KR101031901B1 - Social network analyzing method and system based on communication record - Google Patents

Abstract

The present invention relates to a method and system for analyzing social relationships based on a communication record. The method for analyzing social relationships according to the present invention receives a communication record between nodes and profile information of nodes, and a relationship network between nodes from the input communication record. To generate a relationship strength between nodes in the generated relationship network, extract keywords and attributes from the input communication records and profile information, and manage the relationships between nodes ontologies, and calculate the relationship strength and ontology. By creating social relationships between nodes based on inference rules based on them, we can effectively obtain the relationships between users from the vast communication records generated between users, and utilize ontology and reasoning rules even if they are not provided with information about the relationship between sender and receiver. Analyze and expand relationships Can be.

Description

Social network analyzing method and system based on communication record

The present invention relates to a method and system for analyzing social relations based on a communication record. More particularly, the present invention relates to a method of forming a relationship between users based on a communication record between users, and to analyzing social relationships between users using ontology and inference rules. And to the system.

Traditionally, letters and telephones have been a way of interacting with each other without using face-to-face communication. In recent years, with the development of communication networks and the development of computer technology, there have been great changes in the means and methods of communication between individuals. Mobile phones beyond the limitations of conventional landline phones, E-mail that can be delivered immediately through a computer network without stamps, SMS (short message service) and MMS (multimedia message) that can deliver simple messages through various terminals service).

With the advent of such communication means, exchanges between individuals can be made more quickly and easily in terms of quality, and in terms of quantity, they are incomparable with conventional letters or landlines. In particular, this new communication means, unlike the conventional communication means, can leave a user's communication record in electronic form. For example, when sending a traditional letter to the other party, the contents of the letter and the date and time of transmission and reception will remain only in an analog form, but in the case of an e-mail, it is recorded in an electronic form that can be reproduced and verified at a computer at any time. Furthermore, the aforementioned communication means such as E-mail, SMS, MMS, etc. have no technical problem in keeping records of electronic forms in computer systems such as relay servers. Such records may include sender information, recipient information, transmission and reception times, and may even store the full text of the communication. In addition, due to the development of computer systems, in the case of a voice call using a mobile phone, not only the call log but also the voice call history can be easily stored.

Due to the digitalized communication means in accordance with the development of the technology, efforts are being made to utilize the communication records left by users in corporations and institutions. For example, the communication records between users can be analyzed and reflected in a company's marketing strategy, or used as a basic data for communication billing. The problem is that the amount of this communication record is so large that it is not easy to analyze, and unnecessary information contained in the communication record (which is a kind of noise) hinders this analysis. In particular, the easier the electronic recording of the communication record, the greater the amount of digital data, which causes a problem that it becomes more difficult to extract the necessary information.

Therefore, it is necessary to extract only the information useful to a specific organization including a company and to process it into a form that can be used by appropriately removing unnecessary noise from such a huge communication record. The most basic information that can be extracted from the communication record between users is the relationship between users, and the relationship between users can be expressed by using a specific keyword or topic through the communication record. If we analyze a person's social relations from this, the result can be used as very valuable information.

The technical problem to be solved by the present invention is to solve the problem that the relationship between the user could not be effectively obtained from the vast communication record between the user, and when the information about the relationship between the sender and the receiver is not provided, it is possible to grasp the relationship between the two. By overcoming the limitations that were not there, the social relationship between the sender and the receiver can be analyzed and expanded from the communication records.

In order to solve the technical problem, the social relationship analysis method according to the present invention comprises the steps of receiving a communication record between the nodes belonging to a predetermined domain and the profile information of the nodes; Creating a relationship network between the nodes from the input communication record; Calculating a relationship strength between nodes in the created relationship network; Extracting keywords and attributes from the input communication records and profile information, respectively, and managing the relationships among the nodes by ontology; And generating a social relationship between the nodes by a predetermined inference rule based on the calculated relationship strength and the ontology.

Ontology of managing the relationship between the nodes may include extracting an attribute from the input profile information and extracting a keyword corresponding to the extracted attribute from the input communication record; And extracting a new object from the nodes associated with the keyword and defining the relationship between the nodes.

