JP6419021B2 - Extraction apparatus, extraction method and extraction program - Google Patents

Description

  The present invention relates to an extraction apparatus, an extraction method, and an extraction program.

  2. Description of the Related Art In recent years, techniques for eliminating postings with low credibility are known for posting services such as electronic bulletin boards using a computer network. For example, in such a technique, based on the number of evaluation standard information that is the same topic as the evaluation target information and has the same opinion as the evaluation target information, and the evaluation of the evaluation standard information to the sender, A numerical value indicating credibility is calculated.

JP 2014-134944 A JP2013-77155A

  However, in the above-described conventional technology, it is not always possible to extract a post indicating reliability with high accuracy. For example, in the above-described conventional technique, a numerical value indicating the credibility of the evaluation target information is simply calculated, so that it is difficult to extract a reply post indicating the reliability of the outgoing post with high accuracy.

  The present application has been made in view of the above, and an object thereof is to provide an extraction apparatus, an extraction method, and an extraction program that can extract a post indicating reliability with high accuracy.

  The extraction device according to the present application is extracted by a first extraction unit that extracts a reply post including a predetermined character from reply posts replied to a transmission post sent by a poster, and the first extraction unit. And a second extraction unit for extracting a reply post indicating the reliability of the outgoing post based on the thought related information indicating the thought of the poster who posted the reply post and the degree of the thought. And

  According to one aspect of the embodiment, it is possible to extract a post indicating the reliability with high accuracy.

FIG. 1 is an explanatory diagram illustrating an example of an extraction process performed by the extraction apparatus according to the embodiment. FIG. 2 is a diagram illustrating a configuration example of the extraction device according to the embodiment. FIG. 3 is a diagram illustrating an example of a user information storage unit according to the embodiment. FIG. 4 is a diagram illustrating an example of the posted information storage unit according to the embodiment. FIG. 5 is a diagram illustrating an example of the indication information storage unit according to the embodiment. FIG. 6 is an explanatory diagram for explaining a flow of processing by the extraction device according to the embodiment. FIG. 7 is a flowchart illustrating an extraction processing procedure performed by the extraction apparatus according to the embodiment. FIG. 8 is an explanatory diagram for explaining a flow of extraction processing by the extraction device according to the modification. FIG. 9 is an explanatory diagram for explaining a flow of extraction processing by the extraction device according to the modification. FIG. 10 is a hardware configuration diagram illustrating an example of a computer that realizes the function of the extraction device.

  Hereinafter, an embodiment (hereinafter referred to as “embodiment”) for carrying out an extraction apparatus, an extraction method, and an extraction program according to the present application will be described in detail with reference to the drawings. In addition, the extraction apparatus, the extraction method, and the extraction program which concern on this application are not limited by this embodiment. Moreover, in the following embodiment, the same code | symbol is attached | subjected to the same site | part and the overlapping description is abbreviate | omitted.

[1. Embodiment)
[1-1. Extraction Processing According to Embodiment]
First, an example of the extraction process according to the embodiment will be described with reference to FIG. FIG. 1 is an explanatory diagram illustrating an example of an extraction process performed by the extraction apparatus 100 according to the embodiment. In the example of FIG. 1, the extraction device 100 analyzes the information posted to the posting service and performs an extraction process for extracting a reply post indicating the reliability of the outgoing post.

  The extraction device 100 replies pointing out reliability such as a posting posted to a posting service viewed by an unspecified number of users, such as an electronic bulletin board or SNS (Social Networking Service), by a poster. It is a server device that extracts posts. In the present embodiment, a post referred to by other contributors as a hit is defined as a “buzz post”.

  Generally, in the posting service, various messages are exchanged between users in accordance with the purpose of the bulletin board. However, postings contrary to the purpose may be made. For example, on a stock price bulletin board whose purpose is exchange of information related to stock prices, there are many useful posts such as information that can be used as reference for stock trading and the impressions of individual investors, but it becomes noise such as advertisements called spam and abuse to other users There may be posts. In addition, on the stock price bulletin board, there is a case where a posting is made to induce others to buy or sell by posting information that is contrary to an individual's idea about buying and selling stock. Such talk postings may mislead the user of the bulletin board, impair the reliability of the bulletin board itself, and also cause noise in data analysis. Therefore, it is desirable that the postings are automatically excluded in the same manner as spam. Therefore, the extraction apparatus 100 executes a process of extracting a reply post indicating that the outgoing posting is a hit post.

  Here, it is assumed that the extraction apparatus 100 holds in advance information related to a user who uses the posting service and information related to outgoing postings and reply postings. In the present embodiment, the posting on the stock price bulletin board includes the target brand, tag information indicating the poster's philosophy, title, comment, and information regarding the posting date and time. In the example of FIG. 1, the extraction apparatus 100 holds user display names X1, Y1, and Z1 in association with profile images P1, P2, and P3, respectively, as information about users who use the posting service. Further, the extraction apparatus 100 uses the user ID of the contributor of the display name “X1” as the information regarding the outgoing posting Po1, the title Ti1 “Is the start of downtrend”, and the comment Ar1 “Company A's current stock price is too high "If it collapses in any way, it will crash," posting time Tm1 "January 24 10:02" when the posted post Po1 was posted on the stock bulletin board, and emotion Em1 "strongly" “I want to sell” is stored in association with the post ID “111”.

  Further, the extraction device 100 uses the user ID of the contributor with the display name “Y1”, the title Ti2 “Re: start of downtrend”, and the comment Ar2 “Such sales as information on the reply post Po2 to the outgoing post Po1. , "Post time Tm2" January 24 10:04 "of reply post Po2 and poster's emotion Em2" I want to buy strongly "are associated with post ID" 112 " Hold. Further, the extraction device 100 uses the user ID of the contributor of the display name “Z1”, the title Ti3 “Re: start of downtrend”, and the comment Ar3 “slightly more as information on the reply post Po3 to the outgoing post Po1. ”And the posting time Tm3“ January 24 10:05 ”of the reply posting Po3 and the contributor's emotion Em3“ Looking ”are stored in association with the posting ID“ 113 ”.

  First, the extraction device 100 extracts a reply post Po2 including a predetermined character from reply posts Po2 to Po3 sent back to the outgoing post Po1 sent by the poster “X1”. In the example of FIG. 1, the extraction apparatus 100 extracts a reply post Po2 in which “煽” is included as a predetermined character in a comment from reply posts Po2 to Po3.

