memory card ports, ports for connecting devices equipped with identification modules, audio input/output (I/O) ports, It may include a video input/output (I/O) port, an earphone port, and the like.

The output unit 250 is for outputting an audio signal, a video signal, or an alarm signal, and may include a display unit 251 , a sound output module 252 , an alarm unit 253 , and the like.

The display unit 251 displays and outputs information processed by the user terminal 200 . For example, when the terminal is in a call mode, a user interface (UI) or graphic user interface (GUI) related to a call is displayed. And, when the user terminal 200 is in the video call mode or the shooting mode, the captured and/or received image or UI and GUI are displayed.

Meanwhile, as described above, when the display unit 251 and the touch pad form a layer structure to form a touch screen, the display unit 251 may be used as an input device in addition to an output device. The display unit 251 includes a liquid crystal display, a thin film transistor-liquid crystal display, an organic light-emitting diode, a flexible display, a three-dimensional display ( 3D display). In addition, two or more display units 251 may exist according to an implementation form of the user terminal 200 . For example, the user terminal 200 may be provided with an external display unit (not shown) and an internal display unit (not shown) at the same time.

The sound output module 252 outputs audio data received from the wireless communication unit 210 or stored in the memory 260 in a call signal reception, a call mode or a recording mode, a voice recognition mode, a broadcast reception mode, and the like. Also, the sound output module 352 outputs a sound signal related to a function (eg, a call signal reception sound, a message reception sound, etc.) performed in the user terminal 200 . The sound output module 252 may include a speaker, a buzzer, and the like.

The alarm unit 253 outputs a signal for notifying the occurrence of an event in the user terminal 200 . Examples of events occurring in the terminal include call signal reception, message reception, and key signal input.

The memory 260 may store a program for processing and control of the controller 280, and provides a function for temporary storage of input/output data (eg, phonebook, message, still image, video, etc.). can also be done

The memory 260 may include a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (eg, SD or XD memory), and a RAM. (RAM, Random Access Memory) SRAM (Static Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory) magnetic memory, magnetic disk, It may include at least one type of storage medium among optical disks.

The controller 280 generally controls the overall operation of the terminal. For example, it performs related control and processing for voice calls, data communications, video calls, and the like. Also, the controller 280 may include a multimedia module 281 for playing multimedia. The multimedia module 281 may be implemented within the control unit 280 or may be implemented separately from the control unit 280 .

The controller 280 controls various operations of the terminal for implementing the above-described method of providing marriage information.

The power supply unit 290 receives external power and internal power under the control of the control unit 280 to supply power required for operation of each component.

On the other hand, at least some or all of the operations of the above-described marriage information providing server 100 may be implemented in the user terminal 200 . In this case, the user terminal 200 may be pre-installed with an application for communicating with the marriage information providing server 100 to perform the operation of the above-described marriage information providing server 100 .

15 is a diagram illustrating a wireless communication system that can be applied in a communication process according to an embodiment of the present invention. FIG. 16 is a diagram illustrating a base station in the wireless communication system according to FIG. 15 . 17 is a diagram illustrating a terminal in the wireless communication system according to FIG. 15 . 18 is a diagram illustrating a communication interface in the wireless communication system according to FIG. 15 .

Hereinafter, an example of a wireless communication network system supporting communication between various servers and terminals will be described in detail. In the following description, a first node (device) may be an anchor/donor node or a centralized unit (CU) of an anchor/donor node, and a second node (device) may be an anchor/donor node or a distributed unit (DU) of a relay node. can be

As a part of a node using a radio channel in a wireless communication system, a base station (BS), a terminal, a server, and the like may be included.

A base station is a network infrastructure that provides terminals and wireless access to terminals. A base station has coverage defined as a certain geographic area according to the distance over which signals can be transmitted.

A base station is an "access point (AP)", "enodeb (eNB)", "5th generation (5G) node", "wireless point", " It may be referred to as a "transmission/reception point (TRP)".

The base station, the terminal, and the terminal may transmit and receive radio signals in millimeter wave (mmWave) bands (eg, 28 GHz, 30 GHz, 38 GHz, 60 GHz). In this case, the base station, the terminal, and the terminal may perform beamforming to improve the channel gain. Beamforming may include transmit beamforming and receive beamforming. That is, the base station, the terminal, and the terminal may impart directivity to the transmitted signal and the received signal. To this end, the base station, the terminal, and the terminal may select a serving beam through a beam search procedure or a beam management procedure. Thereafter, communication may be performed using a resource that is in a quasi co-located relationship with a resource carrying the serving beam.

The first antenna port and the second antenna port are considered quasi-co-located if the large-scale properties of the channel through which the symbol of the first antenna port is carried can be inferred from the channel through which the symbol of the second antenna port is carried. The large-scale attribute may include one or more of delay spread, Doppler spread, Doppler shift, average gain, average delay, and spatial Rx parameters.