The step of calculating the relationship strength between the nodes may include extracting a communication time or a communication frequency between the nodes for each communication means from the input communication record; And calculating a strength factor from the extracted communication time communication frequency in consideration of the communication means.

The calculated strength factor is a first strength factor calculated in consideration of a charge ratio for each communication means among the input communication records, or a second strength factor calculated in consideration of communication record statistics of a predetermined group to which the nodes belong. And a combination of the first strength factor and the second strength factor.

The predetermined reasoning rule is a sufficient condition for social relations between the nodes, and is at least one rule derivable from the nodes in consideration of the input communication record, profile information and the ontology.

Further, the following provides a computer readable recording medium having recorded thereon a program for executing the data structuring method described above on a computer.

In order to solve the above technical problem, the social relationship analysis system according to the present invention includes a data input unit for receiving a communication record between the nodes belonging to a predetermined domain and the profile information of the nodes; A relationship network generator for generating a relationship network between the nodes from the input communication record; A relationship strength calculator for calculating a relationship strength between nodes in the generated relationship network; An ontology manager which extracts keywords and attributes from the inputted communication record and profile information, respectively, and manages relationships among the nodes by ontology; And a social relationship generator for generating a social relationship between the nodes by a predetermined reasoning rule based on the calculated relationship strength and the ontology.

According to the present invention, it is possible to effectively obtain the relationship between users from the vast communication records generated between users, and to analyze and expand the relationship between the two by utilizing ontologies and inference rules even if the information on the relationship between the sender and the receiver is not provided. The present invention can also discover new relationships between senders and receivers by extracting keywords and attributes from communication records and profile information.

Prior to describing the embodiments of the present invention, terms and concepts to be used below are briefly described.

Graphs can be used to efficiently express relationships between individuals. A graph is a means of expressing a relationship between objects in a specific group in the field of computer science, and is composed of nodes and edges. A node represents an object (meaning both users or individuals communicating in terms of communication), and an edge represents a relationship between nodes.

Semantic technology is an intelligent technology that allows computers to communicate by setting language and rules that the computer can understand, just as a person reads the screen and understands the meaning. Semantic technology aims to express the relationship-semiteme between objects belonging to the environment in the form of an ontology that can be processed by a machine, that is, a computer, and to process it by an automated machine. Here, ontology is a model that abstracts and shares what people think about things, and is a technology that has been formalized and explicitly defined types of concepts or usage constraints. In the field of computer science, ontology is a data model that represents a specific domain and is defined as a set of formal vocabularies describing the concepts and relationships between them. In particular, ontology is a tool that can implement semantic technology. It is used as a tool that can semantically connect knowledge concepts, and can process and process the concept of human things in the form of a database on a computer. give.

In this field of semantic technology, the expression form of triple is used as a means for expressing a relationship. Triple refers to expressing concepts in the form of subjects, predicates, and objects. Each subject, predicate, and object can be expressed as a Uniform Resource Identifier (URI) in XML. Currently, the standard language describing semantic web ontology is RDF, OWL, and TopicMaps, which are proposed by the W3C.

By utilizing the ontology concept described above, it is possible to express a relationship between a user and a user in a communication environment in a graph structure. In addition, a social relation network can be formed by adding specific rules for identifying social elements inherent to the domain.

Hereinafter, various embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a social relationship analysis system based on a communication record according to an embodiment of the present invention, wherein the data input unit 100, the relationship network generation unit 200, the relationship strength calculation unit 300, and the ontology management unit ( 400 and the social relationship generating unit 500.

The data input unit 100 receives a communication record 110 between nodes belonging to a predetermined domain and profile information 120 for nodes. The communication record 110 means that the history is left as a digitized record when a communication activity between individuals occurs through various communication means. This type of communication record 110 includes a record of communication means available between individuals, such as voice calls, E-mail, SMS and MMS, and instant messenger services. In addition, the communication record 110 may include a sender identifier, a receiver identifier, a transmission / reception time, and transmission / reception contents. Here, the identifier means a unique key given to an individual to distinguish the sender and the receiver in the corresponding communication means. For example, when E-mail and SMS are used as a means of communication between individuals, the corresponding identifiers may be an E-mail address and a mobile phone number, respectively. In other words, these identifiers are utilized as a concept corresponding to the individual user in the real world in the communication record that is digitally recorded. Such an identifier is treated as an entity and has a specific relationship with other objects.

