KR101216155B1 - System and methdo for searching social - Google Patents

Abstract

The social search system includes identification information of each user, a graph indicating a relationship with another user, a database storing a topic of a post posted by the user, a gateway for extracting a topic from a question received from a questioner terminal, and the question from the database. A conversation manager that generates a candidate responder list including identification information of a user including the same topic as a topic extracted from the topic, and a recent topic from posts posted by a user corresponding to the identification information included in the candidate answerer list for a recently designated period A topic collection unit for extracting a and a predetermined weight is applied to each of the topics stored in the database and the recent topic, and a rank calculation unit for calculating the rank by summing the numbers.

Description

Social search system and method {SYSTEM AND METHDO FOR SEARCHING SOCIAL}

The present invention relates to a search technique, and more particularly, to a technique for searching an answerer to answer a question of a questioner on a social network.

Recently, online social network services represented by Twitter and Facebook have become popular. Social network services differ in many ways from the traditional web. In particular, the conventional web has drawn web graphs through links between contents such as web pages, but in online social network services, graphs can be drawn through users and content or links between users and other users. This allows us to consider the property of "intimacy" in addition to the "authority" used in algorithms such as PageRank to evaluate the importance of graph nodes. Also, as online social network services are in the spotlight, social search technology using these social network characteristics has been actively studied.

A representative technology of social search technology is a social question / answer service that analyzes a user's social network when a user asks a question and recommends a user to provide an answer to the question. As a social question / answer service, an algorithm for ranking in consideration of both a user's network and content has been proposed, and it is recognized that this algorithm is effective in a large social network. However, the rank algorithm of the social question / answer service has a disadvantage that is insensitive to the recent change of interest of the user. That is, in the social question / answer service, in recommending a user to provide an answer, there is a problem in that the recommended answerer does not provide a proper answer to the question because the user recommends without considering the user's recent interests.

The present invention proposes a system for searching for a user who will provide an answer in consideration of a user's recent interests.

According to an aspect of the present invention, a database for storing identification information of each user, a graph indicating a relationship with another user, and a topic of a post posted by the user; A gateway for extracting a topic from a question received from the interrogator terminal; A conversation manager configured to generate a candidate answerer list including identification information of a user including a topic identical to a topic extracted from the question in the database; A topic collector configured to extract a recent topic from posts posted by a user corresponding to the identification information included in the candidate responder list during a recently designated period; And a rank calculator configured to calculate a number by applying a predetermined weight to each of the topics stored in the database and the recent topics, and add the numbers to calculate a rank.

The conversation manager may select an answerer among users corresponding to the candidate answerer list according to the rank, transmit the question to the answerer terminal of the selected answerer, and receive an answer corresponding to the question.

The conversation manager may transmit the answer to the interrogator terminal.

The rank setting unit calculates the rank according to the following equation,

는 질문자의 질문에 대한 답변자의 상기 랭크이고,

는 상기 그래프 상에서 질문자와 답변자 간의 거리이고,

는 질문에 포함된 전체 토픽 수와 임의의 토픽 t의 수 간 비율이며,

는 질문에 포함된 전체 토픽 수와 최근 토픽 중 임의의 토픽 t의 수 간 비율일 수 있다.

Is the above rank of the answerer to the questioner's question,

Is the distance between the questioner and the answerer on the graph,

Is the ratio between the total number of topics included in the question and the number of random topics t,

May be the ratio between the total number of topics included in the question and the number of any topic t of recent topics.

The importer for receiving the identification information of each user, the graph, the topic of the post posted by the user from the social network system may further include an importer for storing in the database.

According to another aspect of the present invention, a social search system, the method for searching a social network, comprising: storing identification information of each user, a graph indicating a relationship with another user, a topic of a post posted by the user; Extracting a topic from a question received from the interrogator terminal; Generating a candidate responder list including identification information of a user including a topic identical to a topic extracted from the question in the database; Extracting a recent topic from posts posted by a user corresponding to identification information included in the candidate responder list during a recently designated period; And calculating a number obtained by applying a predetermined weight to each of the topics stored in the database and the recent topics, and adding the numbers to calculate a rank.