Hereinafter, a base station is exemplified in the above-described wireless communication system. The terms "-module", "-unit" or "-er" used hereinafter may mean a unit that processes at least one function or operation, and includes hardware, software, or hardware and software. can be implemented as a combination of

The base station may include a wireless communication interface, a backhaul communication interface, a storage unit and a controller.

The wireless communication interface performs a function of transmitting and receiving signals through a wireless channel. For example, the wireless communication interface may perform a conversion function between a baseband signal and a bit stream according to a physical layer standard of the system. For example, in data transmission, a wireless communication interface encodes and modulates a stream of transmitted bits to produce complex symbols. In addition, upon data reception, the wireless communication interface demodulates and decodes the baseband signal to reconstruct the received bit stream.

The wireless communication interface performs a function of transmitting and receiving signals through a wireless channel. For example, the wireless communication interface may perform a conversion function between a baseband signal and a bit stream according to a physical layer standard of the system. For example, in data transmission, a wireless communication interface encodes and modulates a stream of transmitted bits to produce complex symbols. In addition, upon data reception, the wireless communication interface demodulates and decodes the baseband signal to reconstruct the received bit stream.

In addition, the wireless communication interface up-converts the base band signal into a radio frequency (RF) band signal, transmits the converted signal through the antenna, and then down-converts the RF band signal received through the antenna into a base band signal. To this end, the wireless communication interface includes a transmission filter, a reception filter, an amplifier, a mixer, an oscillator, a digital-to-analog converter (DAC), It may include an analog-to-digital converter (ADC) and the like. Also, the wireless communication interface may include a plurality of transmission/reception paths. Further, the wireless communication interface may include at least one antenna array including a plurality of antenna elements.

In terms of hardware, the wireless communication interface may include a digital unit and an analog unit, and the analog unit may include a plurality of sub-units according to operating power, operating frequency, and the like. The digital unit may be implemented with at least one processor (eg, a digital signal processor (DSP)).

The wireless communication interface transmits and receives signals as described above. Accordingly, a wireless communication interface may be referred to as a "transmitter", "receiver" or "transceiver". In addition, in the following description, transmission/reception performed through a wireless channel may be used to include processing performed in a wireless communication interface as described above.

The backhaul communication interface provides an interface for performing communication with other nodes in the network. That is, the backhaul communication interface converts the bit stream transmitted to another node, for example, another access node, another base station, a higher node or a core network from a base station converts the physical signal received from the other node into a physical signal. convert to stream

The storage unit stores data such as basic programs, applications, and setting information for operation of the base station. The storage unit may include a volatile memory, a non-volatile memory, or a combination of a volatile memory and a non-volatile memory.

The controller controls the overall operation of the base station. For example, the controller transmits and receives signals through a wireless communication interface or a backhaul communication interface. The controller also writes data to the storage and reads the recorded data. The controller can perform the function of the protocol stack required by the communication standard. According to another implementation, the protocol stack may be included in a wireless communication interface. To this end, the controller may include at least one processor.

According to an embodiment, the controller may control the base station to perform an operation according to an embodiment of the present invention.

According to various embodiments, a donor node of a wireless communication system includes at least one processor, a transceiver operatively coupled to the at least one processor, and a plurality of radio bearers for a terminal accessing the relay node. and send to the relay node a first message including first information related to the donor node regarding receive, from the relay node, a second message including second information related to the relay node regarding a plurality of radio bearers for the terminal; Data for the terminal may be transmitted to the relay node. Data may be transmitted to the terminal through a plurality of radio bearers based on the first information and the second information.

According to various embodiments, a radio bearer among a plurality of radio bearers may aggregate a plurality of radio bearers. the at least one processor is further configured to determine a radio bearer for the terminal accessing the relay node and a multiple radio bearer aggregated by the radio bearer; Alternatively, a radio bearer for a terminal accessing the relay node may be determined.

According to various embodiments, the first message may include one or more of the following: identification of a terminal accessing the relay node; display information indicating the type of terminal accessing the relay node; information on a radio bearer of a terminal accessing a relay node; information on the radio bearer delivered by the terminal accessing the relay node; information on the tunnel established for the radio bearer between the donor node and the relay node; information on the aggregated multiple radio bearers; radio bearer mapping information; information about the address on the side of the donor node; information on the address of the relay node side; indication information corresponding to a radio bearer of a terminal accessing a relay node; indication information indicating the relay node to allocate a new address to a radio bearer for a terminal accessing the relay node; a list of address information that cannot be used by a relay node that transmits data of a radio bearer of a terminal accessing the relay node; and information related to security configuration.