The profile information 120 is a type of personal information corresponding to an individual user, and may include age, gender, address, and occupation according to the characteristics of the corresponding domain. Profile information can be used as a basis for identifying relationships between individuals.

The communication record and the profile information described above may be processed in an electronic form and provided to the data input unit 100. If necessary, a process of removing and summarizing unnecessary noise among all input data may be required. . Extracting individual identifiers to represent nodes in graph form from complex communication records, extracting only dates from time of transmission and reception, discarding time information, and summarizing relationships using keywords or topics based on communication data. The process may be such a summary process. The data input unit 100 may be a physical input device such as a keyboard and a mouse, or may be a storage means such as a memory or a disk drive capable of storing a data stream input through a network.

The relationship network generator 200 generates a relationship network 250 between nodes from the communication record 110 input through the data input unit 100. As described above, the communication record 110 includes a sender identifier and a receiver identifier. Thus, if there is a communication record between these identifiers, it can be seen that some relationship is established between them. Although it is not known at this stage whether the relationship is a family, friend, or business relationship, the fact that there is a communication record can establish the relationship between the communication targets. The relational network generator 200 may be implemented as a processor that receives a communication record 110 in electronic form and processes software code that can be read or stored by a conventional computer system. It is also possible to utilize memory if necessary to aid in the processing of software code.

On the other hand, in creating a relationship network, if the communication record 110 input through the data input unit 100 is a communication record such as a telephone call, SMS, E-mail, blog, bulletin board from the contents (contents) The process of extracting and summarizing meaningful information may be preceded. A keyword extraction method using voice recognition from call logs may be utilized, and a keyword extraction method from an electronic string may be utilized. A keyword extraction method using speech recognition and a method of extracting a keyword from a character string are based on various keyword extraction methods well known in the art, and thus detailed description thereof will be omitted. If a specific keyword summarized through the above process is extracted, it is possible to determine on which subject the relation network is formed. This topic-based relationship network can be expressed by adding the extracted theme as attribute information to the edges between the nodes described above.

The relationship strength calculator 300 calculates a relationship strength between nodes in the relationship network 250 generated through the relationship network generator 200. Two relationship networks 250 and 350 are shown in FIG. 1, which means input and output to the relationship strength calculator 300. That is, the relationship strength calculator 300 receives the relationship network 250 and analyzes the strength and weakness of the relationship between the nodes and numerically calculates the degree. In FIG. 1, a thicker line indicates an edge of a relationship network 350 output from the relationship strength calculator 300 with a relatively high relationship strength to aid visual understanding. The relationship strength may calculate a relationship between individuals based on objective numerical data such as communication time, period, frequency, etc. extracted from the communication record 110, and a detailed process will be described later in detail. The relationship strength calculator 300 may be implemented as a processor that receives a relationship network 250 in an electronic form and processes software code that can be read or stored by a conventional computer system. It is also possible to utilize memory if necessary to aid in the processing of software code.

When the above-described keyword summary method is used in the relationship strength calculator 300, it is possible to more specifically express the relationship between the nodes. For example, if a total of one hour of phone calls between two users, if there is no topic analysis process through keyword extraction, a relationship network corresponding to one hour call minutes between two people will be formed. However, if a total of 1 hour of phone calls were 40 minutes on the subject of travel and 20 minutes on the subject of a contract, their relationship would form a relationship network of 40 minutes of travel, In addition, a relationship network of 20 minute calls could be formed separately.

The ontology manager 400 extracts keywords and attributes from the communication record 110 and the profile information 120 input through the data inputter 100, respectively, and manages the relationships between the nodes by the ontology. As described above, ontology means that the relationship between nodes is expressed according to a specific protocol so that the computer system can understand. The ontology manager 400 extracts a keyword from the communication record 110 (more specifically, may be actual communication contents) to identify relationships between nodes and maps them to ontology objects. In addition, the ontology manager 400 may extract the profile items from the profile information 120 and use the same to determine the attributes of the ontology object. For example, if the user's age information is included in the profile information 120, in representing this as an ontology object, the age attribute of the object may be inferred.