  Subsequently, the extraction apparatus 100 transmits based on an emotion (corresponding to an example of thought-related information) and an emotion score (corresponding to an example of the degree of thought) indicating the thought of the poster who posted the extracted reply post. A reply post indicating the reliability of the post Po1 is extracted. Here, the feelings indicating the poster's thought are indicated by five types of tag information: “I want to buy strongly”, “I want to buy”, “I want to sell”, “I want to sell”, and “I want to sell strongly”. Emotion scores are “2”, “1”, “0”, “−1” for emotions of “I want to buy strongly”, “I want to buy”, “Look at”, “I want to sell”, “I want to sell strongly”, respectively. , “−2”. In the example of FIG. 1, the extraction apparatus 100 is based on the emotion Em2 “I want to buy strongly” and the emotion score “2” of “I want to buy strongly” of the poster “Y1” who posted the extracted reply post Po2. Then, a reply post Po2 indicating that the outgoing post Po1 is a hit post is extracted. Specifically, in the extraction device 100, since the emotion Em2 “I want to buy strongly” of the reply post Po2 shows a thought opposite to the emotion Em1 “I want to sell strongly” of the outgoing post Po1, the outgoing post Po1 is a hit post The reply post Po2 is extracted as a reply post pointing out.

  As described above, the extraction apparatus 100 according to the embodiment extracts the reply post Po2 including a predetermined character from the reply posts Po2 to Po3 sent back to the outgoing post Po1 sent by the poster. Further, the extraction apparatus 100 extracts the reply post Po2 indicating the reliability of the outgoing post Po1, based on the emotion Em2 indicating the thought of the poster who posted the extracted reply post Po2 and the degree of the emotion Em2.

  Thereby, since the extraction apparatus 100 can extract a reply post that points out an outgoing post based on a contributor's idea, it can extract a post that points out reliability with high accuracy. For example, the extraction apparatus 100 can extract a reply post having a thought opposite to the idea of the outgoing post, and thus can extract a reply post that denies the outgoing post. For this reason, the extraction apparatus 100 can extract posts that detract from the quality of the bulletin board and maintain the order of the bulletin board.

  1 shows an example in which two reply posts Po2 to Po3 are posted with respect to the outgoing post Po1 for the sake of simplicity. Multiple reply posts are posted.

[1-2. Configuration of Extraction Device According to Embodiment]
Next, the configuration of the extraction apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 2 is a diagram illustrating a configuration example of the extraction apparatus 100 according to the embodiment. As illustrated in FIG. 2, the extraction device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The extraction device 100 includes an input unit (for example, a keyboard and a mouse) that receives various operations from an administrator who uses the extraction device 100, and a display unit (for example, a liquid crystal display) for displaying various types of information. You may have.

(About the communication unit 110)
The communication unit 110 is realized by a NIC or the like, for example. The communication unit 110 is connected to a network by wire or wireless, and transmits / receives information to / from a server device that provides a posting service, a server device that manages the posting service, and the like via the network.

(About the storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage extraction device such as a hard disk or an optical disk. The storage unit 120 includes a user information storage unit 121, a posting information storage unit 122, and an indication information storage unit 123.

(User information storage unit 121)
The user information storage unit 121 stores information about the user for each user who uses the posting service. Here, FIG. 3 shows an example of the user information storage unit 121 according to the embodiment. As illustrated in FIG. 3, the user information storage unit 121 includes items such as “user ID”, “display name”, and “profile image”.

  “User ID” indicates identification information for identifying a user. For example, “user ID” stores a character string formed of alphanumeric characters and uniquely assigned to each user. “Display name” indicates a user name displayed on the posting service. For example, “display name” stores a nickname set by the user himself, a part of the user ID, and the like. “Profile image” indicates image information related to the user. For example, the “profile image” stores an image selected by the user or an uploaded image. The “profile image” is not limited to an image, and may store an address indicating a location where the image is stored.

  That is, FIG. 3 shows an example in which the user with the display name “X1” identified by the user ID “U1” has the profile image “P1”. The user information storage unit 121 may further include various information items regarding the user such as “gender”, “job type”, “mail address”, “address”, “password”, and the like.

(Posted information storage unit 122)
The posted information storage unit 122 stores various types of information posted to the posting service. Specifically, the posting information storage unit 122 stores information related to outgoing posts sent by the poster to the posting service and reply posts to outgoing posts. Here, FIG. 4 shows an example of the posted information storage unit 122 according to the embodiment. As illustrated in FIG. 4, the posting information storage unit 122 includes “post ID”, “original post ID”, “user ID of the poster”, “title”, “comment”, “emotion”, “emotion score”, and It has an item such as “post date”.

  “Post ID” indicates identification information for identifying a post. For example, the “post ID” stores a unique ID assigned to each post in the posting service. The “original posting ID” indicates identification information for identifying the outgoing posting that is the source of the reply posting. If the post is an outgoing post, a blank is stored in the “original post ID”. “Contributor user ID” indicates identification information for identifying a contributor who is a user who has made a posting. “Title” indicates the title of the post. For example, “Title” stores a character string input by a contributor into the title input field of the posting service. “Comment” indicates the posted content. For example, “Comment” stores a character string input by a contributor in the comment input field of the posting service. “Emotion” indicates the philosophy of the poster. For example, “emotion” stores thought-related information indicating the idea selected by the poster in the tag information selection field of the posting service. As an example, when the posting service is a stock price bulletin board, there are five types of “emotions” such as “I want to buy strongly”, “I want to buy”, “I want to sell”, “I want to sell”, and “I want to sell strongly” Of the thought related information, thought related information selected by the poster is stored. The “emotion score” indicates the degree of thought of the poster. For example, the “emotion score” stores a numerical value of the strength of the poster's thought. For example, if the posting service is a stock price bulletin board, the “emotion score” includes emotions of “I want to buy strongly”, “I want to buy”, “I want to sell”, “I want to sell”, “I want to sell strongly” “2”, “1”, “0”, “−1”, and “−2” are stored. “Posting date / time” indicates the date and time when posting is made to the posting service. For example, the “date and time of posting” stores the date and time when the posting was made.

  That is, FIG. 4 shows an example in which the outgoing posting identified by the posting ID “111” is posted by the contributor with the user ID “U1”. In addition, the outgoing posting with the posting ID “111” shows an example in which “Is the start of the downtrend” entered in the title column by the contributor with the user ID “U1”? In addition, the outgoing posting with the posting ID “111” is input by the contributor with the user ID “U1” in the comment column “The current stock price of company A is too high. An example is shown. In addition, the outgoing posting with the posting ID “111” shows an example in which the emotion “I want to sell strongly” is selected by the poster with the user ID “U1” and the emotion score is “−2”. In addition, the outgoing posting with the posting ID “111” is an example posted by the posting person with the user ID “U1” on the posting date “January 24, 10:02”.