The social search method includes selecting an answerer among users corresponding to the candidate answerer list according to the rank; And transmitting the question to an answerer terminal of the selected answerer, and receiving an answer corresponding to the question.

The social search method may further include transmitting the answer to the questioner terminal.

Computing the rank is to calculate the rank according to the following equation,

Computing the rank is to calculate the rank according to the following equation,

는 질문자의 질문에 대한 답변자의 상기 랭크이고,

는 상기 그래프 상에서 질문자와 답변자 간의 거리이고,

는 질문에 포함된 전체 토픽 수와 임의의 토픽 t의 수 간 비율이며,

는 질문에 포함된 전체 토픽 수와 최근 토픽 중 임의의 토픽 t의 수 간 비율일 수 있다.

Is the above rank of the answerer to the questioner's question,

Is the distance between the questioner and the answerer on the graph,

Is the ratio between the total number of topics included in the question and the number of random topics t,

May be the ratio between the total number of topics included in the question and the number of any topic t of recent topics.

According to the present invention, by recommending a user to provide an answer in consideration of the user's recent interests, the probability of the answerer providing an appropriate answer can be increased.

1 is a block diagram illustrating a social search system.
2 illustrates a graph included in social network information received by a social search system from a social network system.
3 is a flow chart illustrating a social search method in which a social search system provides answers to questioners' questions.

The present invention may be variously modified and have various embodiments, and specific embodiments will be illustrated in the drawings and described in detail with reference to the accompanying drawings. It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

Further, in this specification, when an element is referred to as " transmitting " a signal to another element, the element can be directly connected to the other element to transmit a signal, It should be understood that the signal may be transmitted by mediating another component in the middle.

1 is a block diagram illustrating a social search system.

Referring to FIG. 1, the social search system includes an importer 110, a database 120, a gateway 130, a conversation manager 140, a rank calculator 150, and a topic collector 160.

The importer 110 is connected to the social network system 50 so that when an event for the first time a user registers for the social network service or modifies his or her profile occurs, the user identification information from the social network system 50, each user Request and receive a post posted by a user using a social network service and a graph indicating a relationship between the user and other users. The graph will be described in detail later with reference to FIG. 2. For example, the importer 110 may be connected to a social network system 50 that provides a social network service such as FACEBOOK, LINKEDIN, TWITTER, and the like. The importer 110 receives user identification information, graphs, and posts from the social network system 50 when the user performs membership registration or profile modification on the social network system 50 through the terminal. The importer 110 extracts a word from the natural language sentence of the received post. The importer 110 selects the extracted word as a topic, and stores the selected topic and social network information in the database 120. In this case, when there is a duplicate word among the extracted words, the importer 110 may store the duplicated topic for the word in the database 120 or store the duplicate number of the corresponding topic in the database 120. For example, if the word 'social' is duplicated three times in a post, the importer 110 stores the topic 'social' three times in the database 120, or duplicates the topic 'social' and the topic. Store the number 3.

The database 120 stores user identification information, topics, and graphs for each user using the social network service received from the importer 110. The database 120 may provide social network information at the request of the conversation manager 140.

The gateway 130 receives a question or answer input by the user from the social network system 50, and extracts a topic from the question or answer. That is, the gateway 130 extracts a word from a natural language sentence included in the question or answer, and selects the extracted word as a topic of the question or answer. The gateway 130 transmits the selected topic to the conversation manager 140.

The conversation manager 140 extracts identification information of users corresponding to a topic received from the gateway 130 from the database 120. For example, when the conversation manager 140 receives the topic 'food' from the gateway 130, the conversation manager 140 extracts user identification information whose corresponding topic is stored as 'food' among user identification information stored in the database 120. In this case, the extracted user identification information may be one or more, and the conversation manager 140 generates a candidate answerer list including one or more user identification information and transmits the list to the rank calculator 150.