The ontology management unit 400 may model the ontology objects configured as described above using the expression method as described above, and then expand the relationship by continuously accumulating data. On the other hand, in order to construct a more sophisticated ontology schema, it may be possible to define inferred relations or modify existing relations with the help of the domain experts from outside. The ontology management unit 400 may receive ontology definitions through input devices such as a keyboard and a mouse in a typical computer environment, and process software code that allows a computer system to read or store input or existing ontology definitions. It may be implemented as a processor.

The social relationship generation unit 500 generates a social relationship between nodes based on the inference rules based on the relationship strength and the ontology stored in the ontology manager 400 calculated through the relationship strength calculator 300. Inference rules refer to various rules necessary to extend ontology, which is a relationship between already defined objects. These inference rules are constraints that help achieve the inference results derivable from the domain to which the nodes (corresponding to individual users) belong, and are predefined and stored by the domain experts. The detailed process of deriving new relationships between nodes using ontology and inference rules will be described later. The social relation generation unit 500 may be implemented with a processor and a memory required for a task in a typical computer environment, and the computer system reads the social relation network 550 formed by an individual using an ontology and inference rules. Or as electronic data that can be stored. The created social relationship network 550 may be stored and utilized in a storage medium such as a disk or a memory, or provided to an external device through a separate interface.

In the case of using the keyword extraction process described above, in analyzing the social relationship between the nodes through the social relationship generation unit 500, it is possible to more specifically grasp the relationship based on the subject of each relationship. For example, when a total of 100 e-mail records between "Lee Dong Hoon" and "Jung Yong-Dong" are input through the data input unit 100, the e-mail contents are analyzed and 30 e-mail contents are received. Suppose that the topic of 'solution development' is extracted, the theme of 'workshop attending' is extracted from 20 email contents, and the theme of 'coffee club' is extracted from 50 email contents. Through this, three relationship networks will be formed between Lee Dong-hoon and Jung Yong-il under the theme of 'solution development', 'workshop attendance' and 'coffee club', and the relationship between 'coffee club' is the closest relationship between the two. Analysis results can be obtained. As such, through the summary of communication records and the keyword extraction process, the relationship between nodes can be analyzed more diversely, and various relationships can be found therefrom. In analyzing these relationships, it is natural that various relationships can be inferred by the above ontology.

According to this embodiment, it is possible to effectively obtain the relationship between users from the vast communication records generated between users, and to analyze and expand the relationship between the two by using the ontology and inference rules even if the information on the relationship between the sender and the receiver is not provided. Do.

FIG. 2 is a diagram illustrating a process of generating a relationship network and calculating a relationship strength in a social relationship analysis system based on a communication record according to an embodiment of the present invention. In FIG. 1, a data input unit 100 and a relationship network are generated. Only the unit 200 and the relationship strength calculator 300 are shown separately. In FIG. 2, it is assumed that an SMS communication record 110 between individuals is being input to illustrate a process of creating a relationship network and calculating a relationship strength.

As described above with reference to FIG. 1, the communication record 110 may include a sender identifier, a receiver identifier, a transmission / reception time, a message description, and the like. For convenience, FIG. 2 illustrates only the sender identifier and the receiver identifier. Here, P1 to P7 correspond to identifiers. For example, the expression "P1-P2" means that there is an SMS transmission and reception record between P1 and P2. When the communication record 110 is input through the data input unit 100, the relationship network generator 200 determines whether a communication record exists between the identifiers and generates the relationship network 250. 2 shows a relational network 250 in the form of a graph generated from the communication record 110. In the graph, a node is assigned an identifier corresponding to an individual user, and an edge means that there is a communication history between nodes.