According to various embodiments, the second message may include one or more of the following: identification of a terminal accessing the relay node; information on radio bearers accepted by the relay node; information on radio bearers not acknowledged by the relay node; information on radio bearers partially granted by the relay node; radio bearer mapping information; configuration information of a terminal accessing the relay node generated by the relay node; information on the address of the relay node side; and information related to security configuration.

According to various embodiments, the second message may further include information on the integrated multi-radio bearer.

According to various embodiments, the donor node may include a central unit of the donor node, and the relay node may include a distributed unit of the donor node.

According to various embodiments, a relay node of a wireless communication system includes at least one processor, a transceiver operatively coupled to the at least one processor, and a plurality of terminals accessing the relay node from a donor node. configured to receive a first message including first information related to a donor node regarding a radio bearer of ; transmit a second message including second information related to the relay node regarding the plurality of radio bearers to the terminal to the donor node; It is possible to receive data for the terminal from the donor node. Data may be transmitted to the terminal through a plurality of radio bearers based on the first information and the second information.

According to various embodiments, a radio bearer among a plurality of radio bearers may aggregate a plurality of radio bearers. the at least one processor is further configured to determine a radio bearer for the terminal accessing the relay node and a multiple radio bearer aggregated by the radio bearer; Alternatively, multiple radio bearers aggregated by the radio bearer may be determined.

According to various embodiments, the first message may include one or more of the following: identification of a terminal accessing the relay node; display information indicating the type of terminal accessing the relay node; information on a radio bearer of a terminal accessing a relay node; information on the radio bearer delivered by the terminal accessing the relay node; information on the tunnel established for the radio bearer between the donor node and the relay node; information on the aggregated multiple radio bearers; radio bearer mapping information; information about the address on the side of the donor node; information on the address of the relay node side; indication information corresponding to a radio bearer of a terminal accessing a relay node; indication information indicating the relay node to allocate a new address to a radio bearer for a terminal accessing the relay node; a list of address information that cannot be used by a relay node that transmits data of a radio bearer of a terminal accessing the relay node; and information related to security configuration.

According to various embodiments, the second message may include one or more of the following: identification of a terminal accessing the relay node; information on radio bearers accepted by the relay node; information on radio bearers not acknowledged by the relay node; information on radio bearers partially granted by the relay node; radio bearer mapping information; configuration information of a terminal accessing the relay node generated by the relay node; information on the address of the relay node side; and information related to security configuration.

According to various embodiments, the second message may further include information on the integrated multi-radio bearer.

According to various embodiments, the donor node may include a central unit of the donor node, and the relay node may include a distributed unit of the donor node.

Hereinafter, components of a terminal in the above-described wireless communication system are illustrated. The components of the terminal described below may be merged or integrated with the components of the terminal according to the above-described contents as components of a general-purpose terminal supported by the wireless communication system. Contents described with reference may be interpreted as having priority. The terms “-module”, “-unit” or “-er” used below may refer to a unit that processes at least one function.

The terminal includes a communication interface, a storage unit and a controller.

The communication interface performs a function of transmitting and receiving signals through a wireless channel. For example, the communication interface performs a conversion function between a baseband signal and a bit stream according to the physical layer standard of the system. For example, in data transmission, a communication interface encodes and modulates a stream of transport bits to produce complex symbols. In addition, upon data reception, the communication interface demodulates and decodes the baseband signal to reconstruct the received bit stream. In addition, the communication interface up-converts the base band signal into an RF band signal, transmits the converted signal through the antenna, and then down-converts the RF band signal received through the antenna into a base band signal. For example, the communication interface may include a transmission filter, a reception filter, an amplifier, a mixer, an oscillator, and a digital-to-analog converter (DAC). , an analog-to-digital converter (ADC), and the like.

Also, the communication interface may include a plurality of transmission/reception paths. Further, the communication interface may include at least one antenna array including a plurality of antenna elements. On the hardware side, the wireless communication interface may include a digital circuit and an analog circuit (eg, a radio frequency integrated circuit, RFIC). The digital circuit may be implemented with at least one processor (eg, DSP). The communication interface may include a plurality of RF chains. The communication interface may perform beamforming.

The communication interface transmits and receives signals as described above. Accordingly, a communication interface may be referred to as a "transmitter", "receiver" or "transceiver". In addition, in the following description, transmission/reception performed through a wireless channel may be used to include processing performed in a communication interface as described above.

The storage unit stores data such as basic programs, applications, and setting information for the operation of the terminal. The storage unit may include a volatile memory, a non-volatile memory, or a combination of a volatile memory and a non-volatile memory. In addition, the storage unit provides the stored data according to the request of the controller.

The controller controls the overall operation of the terminal. For example, the controller sends and receives signals through a communication interface. The controller also writes data to the storage and reads the recorded data. The controller can perform the function of the protocol stack required by the communication standard. According to another implementation, the protocol stack may be included in the communication interface. To this end, the controller may include at least one processor or microprocessor or may reproduce a part of the processor. Also, a part of the communication interface or controller may be referred to as a communication processor (CP).