The conversation manager 140 receives a rank corresponding to each user identification information included in the candidate responder list from the rank calculator 150. The conversation manager 140 transmits a question through the gateway 130 to the upper-order answerer terminal 20 having a rank corresponding to each user identification information. The conversation manager 140 receives an answer from the answerer terminal 20 and transmits the received answer to the questioner terminal 10. For example, the conversation manager 140 may ask a question with the number of the highest answering device terminals (the answering device terminal having the highest rank and the answering device terminal having the second rank) of the number corresponding to each user identification information (in this embodiment, 2). And transmit an answer from each answerer terminal. The conversation manager 140 may transmit the received answers to the interrogator terminal 10.

The rank calculation unit 150 sets a rank corresponding to each user identification information of the candidate answerer list. The rank calculator 150 transmits the candidate responder list to the topic collector 160 to request a topic (hereinafter, referred to as a recent topic) of interest to respondents corresponding to each user identification information. The rank calculation unit 150 receives a recent topic corresponding to each user identification information from the topic collection unit 160 and calculates a rank of each user identification information included in the candidate answerer list in consideration of the recent topic. The rank calculator 150 calculates a rank by applying a weight to a topic corresponding to each user identification information included in the candidate answerer list and a recent topic among the topics stored in the database 120. The rank calculation unit 150 is set using Equation 1 below.

[Equation 1]

는 답변자,

는 질문자, q는 질문, t는 임의의 토픽을 의미한다. 또한

는 질문자의 질문 q에 대한 답변자의 랭크를 의미한다. 또한

는 그래프 상에서 질문자와 답변자 간의 거리이고,

는 질문에 포함된 전체 토픽 수와 임의의 토픽 t의 수 간 비율이며,

는 질문에 포함된 전체 토픽 수와 최근 토픽 중 임의의 토픽 t의 수 간 비율을 의미한다. At this time,

Is the answerer,

Is a questioner, q is a question, and t is any topic. Also

Means the rank of the answerer to the question q of the questioner. Also

Is the distance between the questioner and the answerer on the graph,

Is the ratio between the total number of topics included in the question and the number of random topics t,

Is the ratio between the total number of topics included in the question and the number of any topic t of recent topics.

 And are defined by Equations 2 and 3 below.

&Quot; (2) "

&Quot; (3) "

는 답변자 의 게시물에 포함된 전체 토픽 수와 임의의 토픽 t의 수 간 비율이며,

는 답변자 의 최근 토픽의 전체 개수와 최근 토픽 중 임의의 토픽 t의 수 간 비율이다.At this time,

Is the ratio between the total number of topics included in the responder's post and the number of random topics t,

Is the ratio between the total number of respondents' recent topics and the number of random topics t in the recent topics.

The rank calculator 150 transmits the rank calculated for each user identification information of the candidate respondent list to the conversation manager 140.

The topic collection unit 160 searches for posts written after a point in time before a specified period from the presenter's posts corresponding to each user identification information included in the candidate responder list. At this time, the topic collection unit 160 may search for posts with respect to a server that provides a predetermined service (eg, blog, SNS, web bookmarking (social bookmarking)). The topic collector 160 extracts a tag or a bookmark included in each searcher's post. The topic collecting unit 160 selects the extracted tag or bookmark as the latest topic of the corresponding answerer. For example, the topic collection unit 160 accesses and searches a post corresponding to one user identification information included in the candidate respondent list to a designated server (a blog and a web bookmarking providing server). At this time, the topic collection unit 160 searches for posts made during a specified period (eg, the last five days) among posts of the blog and web bookmarks corresponding to the user identification information. The topic collecting unit 160 selects a tag corresponding to a post of a searched blog and a searched web bookmark as a recent topic. The topic collecting unit 160 transmits the latest topic to the conversation managing unit 140.

2 is a diagram illustrating a graph included in social network information received by a social search system from a social network system.

는 질문자와 답변자 간의 그래프 상 홉 수를 의미한다. 예를 들어, 사용자 A가 질문자이고, 사용자 G가 답변자인 경우,

는 3이다.
2, a graph included in social network information represents a relationship between users. 1-hop in the graph illustrated in FIG. 2 means that the social network system is set as a friend. For example, user A is set up in a social network system with users B, C, and D, and user D is set up in a friend relationship with users A, B, and F. Therefore, in Equation 1 described above

Is the number of hops in the graph between the questioner and the answerer. For example, if user A is a questioner and user G is an answerer,

Is 3.