Meanwhile, the communication record 110 input to the data input unit 100 includes time information related to a communication action. The relationship strength calculating unit 300 extracts a communication time or communication frequency between nodes for each communication means from the input communication record 110, and calculates an intensity factor from the extracted communication time communication frequency in consideration of the communication means. In this case, the intensity factor is a value introduced to quantify and express the relationship between nodes. This strength factor may be calculated by calculating the communication time or communication frequency extracted from the communication record 110.

In explaining the relationship between individuals, the frequency of communication will be meaningful data in the case of e-mail or SMS because it takes only a moment, but in the case of voice calls, It can be meaningful data. In other words, the more e-mails or SMSs are sent, the longer the talk time is, the deeper the relationship between users is. In FIG. 2, suppose that such communication time or frequency is high in the case of "P1-P2", "P1-P3", "P2-P3", and "P3-P6". As a result, in the relationship network 350 output through the relationship strength calculator 300, the corresponding edges are displayed in bold lines to visually indicate that the relationship between the related nodes is relatively deeper. This representation of FIG. 2 is presented for ease of understanding and may be embodied substantially as a high numerical strength coefficient imparted to the edge connecting the nodes. The process will be described in more detail below.

As an example of communication, mobile phone subscribers in Korea exceeded 40 million in 2006. This means that in processing cellular subscriber data, typical ages and genders should be divided by demographic rates. That is, a detailed and individual analysis is required by generating a plurality of classification rules such as dividing the user's age into sections such as youth, adolescent, young man, old age, and old age, and additionally dividing gender between male and female. As such, the user's mobile phone communication records can be extracted and summarized based on a plurality of classification rules.

3A to 3C are diagrams for describing various methods of calculating a relationship strength in a social relationship analysis system based on a communication record according to an embodiment of the present invention, and using classification criteria by communication means, sender, and recipient. The method of calculating the relationship strength is illustrated.

In FIG. 3A, the communication record 110 assumes a situation in which communication records are input through various forms such as voice call records, SMS, MMS, E-mail, blogs, bulletin boards, and messengers. The relationship strength calculator 300 receives the various types of communication records 110 for each communication means. Each communication record can be aggregated by sender and recipient. Here, the communication record aggregation by sender may be the total call volume, SMS sending amount, MMS sending amount, E-mail sending amount, etc. sent from one sender, and the communication record aggregation by recipient is based on one sender. It may be the total amount of calls received, the amount of SMS sent, the amount of MMS sent, the amount of E-mail sent, etc. for each corresponding recipient. That is, the communication record aggregation by sender can be used as statistics about the entire communication record of the sender, and the communication record aggregation by receiver can be used as data for determining which recipients have more interactions with the receiver. have.

On the other hand, the relationship strength calculation unit 300 calculates the strength coefficient between nodes from the extracted and summarized communication records. This is done taking into account the statistical information between the sender and the receiver. This calculation process can be configured freely and flexibly according to the characteristics of the domain to which the nodes belong. In collecting the communication records for each communication means illustrated in FIG. 3A, the weights of the respective communication means may be different. For example, sending one message through SMS and writing and sending one e-mail can be given different weights in measuring the intimacy between individuals. More specifically, in three communications, a higher weight may be given to three voice calls than to send two SMS and one E-mail, resulting in a higher strength factor value. There will be.

On the other hand, the flexible strength factor calculation method may be used in a situation where more specific classification criteria are applied. For example, a high SMS usage rate of 10 teenagers or a small call volume of 60 generations or more are characteristics representing each age section, and thus the strength factor may be calculated in consideration of this. Therefore, if necessary, the nodes are classified by age and gender to find factors representing each classification, and then the strength of the relationship in consideration of the weight of each factor may be derived.

In the following, two additional methods of calculating the strength factor are introduced.

First, as a method of calculating a subjective intensity factor considering the individual difference (conventionally referred to as the first intensity factor), the content thereof will be described with reference to FIG. 3B. The first strength factor is calculated in consideration of the charge ratio for each communication means among the input communication records. To this end, the first strength factor calculator 310 extracts the statistics and the charge ratio 315 for the communication record for each communication means that the caller likes to use based on one caller. Here, the charge rate 315 for each communication means means a kind of occupancy occupied by each communication means among all communication means. Based on this, the first strength factor calculator 310 calculates the first strength factor by varying the weight of the communication record according to the caller's communication pattern.

