KR101733911B1 - Module for analyzing of subscriber's tendency by uploaded contents to social network - Google Patents

Module for analyzing of subscriber's tendency by uploaded contents to social network Download PDF

Info

Publication number
KR101733911B1
KR101733911B1 KR1020160016083A KR20160016083A KR101733911B1 KR 101733911 B1 KR101733911 B1 KR 101733911B1 KR 1020160016083 A KR1020160016083 A KR 1020160016083A KR 20160016083 A KR20160016083 A KR 20160016083A KR 101733911 B1 KR101733911 B1 KR 101733911B1
Authority
KR
South Korea
Prior art keywords
content
vocabulary
subscribers
unit
keyword
Prior art date
Application number
KR1020160016083A
Other languages
Korean (ko)
Inventor
이경순
정광용
조승현
Original Assignee
전북대학교산학협력단
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 전북대학교산학협력단 filed Critical 전북대학교산학협력단
Priority to KR1020160016083A priority Critical patent/KR101733911B1/en
Application granted granted Critical
Publication of KR101733911B1 publication Critical patent/KR101733911B1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/01—Social networking
    • G06F17/27—
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/30—

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Human Resources & Organizations (AREA)
  • General Health & Medical Sciences (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Computing Systems (AREA)
  • Operations Research (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Game Theory and Decision Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The present invention relates to a method and apparatus for extracting a main forwarder among a user's subscribers, Analyze the overall response of subscribers to classify subscribers 'propensity as supportive / non-supportive / neutral, and compare the uploaded content with the vocabulary emotional database to analyze subscribers' opinions on the corresponding content as affirmative / negative / neutral A subscribers tendency analysis module. The present invention extracts a main content through a main messenger and thus can identify a content having a high influence on the user's content and can classify the tendency of the subscribers into support / non-support / neutral through the reaction of the main messenger . Also, by grasping the subscribers' opinions on the contents as affirmative / negative / neutral, it is possible to grasp only the opinions of the specific contents as affirmative / negative / neutral without regard to support or non-supportive tendency. This has the effect of enabling subscribers to exercise marketing influence and social influence on content.

Description

{MODULE FOR ANALYZING OF SUBSCRIBER'S TENDENCY BY UPLOADED CONTENTS TO SOCIAL NETWORK}

The present invention relates to a module for analyzing a tendency of a subscriber according to a subject of contents uploaded to a social network. More particularly, the present invention relates to a module for extracting a main sender among subscribers of a user, Analyze the overall response of subscribers to classify subscribers 'propensity as supportive / non-supportive / neutral, and compare the uploaded content with the vocabulary emotional database to analyze subscribers' opinions on the corresponding content as affirmative / negative / neutral A subscribers tendency analysis module.

As the Internet technology develops, the field of mutual communication on the Internet network is gradually widening. Social network services such as facebook and twitter are popularly used, and they express their thoughts and information in the social network service area. Therefore, the contents uploaded in the social network reflects recent social issues and trends in real time. By analyzing this, it is possible to predict the future, not only to find the optimum countermeasures but also to link the information to the profit. In recent years, as the number of Twitter users has increased, the interaction between users has become very large, and the influence of users with a large number of followers has increased dramatically. As the use of social network services is rapidly increasing, networking on the Internet has created a need for systematic research on their social and cultural influence.

Korean Patent No. 10-1336293 (hereinafter referred to as " precedent document ") describes the intimacy of users who perform network activities related to an individual on a social network and expresses the intimacy of the users in a network form so that the relationship between users Based personal network extraction system for allowing a user to easily identify users with high information and intimacy. Preceding documents have the effect of knowing users who are positive to themselves by quantifying the intimacy of users within the social network.

However, the precedent literature can only grasp the intimacy of followers in the social network, and can not grasp the proponents' tendencies such as support / non-support / neutrality. In addition, it is necessary to grasp the opinions of the followers of the contents with the structure in which each opinion of the followers of the contents can not be grasped. For example, a supportive proponent may have a negative opinion on a particular content, and even a non-proponent follower may have a positive opinion on a particular content. Therefore, there is a need for a technology that allows followers to have a certain opinion on specific content (specific social issues), what information they want and need, and what purpose they follow.