According to an embodiment of the present invention, the controller may control the terminal to perform the operation according to the embodiment of the present invention.

Hereinafter, a communication interface in a wireless communication system is exemplified.

The communication interface includes encoding and modulation circuitry, digital beamforming circuitry, a plurality of transmission paths, and analog beamforming circuitry.

The encoding and modulation circuitry performs channel encoding. At least one of a low-density parity check (LDPC) code, a convolutional code, and a polar code may be used for channel encoding. The encoding and modulation circuitry generates modulation symbols by performing constellation mapping.

The digital beamforming circuit performs beamforming on a digital signal (eg, a modulation symbol). To this end, the digital beamforming circuit multiplexes the modulation symbols by the beamforming weight value. The beamforming weight may be used to change the size and phrase of a signal, and may be referred to as a “precoding matrix” or a “precoder”. The digital beamforming circuit outputs digital beamformed modulation symbols to a plurality of transmission paths. In this case, modulation symbols may be multiplexed or the same modulation symbol may be provided to a plurality of transmission paths according to a multiple-antenna technology (multiple input multiple output, MIMO) transmission method.

The plurality of transmission paths converts the digital beamformed digital signal into an analog signal. To this end, each of the plurality of transmission paths may include an inverse fast Fourier transform (IFFT) calculation unit, a cyclic prefix (CP) insertion unit, a DAC, and an up-conversion unit. The CP insertion unit is for an orthogonal frequency division multiplexing (OFDM) method and may be omitted when other physical layer methods (eg, a filter bank multi-carrier: FBMC) are applied. That is, the plurality of transmission paths provide independent signal processing processes for a plurality of streams generated through digital beamforming. However, depending on the implementation, some elements of a plurality of transmission paths may be used in common.

The analog beamforming circuit performs beamforming on an analog signal. To this end, the digital beamforming circuit multiplexes the analog signal by the beamforming weight value. Beamformed weights are used to change the size and text of the signal. More specifically, the analog beamforming circuit may be configured in various ways according to a connection structure between the plurality of transmission paths and the antenna. For example, each of the plurality of transmission paths may be connected to one antenna array. In another example, a plurality of transmission paths may be coupled to one antenna array. In another example, the plurality of transmission paths may be adaptively coupled to one antenna array or may be coupled to two or more antenna arrays.

The steps of the method or algorithm described in relation to the embodiment of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or implemented by a combination thereof. A software module may include random access memory (RAM), read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, hard disk, removable disk, CD-ROM, or It may reside in any type of computer-readable recording medium well known in the art to which the present invention pertains.

As mentioned above, although embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art to which the present invention pertains can realize that the present invention can be embodied in other specific forms without changing its technical spirit or essential features. you will be able to understand Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive.

Claims (1)

A marriage information matching method using a Bi-LSTM-based preference prediction model performed in a marriage information providing server, comprising:
Receiving profile information and social activity information of the first member from a user terminal or a social activity server;
inputting the profile information and social activity information into a propensity prediction model to predict a propensity value of the first member with respect to at least one criterion;
inputting the profile information and social activity information into a preference prediction model to predict a first preference that is a preference preference range of the first member with respect to the at least one criterion;
generating member matching coordinates for the at least one criterion;
determining a counterpart member to be matched with the first member by using the member matching coordinates; and
Comprising the step of transmitting the profile of the other member to the user terminal,
The first preference indicates a range of the preference of the first member with respect to the other member based on the at least one criterion,
The member matching coordinates are composed of a coordinate system having a dimension corresponding to the at least one predefined criterion, and in the member matching coordinates, propensity coordinates corresponding to each of a plurality of members, and a center corresponding to the range of the propensity The first preference expressed as a region having a radius of
The propensity coordinates corresponding to each of the plurality of members and the first preference are periodically updated for each unit time,
The step of determining the counterpart member,
determining, among the plurality of members, a member whose propensity coordinates are located within the region of the first preference as the counterpart member;
Among the plurality of members, a first disposition coordinate according to a predetermined N unit time is included in the area of the second preference, and in the second disposition coordinate according to an N-1 unit time corresponding to the past than the N unit time. determining, as the counterpart member, a member whose direction of movement in the first orientation coordinates faces the region of the first preference;
The closer the propensity vector distance between the propensity coordinates of the counterpart member and the propensity coordinates of the first member, the higher priority is set for the opposing member,
The preference prediction model is
It is an artificial neural network based on Bi-LSTM (Bidirectional LSTM), in which each member's profile information and social activity information are taken as input values and supervised learning is performed in advance using learning data in which preference information corresponding to the input values is used as an output value. ,
Marriage information matching method using Bi-LSTM-based preference prediction model.

Images