In FIG. 3B, it is assumed that the charge rate 315 of the caller by communication means for voice call, SMS, MMS, and e-mail is 50%, 30%, 10%, and 10%, respectively, based on the total 100%. Suppose there are "Lee Dong Hoon" and "Jung Yong Il". For example, suppose that the number of voice calls, SMS, MMS, and e-mails sent to Lee Dong-hoon and Jung Yong-il is shown in Table 1 below.

Voice call SMS MMS E-mail Recipient "Dong Hoon" 60 cases 30 cases 0 Matches 10 cases Recipient "Date Date" 40 cases 30 cases 20 Matches 10 cases

Note the voice call and MMS in Table 1. "Lee Dong Hoon" had 10 more voice calls than average and 10 less MMS. On the other hand, "Dong Yong-il" had 10 fewer voice calls and 10 more MMSs than average. In this case, when considering the caller's communication pattern, it is determined that the caller is more intimate with " Go Hun. &Quot; This is because the proportion of voice calls, which occupy a greater proportion of the caller's communication pattern, is higher in "Dong Hoon" than "Jung Yong Il". As described above, the first strength factor may be calculated, and this value corresponds to a subjective value considering the caller's communication pattern or characteristic.

The second method is a method of calculating an objective strength factor (conventionally referred to as a second strength factor) in consideration of a communication pattern of a general user, which will be described with reference to FIG. 3C. The second strength factor is calculated taking into account the communication record statistics 325 of the group to which the nodes belong. To this end, the second strength factor calculation unit 320 records the average communication pattern or statistics 325 of the group to which the sender belongs (based on the specific classification described above) based on one caller and the sender. The second strength factor is calculated by measuring the communication strength between the sender and the receiver. The communication record statistics 325 of such a group may include a communication record average or standard deviation extracted from the input communication record.

In FIG. 3C, assume that the rates of voice call, SMS, MMS and E-mail sending of teenage girls are 30%, 50%, 10%, and 10%, respectively, based on the total 100%. For example, if the caller is a teenage girl and she sends 40, 40, 10, and 10 voice calls, SMS, MMS, and e-mails to a specific recipient, the girl's voice call usage score is the same age. Will be calculated relatively high, and the SMS usage score will be calculated relatively low. From these values, a second intensity factor can be calculated, which corresponds to an objective value obtained through comparison with callers having the same condition.

The method of calculating the first strength factor and the method of calculating the second intensity factor described above are merely examples of various intensity factor calculation methods, and it is natural that a third intensity factor or other strength factor obtained by arithmetically combining both may be used. . In each process, factor analysis can be used or a decision tree can be used to calculate a numerical strength factor.

4A and 4B are diagrams for explaining a process of inferring a relationship between nodes using an ontology and an inference rule in a social relationship analysis system based on a communication record according to an embodiment of the present invention.

As briefly described with reference to FIG. 1, the ontology manager 400 extracts an attribute from the input profile information 120, extracts a keyword corresponding to the attribute extracted from the input communication record 110, and associates the keyword with the keyword. Extract new objects from nodes and define them as relationships between nodes. If the profile information 120 includes the name, sex, address, age, and family relationship, it may be possible to derive the relationship between the nodes from the plurality of profile information. In this case, an attribute called family relationship may be extracted from the profile information. On the other hand, if the input communication record 110 is an e-mail, if a keyword is found, such as the title "superman" in the mail content, it will be possible to estimate the relationship between the relationship between the boss and the subordinates. In order to estimate such a relationship, it is possible to extend the relationship between objects by calculating the subject relevance between the object and the keyword in consideration of the above-described strength factor and extracting the new relationship.

In order to support this reasoning, it is only natural that various reasoning rules and validity can be used in the domain. Here, the inference rule is a sufficient condition for social relations between nodes, and means at least one or more rules derivable from nodes in consideration of input communication records, profile information, and ontology. This reasoning rule can be given in the form of a conditional statement. In other words, if the condition is satisfied, the conclusion of inference can be reached, but vice versa.

Table 2 below illustrates inference rules for inferring that the sender and the receiver are in a parental relationship.