3 is a flowchart illustrating a social search method in which a social search system provides an answer to a question of a questioner.

Referring to FIG. 3, in step 310, the social search system receives social network information and posts posted by each user from the social network system.

In operation 320, the social search system extracts a word from the natural language sentence of the post received in operation 310, and selects the extracted word as a topic of the corresponding user.

In step 330 the social search system receives a question from the caller terminal.

In operation 340, the social search system extracts a word from the natural language sentence of the question received in operation 330, and selects the extracted word as a topic of the question.

In step 350, the social search system searches for users who have posted posts about the topic that corresponds to the topic of the question. The social search system generates a list of candidate respondents containing the retrieved user's identification information.

In step 360, the social search system extracts a recent topic of each user included in the candidate responder list.

In operation 370, the social search system calculates a rank of each user (candidate responder) included in the candidate responder list. The process of calculating the rank of each user has been described above with reference to Equation 1.

In step 380, the social search system selects the top-ranked number of the top users among the users included in the candidate responder list as the responders.

In operation 390, the social search system transmits the question received in operation 330 to the answerer terminal 20 of the answerer selected in operation 380, and receives an answer from the corresponding answerer terminal 20.

In operation 395, the social search system transmits the answer received in operation 390 to the interrogator terminal 10.

So far I looked at the center of the embodiment for the present invention. Many embodiments other than the above-described embodiments are within the claims of the present invention. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. The disclosed embodiments should, therefore, be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

Claims (9)

A database for storing identification information of each user, a graph indicating a relationship with another user, and a topic of a post posted by the user;
A gateway for extracting a topic from a question received from the interrogator terminal;
A conversation manager configured to generate a candidate responder list including identification information of a user including a topic identical to a topic extracted from the question by referring to the database;
A topic collector configured to extract a recent topic from posts posted by a user corresponding to the identification information included in the candidate responder list during a recently designated period; And
A rank calculator configured to calculate a number obtained by applying a predetermined weight to each of the topics stored in the database and the most recent topic, and calculate the rank by adding the numbers;
Social search system comprising a.
The method according to claim 1,
The conversation manager selects an answerer among users corresponding to the candidate answerer list according to the rank,
And transmitting the question to an answerer terminal of a selected answerer and receiving an answer corresponding to the question.
The method of claim 2,
The conversation manager transmits the answer to the questioner terminal.
The method according to claim 1,
The rank setting unit calculates the rank according to the following equation,

Is the above rank of the answerer to the questioner's question,

Is the distance between the questioner and the answerer on the graph,

Is the ratio between the total number of topics included in the question and the number of random topics t,

Is a ratio between the total number of topics included in the question and the number of any topic t of recent topics.
The method according to claim 1,
And an importer for receiving the identification information of the user, the graph, and a topic of a post posted by the user from a social network system and storing the identification information in the database.
The social search system is a method for searching a social network,
Storing identification information of each user, a graph indicating a relationship with another user, and a topic of a post posted by the user in a database;
Extracting a topic from a question received from the interrogator terminal;
Generating a candidate responder list including identification information of a user including a topic identical to a topic extracted from the question by referring to the database;
Extracting a recent topic from posts posted by a user corresponding to identification information included in the candidate responder list during a recently designated period; And
Calculating a number by applying a predetermined weight to each of the topics stored in the database and the most recent topic, and calculating the rank by summing the numbers
Social search method comprising a.
The method of claim 6,
Selecting an answerer among users corresponding to the candidate answerer list according to the rank; And
And transmitting the question to an answerer terminal of the selected answerer, and receiving an answer corresponding to the question.
The method of claim 7, wherein
And transmitting the answer to the questioner terminal.
The method of claim 6,
Computing the rank is to calculate the rank according to the following equation,

Is the above rank of the answerer to the questioner's question,

Is the distance between the questioner and the answerer on the graph,

Is the ratio between the total number of topics included in the question and the number of random topics t,

Is a ratio between the total number of topics included in the question and the number of any topic t of recent topics.

Images