  FIG. 4 shows an example in which the reply post identified by the post ID “112” is a post for the outgoing post with the post ID “111”. Further, the reply post identified by the post ID “112” shows an example in which the emotion “I want to buy strongly” is selected by the contributor with the user ID “U2” and the emotion score is “2”.

(About the indication information storage unit 123)
The indication information storage unit 123 stores information related to the reply posting indicating the reliability of the outgoing posting. Specifically, the indication information storage unit 123 stores information related to a combination of an outgoing post and a reply post indicating the reliability of the outgoing post. For example, the indication information storage unit 123 stores a posting ID of an outgoing posting that is a hit post and a posting ID of an indication post that points out a hit post. Here, FIG. 5 illustrates an example of the indication information storage unit 123 according to the embodiment. As illustrated in FIG. 5, the indication information storage unit 123 includes items such as “buzzing posting ID” and “pointing posting ID”.

  The “posting post ID” indicates identification information for identifying a outgoing post that is a post. For example, the “posted post ID” stores the post ID of the outgoing post determined to be a post. “Pointed post” indicates identification information for identifying a reply post that is a pointed post that points out a hit post. For example, the “posted post ID” stores the post ID of a reply post extracted as being a post.

  That is, FIG. 5 shows an example in which the reply post identified by the post ID “112” is an indication post pointing out that the outgoing post identified by the post ID “111” is a hit post.

(About the control unit 130)
Returning to FIG. 2, the control unit 130 executes, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like using an extraction program stored in a storage device inside the extraction device 100 as a work area. It is realized by doing. The control unit 130 is realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

  As shown in FIG. 2, the control unit 130 includes a reception unit 131, a first extraction unit 132, a second extraction unit 133, and a provision unit 134, and functions and functions of information processing described below. Realize or execute. Note that the internal configuration of the control unit 130 is not limited to the configuration illustrated in FIG. 2, and may be another configuration as long as it performs an extraction process described later. In addition, the connection relationship between the processing units included in the control unit 130 is not limited to the connection relationship illustrated in FIG. 2, and may be another connection relationship.

(About the receiver 131)
The receiving unit 131 receives information about the user from a server device or the like that provides a posting service. For example, the receiving unit 131 receives various types of information such as “user ID”, “display name”, and “profile image” as information about the user. Then, the reception unit 131 stores information regarding the received user in the user information storage unit 121. For example, the reception unit 131 stores “display name” and “profile image” in the user information storage unit 121 in association with “user ID”.

  The receiving unit 131 receives information related to posting posted by the user to the posting service from a server device or the like that provides the posting service. For example, the receiving unit 131 includes “post ID”, “original post ID”, “user ID of the contributor”, “title”, “comment”, “emotion”, “emotion score”, and “post” as information regarding the post. Various information such as “date and time” is received. Then, the reception unit 131 stores information regarding the received posting in the posting information storage unit 122. For example, the receiving unit 131 associates “original post ID”, “user ID of the poster”, “title”, “comment”, “emotion”, “emotion score”, and “post date” with “post ID”. And stored in the user information storage unit 121.

(About the first extraction unit 132)
The first extraction unit 132 extracts a reply post that includes a predetermined character from reply posts that are replied to the outgoing post that is sent by the poster. Specifically, the first extraction unit 132 extracts a reply post including a predetermined character from reply posts viewed from an unspecified number of users (for example, reply posts on an electronic bulletin board or SNS). For example, the first extraction unit 132 refers to “comments” and “titles” stored in the post information storage unit 122 and extracts reply posts in which a predetermined character is included in “comments” or “titles”. As an example, the first extraction unit 132 extracts a reply post that includes a character that denies an outgoing post as a reply post that includes a predetermined character.

  Here, an example in which the target is a stock price bulletin board will be described. When the stock price bulletin board is a target, there is a case where “seriously posting” that induces buying and selling to other users by intentionally posting information contrary to the true idea of the poster may be made. Here, it is difficult to define “sent post” at a level that can be clearly identified. For example, it is difficult to determine whether the posting of the comment “This stock will definitely go up” is a mere impression of a beginner or the like, or a “scoring post” that seeks to gain a profit by raising the stock price. In this case, it is necessary to determine whether or not the posting is made by comprehensively considering stock-related knowledge, the market situation, the individuality of the user who is the poster, and the like. Therefore, in this embodiment, the judgment as to whether or not the posting is made is referred to by other contributors as “buzz” because the judgment by the user of the stock price bulletin board is close to the judgment based on comprehensive knowledge. It is assumed that the posted post is a “sent post”. Specifically, the first extraction unit 132 extracts a reply post that includes “煽” as a character that denies an outgoing post. For example, the first extraction unit 132 extracts a reply post that does not include the character “煽” in the outgoing post that is the source of the reply post and includes the character string “煽” in the reply post. In the example of FIG. 4, the first extraction unit 132 extracts a reply post with a post ID “112” in which “煽” is included in the comment “I won't get on such a sale”.

(About the second extraction unit 133)
The second extraction unit 133 generates a reply post indicating the reliability of the outgoing post based on the thought related information indicating the thought of the poster who posted the reply post extracted by the first extraction unit 132 and the degree of the thought. Extract. For example, since the extraction result by the first extraction unit 132 merely extracts a reply post that includes characters such as “煽”, is it an indication for an outgoing post that is an original post? It cannot be accurately determined whether or not. As an example, there is a case where “it is not an indication of resentment to an outgoing post” such as “meta topic about resentment” or “indication of resentment to a post other than an outgoing post”. For this reason, the second extraction unit 133 extracts a reply post indicating the reliability of the outgoing post based on the idea of the post in order to increase the extraction accuracy of the “posted post”.

  Specifically, the second extraction unit 133 extracts a reply post in which the thought-related information of the reply post indicates a thought opposite to the thought of the outgoing post as a reply post indicating the reliability of the outgoing post. For example, the second extraction unit 133 refers to the “emotion” of the outgoing post and the reply post stored in the post information storage unit 122, and the “emotion” of the reply post reverses the “emotion” of the outgoing post. Extract the reply posts shown. In the example of FIG. 4, the second extraction unit 133 determines that the emotion “I want to buy strongly” of the reply post with the post ID “112” extracted by the first extraction unit 132 is the emotion “ Since the idea of “I want to sell strongly” is indicated in the opposite direction, the reply post with the post ID “112” is extracted.