Condition 1 Is the recipient 20 years older than the sender? Condition 2 Is the gender of the recipient male? Condition 3 Does the sender have a strength factor of 30 or more? Condition 4 Do the receiver and sender have the same family name? Condition 5 Is the recipient married?

As shown in Table 2 above, if all five conditions are satisfied, the probability that the sender and the receiver are rich is very high. This rule of inference might give a clear conclusion, such as "If all five conditions are true, then we have a rich relationship," but we assign different weights to each condition, and arithmetically, "The probability of being rich is 78%." Inference can also be derived from the probability.

On the other hand, these conditions can be checked through the attributes extracted from the input profile information, keywords extracted from the communication record, or can also be checked using the ontology already established.

In FIG. 4A, the relationship between P1 and P2 is a parental relationship 510, and the relationship between P1 and P2 is also a parental relationship 520. In addition, the edge between P1-P2 and the edge between P1-P3 are indicated by a thick line to show that they have a very strong relationship with each other. Suppose that under these circumstances, the intensity factor between P2-P3 is also very high. However, there is no known relationship between the two. In this case, a new relationship 530 may be given to an edge between P2 and P3 by utilizing the already established ontology 341 and the inference rule 342.

For example, assume that there are various relationships that define family relationships in the ontology and that the inference rules in Table 2 described above can be applied. In FIG. 4A, as a result of examining the conditions of Table 2 as P1 and P2, a probability of 95% is calculated. Further, suppose that P1 has a 89% probability that P3 has a rich relationship as a result of examining the conditions in Table 2. As a result, P2 and P3 may be siblings or grandchildren. If the result of the calculation is inferred through the ontology, the relationship 530 between P2 and P3 may be revealed. If there is an additional reasoning rule about the age difference between P2 and P3 and the order of age between P1, P2, and P3, the relationship can be deduced more clearly. 4B illustrates that a new relationship is given by determining that the relationship 530 between P2 and P3 becomes a grandchild relationship through this reasoning process.

Various methods may be used to implement the relationship formation process based on the inference rule described above. In particular, embodiments of the present invention configure a graph composed of nodes and edges using triples representing ontology, and profile By defining attributes from information and inferencing rules with F-LOGIC, it is possible to express ontology and individualize social relationships between individuals. You can also add inference rules and extend relationships using ontology-specific expressive powers such as TRANSITIVE, INVERSE, and PARTSOF. In addition, the ontology query language such as SPARQL protocol and RDF Query Language (SPARQL) can be used to query relationships between objects in an already built ontology.

According to the present embodiment, even if the information on the relationship between the sender and the receiver is not provided, it is possible to analyze and expand the relationship between the two by using the ontology and the inference rule. In particular, the present invention can discover new relationships between senders and receivers by extracting keywords and attributes from communication records and profile information.

FIG. 5 is a flowchart illustrating a social relationship analysis method based on a communication record according to an embodiment of the present invention. Each step corresponds to the elements of FIG.

In step 51, communication records between nodes belonging to a predetermined domain and profile information of the nodes are received. This corresponds to the data input unit 100 described above with reference to FIG. 1.

In step 52, a relationship network between nodes is generated from the communication records input in step 51. This corresponds to the relationship network generator 200 described above with reference to FIG. 1.

In step 53 to 52, the relationship strength between nodes in the relationship network generated in step 52 is calculated. This corresponds to the relationship strength calculator 300 described above with reference to FIG. 1.

In step 54 and 51, keywords and attributes are extracted from the communication records and profile information input in step 54 to manage the relationships between nodes ontologies. This corresponds to the ontology manager 400 described above with reference to FIG. 1.

Social relationships between nodes are generated by the inference rules based on the relationship strengths calculated in steps 55 through 53 and the ontology of 54 steps. This corresponds to the social relationship generator 500 described above with reference to FIG. 1.

Meanwhile, the present invention can be embodied as computer readable codes on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored.

Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disks, optical data storage devices, and the like, which may also be implemented in the form of carrier waves (for example, transmission over the Internet). Include. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. In addition, functional programs, codes, and code segments for implementing the present invention can be easily deduced by programmers skilled in the art to which the present invention belongs.

The present invention has been described above with reference to various embodiments thereof. Those skilled in the art will understand that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

1 is a diagram illustrating a social relationship analysis system based on a communication record according to an embodiment of the present invention.

2 is a diagram illustrating a process of generating a relationship network and calculating relationship strength in a social relationship analysis system based on a communication record according to an embodiment of the present invention.

3A to 3C are diagrams for describing various methods of calculating relationship strength in a social relationship analysis system based on a communication record according to an embodiment of the present invention.

4A and 4B are diagrams for explaining a process of inferring a relationship between nodes using an ontology and an inference rule in a social relationship analysis system based on a communication record according to an embodiment of the present invention.

5 is a flowchart illustrating a social relationship analysis method based on a communication record according to an embodiment of the present invention.

Description of the main parts of the drawing

100: data input unit

110: communication record 120: profile information

200: relationship network generator 250: relationship network

300: relationship strength calculation unit

350: Relationship network with relationship strength

400: Ontology Management Department

500: social relationship generation unit 550: social relationship network

Claims (11)

At a data input unit, receiving communication records between nodes belonging to a predetermined domain and profile information of the nodes; Generating, at a relationship network generation unit, a relationship network between the nodes from the input communication record; Calculating, by a relationship strength calculator, relationship strengths between nodes in the generated relationship network; In the ontology manager, extracting keywords and attributes from the inputted communication record and profile information, respectively, and managing the relationships among the nodes by ontology; And Generating a social relationship between the nodes in a social relationship generating unit by a predetermined reasoning rule based on the calculated relationship strength and the ontology, Ontology of managing the relationship between the nodes, Extract an attribute from the input profile information, extract a keyword corresponding to the extracted attribute from the input communication record, And extracting a new object from the nodes associated with the keyword and defining the new object as a relationship between the nodes. delete The method of claim 1, Calculating the strength of the relationship between the nodes, Extracting a communication time or communication frequency between the nodes for each communication means from the input communication record; And Calculating a strength factor from the extracted communication time communication frequency in consideration of the communication means. The method of claim 3, wherein The calculated strength factor is, Or a first intensity factor calculated in consideration of the charge ratio for each communication means among the input communication records; A second strength factor calculated by taking into account communication log statistics of a predetermined group to which the nodes belong; And a method of combining the first intensity factor and the second intensity factor. The method of claim 1, The predetermined reasoning rule is, Is a sufficient condition for social relations between the nodes, And at least one rule derivable from the nodes in consideration of the input communication record, profile information, and the ontology. A non-transitory computer-readable recording medium having recorded thereon a program for executing the method of claim 1. A data input unit for receiving communication records between nodes belonging to a predetermined domain and profile information of the nodes; A relationship network generator for generating a relationship network between the nodes from the input communication record; A relationship strength calculator for calculating a relationship strength between nodes in the generated relationship network; An ontology manager which extracts keywords and attributes from the inputted communication record and profile information, respectively, and manages relationships among the nodes by ontology; And A social relationship generating unit generating a social relationship between the nodes by a predetermined reasoning rule based on the calculated relationship strength and the ontology, The ontology management unit, Extract an attribute from the input profile information, extract a keyword corresponding to the extracted attribute from the input communication record, And extracting a new object from the nodes associated with the keyword to define a relationship between the nodes. delete The method of claim 7, wherein The relationship strength calculation unit, Extracting a communication time or a communication frequency between the nodes for each communication means from the input communication record; And a social factor analysis system for calculating a strength factor from the extracted communication time communication frequency in consideration of the communication means. The method of claim 9, The calculated strength factor is, Or a first intensity factor calculated in consideration of the charge ratio for each communication means among the input communication records; A second strength factor calculated by taking into account communication log statistics of a predetermined group to which the nodes belong; And a social intensity analysis system that is a combination of the first intensity factor and the second intensity factor. The method of claim 7, wherein The predetermined reasoning rule is, Is a sufficient condition for social relations between the nodes, And at least one rule derivable from the nodes in consideration of the input communication record, profile information, and the ontology.

Images