  Moreover, the 2nd extraction part 133 extracts the transmission posting applicable to the original posting of the reply posting which points out the reliability of transmission posting. In the example of FIG. 4, the second extraction unit 133 extracts the outgoing posting with the posting ID “111” corresponding to the original posting of the reply posting with the posting ID “112”. Then, the second extraction unit 133 stores the extracted reply posting in the indication information storage unit 123 in association with the outgoing posting. In the example of FIGS. 4 and 5, the second extraction unit 133 associates the reply post with the post ID “112” with the outgoing post with the post ID “111” and associates the “posted post ID” and “ Stored in “Summary Post ID”.

  Another example of the extraction process by the first extraction unit 132 and the second extraction unit 133 will be described with reference to FIG. FIG. 6 is an explanatory diagram for explaining a flow of processing by the extraction device according to the embodiment. In the example of FIG. 6, the extraction device 100 holds the display names X1 and Y1 of the users in association with the profile images P1 and P2 as information about the users who use the posting service. In addition, the extraction device 100 holds information related to the outgoing post Po21 and the reply post Po22 to the outgoing post Po21. Specifically, the extraction device 100 uses the user ID of the contributor of the display name “X1”, the title Ti21 “uptrend”, and the comment Ar21 “A company stock suddenly rises as information related to the outgoing posting Po21. It ’s really amazing. I wonder if I will buy it again tomorrow. ”, Posting time Tm21“ February 24 10:02 ”of outgoing post Po21, and poster ’s emotion Em21“ Looking at the scene ”with post ID“ 211 ” Hold in association. Further, the extraction apparatus 100 uses the user ID of the poster having the display name “Y1”, the title Ti22 “Re: Uptrend”, and the comment Ar22 “Look at today's state as information on the reply post Po22 to the outgoing post Po21. I'd like to buy a stock from Company A. I'd like to sell it, but I'd like it to drop in the market. "Reply post Po22 post time Tm22" February 24 10:04 "and contributor's emotion Em22" Look at the scene " Are associated with the post ID “212”.

  Then, in the example of FIG. 6, the first extraction unit 132 reads the comment Ar22 of the reply post Po22 sent back to the outgoing post Po21 sent by the contributor “X1”. “I want to drop in the market, but I want to drop in the market.” Since “煽” is included, the reply post Po22 is extracted.

  Subsequently, in the second extraction unit 133, the feeling Em22 “Looking” of the reply post Po22 extracted by the first extracting unit 132 is the same idea as the feeling Em21 “Looking” of the outgoing post Po21, and the opposite direction Therefore, the reply post Po22 is not extracted as a reply post indicating that the outgoing post Po21 is a hit post. Thereby, the extraction apparatus 100 does not extract, for example, a post including the character “煽” as a meta topic about resentment, so that it is possible to improve the accuracy of extracting a reply post indicating the reliability of the outgoing post. Moreover, since the extraction apparatus 100 does not extract a reply post having the same idea as that of the outgoing post, the extraction accuracy of the reply post pointing out the reliability of the outgoing post can be improved.

(About the providing unit 134)
The providing unit 134 provides information related to a reply post indicating the reliability of the outgoing post. Specifically, the providing unit 134 refers to the “pointed post ID” and the “buzzed post ID” stored in the pointed information storage unit 123, and refers to the reply post pointing out the reliability of the outgoing post, and the reply post Provide information on combinations with outgoing posts corresponding to original posts. For example, the providing unit 134 extracts the ID and comment of the indication post and the ID and comment of the posting post from the post information storage unit 122 and transmits them to a server device or the like that manages the post service. Thereby, the administrator of the posting service can grasp a posting that impairs the quality of the bulletin board.

[1-3. Extraction Processing Procedure According to Embodiment]
Next, a processing procedure performed by the extraction apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 7 is a flowchart illustrating an extraction processing procedure performed by the extraction apparatus 100 according to the embodiment.

  As illustrated in FIG. 7, the extraction apparatus 100 determines whether information related to posting has been received (step S <b> 101). For example, the extraction apparatus 100 may include “post ID”, “original post ID”, “user ID of the poster”, “title”, “comment”, “emotion”, “ It is determined whether various types of information such as “emotion score” and “post date” have been received. Here, when the information regarding the posting is not received (step S101; No), the extraction apparatus 100 stands by until the information regarding the posting is received.

  On the other hand, when receiving information related to posting (step S101; Yes), the extraction device 100 stores the received information related to posting in the posted information storage unit 122 (step S102). Thereby, the extraction apparatus 100 holds information related to posting in the posting service.

  Thereafter, the extraction apparatus 100 extracts a reply post including a predetermined character from reply posts replied to the outgoing post sent by the poster (step S103). For example, the extraction apparatus 100 extracts a reply post that includes a character that denies an outgoing post as a reply post that includes a predetermined character. As an example, the extraction apparatus 100 extracts a reply post including “文字” in the “comment” or “title” from the posted information storage unit 122.

  Subsequently, the extraction apparatus 100 extracts a reply post indicating the reliability of the outgoing post based on the thought related information indicating the thought of the contributor who posted the extracted reply post and the degree of the thought (step S104). ). For example, the extraction apparatus 100 extracts a reply post in which the thought-related information of the reply post indicates a thought in the opposite direction to the thought of the outgoing post as a reply post pointing out the reliability of the outgoing post.

  After that, the extraction apparatus 100 stores the extracted reply posting in the indication information storage unit 123 in association with the original posting outgoing posting (step S105). For example, the extraction apparatus 100 stores the extracted post ID of the reply post in the indication information storage unit 123 in association with the post ID of the outgoing post.

Then, the extraction apparatus 100 provides information regarding the extracted reply posting (step S1).
06). For example, the extraction device 100 refers to the “pointed post ID” and the “sent post ID” stored in the pointed information storage unit 123, and refers to the reply post pointing out the reliability of the outgoing post and the original post of the reply post. Provide information on applicable outgoing posts. As an example, the extraction device 100 extracts the ID and comment of the pointed posting and the ID and comment of the posting post from the posting information storage unit 122 and transmits them to a server device or the like that manages the posting service.

[1-4. Effects of the embodiment
As described above, the extraction device 100 according to the embodiment includes the first extraction unit 132 and the second extraction unit 133. The first extraction unit 132 extracts a reply post that includes a predetermined character from reply posts that are replied to the outgoing post that is sent by the poster. The second extraction unit 133 generates a reply post indicating the reliability of the outgoing post based on the thought related information indicating the thought of the poster who posted the reply post extracted by the first extraction unit 132 and the degree of the thought. Extract.

  Thereby, since the extraction apparatus 100 can extract a reply post that points out an outgoing post based on a contributor's idea, it can extract a post that points out reliability with high accuracy. For this reason, since the extraction apparatus 100 can provide a posting indicating the reliability to the posting service manager, the quality of the posting service can be improved. In addition, the extraction apparatus 100 considers the philosophy of the poster in addition to the linguistic information indicating the reliability, so that the accuracy of extracting the posting indicating the reliability can be improved.

  Further, in the extraction device 100 according to the embodiment, the second extraction unit 133 is a reply posting in which the thought-related information of the reply post indicates a thought opposite to the thought of the outgoing post as a reply post pointing out the reliability of the outgoing post. To extract.

  As a result, the extraction apparatus 100 can extract a reply post having a thought opposite to the idea of the outgoing post, and thus can extract a reply post pointing out the reliability of the outgoing post with high accuracy. For this reason, for example, the extraction apparatus 100 can provide the posting service administrator with a post that damages the quality of the bulletin board and can maintain the order of the bulletin board.

  Further, in the extraction device 100 according to the embodiment, the first extraction unit 132 extracts a reply post that includes a character that denies an outgoing post as a reply post that includes a predetermined character.

  As a result, the extraction apparatus 100 can extract a reply post that denies an outgoing post, and thus can select a reply post that points to an outgoing post that impairs the quality of the posting service.

  In the extraction device 100 according to the embodiment, the first extraction unit 132 extracts a reply post including a predetermined character from reply posts viewed from an unspecified number of users.

  Thereby, since the extraction apparatus 100 can extract, for example, a post that has a great influence on the user of the posting service, the order of the posting service can be maintained and the satisfaction of the user can be increased. Can do.

[2. (Modification)
The extraction device 100 according to the above-described embodiment may be implemented in various different forms other than the above-described embodiment. Therefore, in the following, another embodiment of the extraction device 100 will be described.

[2-1. (Responding posts are extracted based on the degree of thought)
In the above embodiment, the extraction device 100 points out the reliability of the outgoing post based on the thought-related information and the degree of the thought indicating the thought of the poster who posted the reply post extracted by the first extraction unit 132. An example of extracting reply posts was given. Here, the extraction apparatus 100 may extract a reply post based on a difference between the degree of the idea of the reply post and the degree of the idea of the outgoing post.

  Specifically, the second extraction unit 133 of the extraction apparatus 100 determines that the difference between the degree of the thought of the reply post and the degree of the thought of the outgoing post satisfies a predetermined condition as a reply post pointing out the reliability of the outgoing post. Extract reply posts. For example, the second extraction unit 133 is a reply post indicating the reliability of the outgoing post as a reply post pointing out the reliability of the outgoing post, and the thought-related information is a reply post showing the idea in the opposite direction to the outgoing post idea. Reply posts whose difference from the degree of the idea is equal to or greater than a predetermined threshold are extracted. This point will be described with reference to FIG.

  FIG. 8 is an explanatory diagram for explaining the flow of extraction processing by the extraction apparatus 100 according to the modification. In the example of FIG. 8, the extraction apparatus 100 holds user display names X1, Y1, and Z1 in association with profile images P1, P2, and P3, respectively, as information about users who use the posting service. In addition to the information related to the outgoing post Po1 and the reply post Po2 shown in FIG. 1, the extraction apparatus 100 holds information related to the return post Po33 for the outgoing post Po1. Specifically, the extraction device 100 wants to buy the user ID of the contributor with the display name “Z1”, the title Ti33 “Re: start of downtrend”, and the comment Ar33 “a little bit of information as the reply post Po33”. “I wonder if it ’s selling.” Post time Tm33 “January 24th 10:07” of reply post Po33 and poster's emotion Em33 “I want to buy” are associated with post ID “313”. Hold.

  First, the extraction apparatus 100 extracts a reply post including “煽” as a predetermined character in a comment from reply posts Po2 and Po33 sent back to the outgoing post Po1 sent by the poster “X1”. . In the example of FIG. 8, the extraction device 100 extracts the reply post Po2 because the comment “Reply post Po2 does not get on such a sale” includes the character “煽”. Further, the extraction device 100 extracts the reply post Po33 because the comment “I want to buy it a little. I wonder if it is selling” in the comment of the reply post Po33.

  Subsequently, the extraction device 100 is a reply post indicating the reliability of the outgoing post as a reply post pointing out the reliability of the outgoing post. Reply posts whose difference from the degree of thought is equal to or greater than a predetermined threshold are extracted. The predetermined threshold is set to an appropriate value according to, for example, a target value of the number of reply posts to be extracted as reply posts indicating the reliability of outgoing posts. As an example, a higher value is set for the predetermined threshold as the number of reply posts to be extracted is reduced. On the other hand, a lower value is set for the predetermined threshold as the number of reply posts to be extracted is increased. In the example of FIG. 8, the extraction apparatus 100 extracts a reply post in which the difference between the emotion score of the reply post and the emotion score of the outgoing post is “4” or more as a predetermined threshold. Here, the difference between the emotion score “2” of the emotion Em2 “I want to buy strongly” of the reply post Po2 and the emotion score “−2” of the emotion Em1 “I want to sell strongly” of the outgoing post Po1 is “4”. . Therefore, the extraction apparatus 100 extracts the reply post Po2 as a reply post indicating the reliability of the outgoing post Po1.

  On the other hand, the difference between the emotion score “1” of the emotion Em33 “I want to buy” of the reply post Po33 and the emotion score “−2” of the emotion Em1 “I want to sell strongly” of the outgoing post Po1 is “3”. Therefore, the extraction device 100 does not extract the reply post Po33 as a reply post indicating the reliability of the outgoing post Po1.

  As described above, the extraction device 100 according to the modified example is a reply post that points out the reliability of the outgoing post, and a reply post that satisfies a predetermined condition in which the difference between the degree of the thought of the reply post and the degree of the thought of the outgoing post is satisfied. Extract. For example, the extraction apparatus 100 is a reply post in which thought-related information indicates a thought opposite to the thought of the outgoing post as a reply post pointing out the reliability of the outgoing post. Reply posts whose difference from the degree of is greater than or equal to a predetermined threshold are extracted.

  Thereby, the extraction apparatus 100 can extract the reply posts in which the difference between the degree of the thought of the reply posts and the degree of the thought of the outgoing posts is equal to or greater than a predetermined threshold. Can be adjusted. For example, the extraction device 100 can improve the quality of reply posts that are extracted as reply posts that indicate the reliability of outgoing posts by setting a high predetermined threshold. For this reason, since the extraction apparatus 100 can provide information related to a high-quality reply posting, the quality of the posting service can be improved. Further, the extraction apparatus 100 can increase the number of reply posts to be extracted as reply posts indicating the reliability of outgoing posts by setting the predetermined threshold value low. Therefore, the extraction device 100 can provide an amount of information necessary for big data analysis and the like.

[2-2. (Extract reply posts that affirm outgoing posts)
In the above embodiment, the outgoing posting is a hit post based on the thought-related information indicating the thought of the poster who posted the reply post extracted by the first extraction unit 132 and the degree of the thought. An example of extracting a reply post that points out was explained. Here, the extraction device 100 may extract not only a reply post that points out that the outgoing post is a hit post but also a reply post that makes various indications such as an indication that gives up the outgoing post.

  For example, the first extraction unit 132 of the extraction apparatus 100 extracts a reply post that includes a character that affirms an outgoing post as a reply post that includes a predetermined character. The second extraction unit 133 of the extraction device 100 extracts a reply post in which the thought-related information of the reply post indicates a thought in the same direction as the thought of the outgoing post as a reply post indicating the reliability of the outgoing post. This point will be described with reference to FIG.

  FIG. 9 is an explanatory diagram for explaining the flow of extraction processing by the extraction apparatus 100 according to the modification. In the example of FIG. 9, the extraction apparatus 100 holds the display names X1 and Y1 of the users in association with the profile images P1 and P2 as information related to the users who use the posting service. Further, the extraction device 100 holds information regarding the outgoing posting Po41 and the reply posting Po42 to the outgoing posting Po41. Specifically, the extraction device 100 uses the user ID of the poster with the display name “X1”, the title Ti41 “at the time of purchase”, and the comment Ar41 “Company A's current stock price is too low as information regarding the outgoing posting Po41. "It's time to buy!", Posting time Tm41 “January 24, 14:02” of outgoing post Po41, and poster's emotion Em41 “I want to buy strongly” in association with post ID “411” To do. In addition, the extraction device 100 uses the user ID of the contributor with the display name “Y1”, the title Ti42 “Re: at the time of purchase”, and the comment Ar42 “soaring as information regarding the reply post Po42 to the outgoing post Po41. “Thank you for buying! Thank you for your purchase!”, Posting time Tm42 of reply post Po42 “January 24, 15:04” and contributor's emotion Em42 “I want to buy strongly” with post ID “ 412 ".

  First, the extraction device 100 extracts a reply post that includes characters that affirm the outgoing post. For example, the extraction apparatus 100 extracts a reply post including “agreement” as a character to affirm the outgoing post. In the example of FIG. 9, the extraction device 100 includes the text “Agree” in the comment “Reply post Po42. You ’ve agreed. Thank you for your purchase!” The reply post Po42 is extracted.

  Subsequently, the extraction apparatus 100 extracts a reply post in which the thought-related information of the reply post indicates a thought in the same direction as the thought of the outgoing post as a reply post indicating the reliability of the outgoing post. In the example of FIG. 9, the feeling Em42 “I want to buy strongly” of the reply post Po42 is in the same direction as the emotion Em41 “I want to buy strongly” of the outgoing post Po41. Therefore, the extraction device 100 extracts the reply post Po42 as a reply post that points out the reliability of the outgoing post Po41.

  As described above, the extraction device 100 according to the modified example extracts a reply post including a character that affirms an outgoing post as a reply post including a predetermined character. Further, the extraction apparatus 100 extracts a reply post in which the thought-related information of the reply post indicates a thought in the same direction as the thought of the outgoing post as a reply post indicating the reliability of the outgoing post.

  Thereby, since the extraction apparatus 100 can extract the reply post that agrees with the outgoing post, it is possible to provide a highly reliable combination of the outgoing post and the reply post. For example, the extraction apparatus 100 can extract a reply post that gives up outgoing postings, and thus can provide information on a contributor who made a high-quality post. For this reason, the extraction apparatus 100 can improve the quality of the posting service.

  In addition, the extraction device 100 according to the modified example is a reply post indicating the reliability of the outgoing post as a reply post in which the thought related information is a reply post indicating the same direction as the outgoing post idea, Reply posts whose difference from the degree of thought of outgoing posts is less than a predetermined threshold may be extracted. For example, the extraction apparatus 100 extracts a reply post in which the difference between the degree of the thought of reply posting and the degree of the idea of outgoing posting is “1” or less. As an example, it is assumed that the feeling of reply posting is “I want to buy strongly” and the feeling of outgoing posting is “I want to buy”. In this case, the extraction apparatus 100 has the same feeling as the sentiment sentiment “I want to buy” in the same direction as the sentiment sentiment “I want to buy”, and the sentiment sentiment score “2” and the sentiment sentiment score Since the difference from “1” is “1” which is equal to or less than “1”, the reply post is extracted.

  As a result, the extraction apparatus 100 is a reply post in the same direction as the idea of the outgoing post, and a reply post in which the difference between the degree of the post of the reply post and the degree of the outgoing post is equal to or less than a predetermined threshold value. Since it can be extracted, the number and quality of reply posts to be extracted can be adjusted. For example, the extraction device 100 can improve the quality of reply posts that are extracted as reply posts that indicate the reliability of outgoing posts by setting a predetermined threshold value low. For this reason, since the extraction apparatus 100 can provide information related to a high-quality reply posting, the quality of the posting service can be improved. Further, the extraction apparatus 100 can increase the number of reply posts to be extracted as reply posts indicating the reliability of outgoing posts by setting the predetermined threshold value high. Therefore, the extraction device 100 can provide an amount of information necessary for big data analysis and the like.

[2-3. (Applicable)
In the above embodiment, an example has been described in which the extraction apparatus 100 extracts a reply post indicating the reliability of an outgoing posting for a stock price bulletin board of a posting service. Here, the extraction device 100 is not limited to the stock price bulletin board, and may target various services.

  For example, the extraction device 100 may extract a reply post that points out the reliability of an outgoing post for an election bulletin board that is related to an election. As an example, it is assumed that an outgoing posting with a comment in favor of My Number is posted on the election bulletin board. It is assumed that the thought related information indicating the thought of the outgoing posting is “supporting political party A”. Further, it is assumed that a reply post in which a comment against my number is written is posted on the election bulletin board as a reply post to the outgoing post. It is assumed that the thought related information indicating the thought of such a reply posting is “supporting political party B”. Here, “Political Party A” and “Political Party B” are political parties having opposite thoughts. In this case, the extraction device 100 extracts a reply post in which a comment that opposes My Number is described as a reply post indicating the reliability of the outgoing post.

  In another example, the extraction apparatus 100 may extract a reply post that points out the reliability of an outgoing post for a sports bulletin board that is about sports. As an example, it is assumed that an outgoing posting in which a comment predicting “Team G” is posted on the sports bulletin board as a team winning this year is posted. It is assumed that the thought related information indicating the thought of the outgoing posting is “support team G”. In addition, it is assumed that a reply post in which a comment predicting “Team H” as a team winning this year is written as a reply post to the outgoing posting on the sports bulletin board. It is assumed that the thought related information indicating the thought of the reply posting is “support team H”. Here, it is assumed that the thought related information “support team G” and “support team H” are thoughts in opposite directions. In this case, the extraction apparatus 100 extracts a reply post in which a comment that predicts “team H” is written as a team whose support team is H and wins this year as a reply post indicating the reliability of the outgoing posting.

  In another example, the extraction apparatus 100 may extract a reply post that points out the reliability of an outgoing post for an idle bulletin board whose theme is related to an idol. As an example, it is assumed that an outgoing posting in which a comment that predicts “idol R” is written as an idol to break next year is posted on the idol bulletin board. It is assumed that the thought related information indicating the thought of the outgoing posting is “supporting idol R”. Further, it is assumed that a reply post in which a comment predicting “idol K” as an idol to be broken next year is posted as a reply post to the outgoing post on the idle bulletin board. It is assumed that the thought related information indicating the thought of the reply posting is “supporting idol K”. Here, it is assumed that the thought related information “supporting idol R” and “supporting idol K” are thoughts in opposite directions. In this case, the extraction apparatus 100 extracts a reply post in which a comment that predicts “idol K” is described as a idol who is supported by the K and breaks next year as a reply post that points out the reliability of the outgoing post.

  As described above, the extraction apparatus 100 according to the modification extracts reply posts indicating the reliability of outgoing posts for various posting services. As a result, the extraction apparatus 100 can extract reply posts indicating the reliability of outgoing posts regardless of the theme handled by the posting service, for example, so that the quality of various types of posting services can be improved.

[2-4. (Extract reply posts based on user information)
In the above-described embodiment, an example in which the extraction apparatus 100 extracts a reply post indicating the reliability of an outgoing post has been described. Here, the extraction apparatus 100 may extract a reply post that points out the reliability of the outgoing post based on user information indicating the characteristics of the poster.

  Specifically, the extraction apparatus 100 extracts a reply post indicating the reliability of the outgoing post based on the support level indicating the level of support from other users in the user information of the poster. For example, the degree of support is the number of times other users have pressed the “I think so” button for a post made by the poster in the past. As an example, the extraction device 100 preferentially extracts a reply post posted by a poster who has a high degree of support as a reply post indicating the reliability of the outgoing post.

  As described above, the extraction apparatus 100 according to the modification extracts a reply post indicating the reliability of the outgoing post based on the user information indicating the characteristics of the poster. Thereby, the extraction apparatus 100 can extract a reply post posted by a highly credible contributor, and thus can provide a high-quality reply post.

[2-5. (Retrieval posts are extracted considering actual stock price)
In the above-described embodiment, an example in which the extraction device 100 extracts a reply post indicating that the outgoing posting is a hit post for the stock price bulletin board has been described. Here, the extraction apparatus 100 may extract a reply post indicating the reliability of the outgoing post in consideration of the actual stock price.

  Specifically, the extraction apparatus 100 extracts a reply post indicating the reliability of the outgoing post based on the degree of coincidence between the actual stock price fluctuation and the poster's idea. For example, when the actual stock price is soaring, the extraction apparatus 100 preferentially extracts reply posts whose contributor's idea is “I want to buy strongly”.

  Thus, the extraction apparatus 100 according to the modification extracts a reply post that points out the reliability of the outgoing post in consideration of the actual stock price. Thereby, since the extraction apparatus 100 can extract a highly reliable reply posting indicating the reliability of the outgoing posting with high accuracy, it can provide a high-quality reply posting.

[2-6. (Retrieval posts are extracted by machine learning)
In the above embodiment, the extraction device 100 points out the reliability of the outgoing post based on the thought-related information and the degree of the thought indicating the thought of the poster who posted the reply post extracted by the first extraction unit 132. An example of extracting reply posts was given. Here, the extraction apparatus 100 may extract a reply post indicating the reliability of the outgoing post by machine learning.

  For example, the extraction apparatus 100 sets the reply post extracted by the second extraction unit 133 as a correct answer, performs machine learning using, for example, a language expression such as a maximum partial character string as a feature, and indicates a reply indicating the reliability of the outgoing post Extract posts. As an example, the extraction apparatus 100 performs binary document classification by SVM (Support Vector Machine) using the maximum partial character string for a post title and a comment as a feature.

  As described above, the extraction apparatus 100 according to the modification extracts a reply post indicating the reliability of the outgoing post by machine learning. Thereby, since the extraction apparatus 100 can extract the reply post indicating the reliability of the outgoing posting with high accuracy, it can provide a high-quality reply post.

[2-7. Effect of modification)
As described above, in the extraction apparatus 100 according to the modification, the second extraction unit 133 determines the difference between the degree of the thought of the reply post and the degree of the thought of the outgoing post as a reply post indicating the reliability of the outgoing post. Extracts reply posts that satisfy a predetermined condition. For example, the second extraction unit 133 is a reply post indicating the reliability of the outgoing post as a reply post pointing out the reliability of the outgoing post, and the thought-related information is a reply post showing the idea in the opposite direction to the outgoing post idea. Reply posts whose difference from the degree of the idea is equal to or greater than a predetermined threshold are extracted.

  Thereby, the extraction apparatus 100 according to the modification can extract a reply post in which the difference between the degree of the idea of the reply post and the degree of the idea of the outgoing post is equal to or more than a predetermined threshold. The number and quality can be adjusted. For example, the extraction device 100 can improve the quality of reply posts that are extracted as reply posts that indicate the reliability of outgoing posts by setting a high predetermined threshold. For this reason, since the extraction apparatus 100 can provide information related to a high-quality reply posting, the quality of the posting service can be improved. Further, the extraction apparatus 100 can increase the number of reply posts to be extracted as reply posts indicating the reliability of outgoing posts by setting the predetermined threshold value low. Therefore, the extraction device 100 can provide an amount of information necessary for big data analysis and the like.

  In addition, in the extraction device 100 according to the modification, the first extraction unit 132 extracts a reply post including a character that affirms an outgoing post as a reply post including a predetermined character. The second extraction unit 133 extracts a reply post in which the thought-related information of the reply post indicates a thought in the same direction as the thought of the outgoing post as a reply post indicating the reliability of the outgoing post. For example, the second extraction unit 133 is a reply post indicating the reliability of the outgoing post as a reply post pointing out the reliability of the outgoing post. Reply posts whose difference from the degree of the idea is below a predetermined threshold are extracted.

  Thereby, since the extraction apparatus 100 according to the modification can extract a reply post that agrees to the outgoing post, it is possible to provide a highly reliable combination of the outgoing post and the reply post. For example, the extraction apparatus 100 can extract a reply post that gives up outgoing postings, and thus can provide information on a contributor who made a high-quality post. For this reason, the extraction apparatus 100 can improve the quality of the posting service.

[3. Others]
In addition, among the processes described in the above embodiment, all or part of the processes described as being automatically performed can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known extraction method. In addition, the processing procedures, specific names, and information including various data and parameters shown in the document and drawings can be arbitrarily changed unless otherwise specified.

  Further, each component of each illustrated apparatus is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each extraction device is not limited to the one shown in the figure, and all or a part thereof is functionally or physically distributed in arbitrary units according to various loads or usage conditions. -Can be integrated and configured.

  For example, the posting information storage unit 122 illustrated in FIG. 2 may be stored in a storage server or the like without the extraction device 100. In this case, the extraction apparatus 100 acquires information related to posting by accessing the storage server.

  Further, the extraction apparatus 100 may be an extraction apparatus that performs only the extraction process without performing the providing process. In this case, the extraction device does not have the providing unit 134. Then, the providing device having the providing unit 134 provides information related to the reply posting extracted by the extracting device 100.

  Further, the extraction apparatus 100 according to the embodiment described above is realized by a computer 1000 having a configuration as shown in FIG. 10, for example. Hereinafter, the extraction apparatus 100 will be described as an example. FIG. 10 is a hardware configuration diagram illustrating an example of a computer 1000 that implements the functions of the extraction device 100. The computer 1000 includes a CPU 1100, RAM 1200, ROM 1300, HDD 1400, communication interface (I / F) 1500, input / output interface (I / F) 1600, and media interface (I / F) 1700.

  The CPU 1100 operates based on an extraction program stored in the ROM 1300 or the HDD 1400 and controls each unit. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is activated, an extraction program depending on the hardware of the computer 1000, and the like.

  The HDD 1400 stores an extraction program executed by the CPU 1100, data used by the program, and the like. The communication interface 1500 receives data from other devices via the communication network 50 and sends the data to the CPU 1100, and transmits the data generated by the CPU 1100 to other devices via the communication network 50.

  The CPU 1100 controls an output device such as a display and a printer and an input device such as a keyboard and a mouse via the input / output interface 1600. The CPU 1100 acquires data from the input device via the input / output interface 1600. Further, the CPU 1100 outputs the generated data to the output device via the input / output interface 1600.

  The media interface 1700 reads the extraction program or data stored in the recording medium 1800 and provides it to the CPU 1100 via the RAM 1200. The CPU 1100 loads the extraction program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable disk), a magneto-optical recording medium such as an MO (Magneto Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. It is.

  For example, when the computer 1000 functions as the extraction device 100 according to the embodiment, the CPU 1100 of the computer 1000 implements the function of the control unit 130 by executing an extraction program loaded on the RAM 1200. The HDD 1400 stores data in the storage unit 120. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them. However, as another example, these programs may be acquired from another extraction device via the communication network 50.

  As described above, some of the embodiments of the present application have been described in detail with reference to the drawings. However, these are merely examples, and various modifications, including the aspects described in the disclosure section of the invention, based on the knowledge of those skilled in the art, It is possible to implement the present invention in other forms with improvements.

  Further, the above-described extraction device 100 may be realized by a plurality of server computers, and depending on functions, an external platform or the like may be realized by calling an API (Application Programming Interface), network computing, or the like. Can be changed flexibly.

  Further, “section (module, unit)” described in the claims can be read as “means”, “circuit”, and the like. For example, the receiving unit can be read as receiving means or a receiving circuit.

DESCRIPTION OF SYMBOLS 100 Extraction device 120 Storage part 121 User information storage part 122 Posting information storage part 123 Pointed information storage part 131 Receiving part 132 1st extraction part 133 2nd extraction part 134 Provision part

Claims (11)

A first extraction unit for extracting a reply post including a predetermined character from reply posts replied to a send post sent by a poster;
A second extraction for extracting a reply post indicating the reliability of the outgoing post based on thought related information indicating the thought of the poster who posted the reply post extracted by the first extraction unit and the degree of the thought And an extraction device. The second extraction unit includes
The reply post that satisfies a predetermined condition in which a difference between the degree of the thought of the reply post and the degree of the thought of the outgoing post is extracted as a reply post pointing out the reliability of the outgoing post. The extraction device described. The second extraction unit includes
The reply post in which the thought related information of the reply post indicates a thought opposite to the thought of the outgoing post is extracted as a reply post pointing out the reliability of the outgoing post. Extraction device. The second extraction unit includes
4. A reply post in which the difference between the degree of the thought of the reply post and the degree of the thought of the outgoing post is a predetermined threshold or more is extracted as a reply post pointing out the reliability of the outgoing post. The extraction device described in 1. The first extraction unit includes:
The extraction device according to claim 3 or 4, wherein a reply post including a character that denies the outgoing post is extracted as a reply post including the predetermined character. The second extraction unit includes
The reply post indicating the thought in the same direction as the thought of the outgoing post is extracted as a reply post pointing out the reliability of the outgoing post. The extraction device according to one. The second extraction unit includes
The reply post in which a difference between the degree of the thought of the reply post and the degree of the thought of the outgoing post is equal to or less than a predetermined threshold is extracted as a reply post pointing out the reliability of the outgoing post. The extraction device described in 1. The first extraction unit includes:
The extraction apparatus according to claim 6 or 7, wherein a reply post including a character that affirms the outgoing post is extracted as a reply post including the predetermined character. The first extraction unit includes:
The extraction device according to any one of claims 1 to 8, wherein a reply post including a predetermined character is extracted from reply posts viewed by an unspecified number of users. An extraction method executed by an extraction device,
A first extraction step of extracting a reply post including a predetermined character from reply posts replied to the outgoing post sent by the poster;
A second extraction for extracting a reply post indicating the reliability of the outgoing post based on thought related information indicating the thought of the poster who posted the reply post extracted in the first extraction step and the degree of the thought The extraction method characterized by including the process. A first extraction procedure for extracting a reply post including a predetermined character from reply posts replied to a send post sent by a contributor;
A second extraction for extracting a reply post indicating the reliability of the outgoing post based on thought-related information indicating the thought of the poster who posted the reply post extracted by the first extraction procedure and the degree of the thought An extraction program that causes a computer to execute the procedure.

Images