Korean Patent No. 10-1336293 (Intimacy-based personal network extraction system of social network, registration date: November 25, 2013)

In order to solve such a problem, the present invention has an object of extracting a main deliverer among subscribers who subscribe to contents of a user using a social network, and extracting main contents through a main deliverer.

In addition, the present invention aims at clustering user's content and comparing clustering contents with a vocabulary emotion database to identify the tendency of the subscribers to the content as positive / negative / neutral.

The subscribers' propensity analysis module according to the present invention includes a data extracting unit for extracting content response information for a content uploaded by the user of at least one subscriber subscribing to a content of a user using a social network, A subscriber analysis unit for analyzing the subscriber propensity of the user by comparing the subscriber's preference with a vocabulary emotion database, and a subscriber analysis unit for analyzing the subscribers' And a sender extracting unit for extracting a main sender among the plurality of senders.

The vocabulary emotion database according to the present invention includes an emotional data storage unit for storing basic emotional vocabularies for positive and negative formed in predetermined syllable units and storing the classified emotional vocabularies, A vocabulary identification unit for identifying a specific vocabulary corresponding to the cut syllable by comparing the extracted vocabulary with the basic vocabulary word; And a vocabulary adding unit for storing the vocabulary.

The subscriber's propensity analyzing module according to the present invention includes a content extracting unit for extracting a valid content from the content uploaded by the user by removing the content without the content response information, A keyword recognition unit for recognizing a main keyword referred to another user by a predetermined ratio or more of the valid keywords by the main communicator; A content classifying unit for setting a content as a central content and classifying the effective content by a section by applying the number of effective keywords of the central content to a predetermined clustering method, Extracts each extracted keyword reaction information, Compared with the emotional vocabulary database includes more keyword analysis unit that analyzes the subscriber nature of the valid keywords.

The present invention extracts a main forwarder among subscribers who subscribe to contents of a user using a social network and extracts main contents through a main forwarder to identify contents having high influence among the contents of the user, Analysis of subscribers 'responses to the content of the main forwarder has the effect of classifying the subscribers' tendency as support / non support / neutral.

 In addition, the present invention clusters user's content, compares the clustering content with the vocabulary emotion database, and grasps the subscribers' opinions on the content as positive / negative / neutral, Only opinions can be identified as positive / negative / neutral. This has the effect of enabling subscribers to exercise marketing influence and social influence on content.

1 is a block diagram of a tendency analysis module according to the present invention.
2 is a diagram for explaining the addition of emotional vocabulary according to the present invention.
FIG. 3 is an exemplary diagram for explaining a main messenger extracting method according to the present invention.
4 is an exemplary diagram for explaining a main keyword extraction method according to the present invention.
FIG. 5 is an exemplary diagram showing a tendency of subscribers according to the effective keyword according to the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail.

1 is a block diagram of a tendency analysis module according to the present invention.

1, the propensity analysis module includes a data extraction unit 100, a subscriber analysis unit 200, a vocabulary evaluation database 300, a sender extraction unit 400, a content extraction unit 500, a keyword extraction unit 600 A keyword recognition unit 700, a content classification unit 800, and a keyword analysis unit 900.

The data extraction unit 100 extracts content response information of at least one subscriber who subscribes to a user's content using a social network. The content response information includes the subscriber's comments, transmissions, comments, and comments on the content uploaded by the user. Here, the content uploaded by the user includes a text, an image, an image, and the like. For example, if a user A uploads information about a particular entertainer to a social network, subscribers who subscribe to A may reply to the uploaded content (information about a particular entertainer) And creates reaction information such as retweet. The data extracting unit 100 extracts information about subscribers' ripple, delivery, and reference to the contents (information on a specific entertainer). On the other hand, a subscription can be the same as 'follow' used in Twitter, and 'subscribers' in 'follower'.

[Analysis of subscribers' tendency]

The subscriber analysis unit 200 compares the extracted content response information with the vocabulary emotion database 300 to analyze a user's subscriber propensity. The vocabulary emotion database 300 stores vocabulary expressing emotions for positive and negative persons. For example, if user A has subscribers B, C, and D, and subscriber B submits the content of A to other users and leaves affirmative replies, they are classified as positive subscribers. On the other hand, if C subscribers leave a negative reflections on B's content replies, classify B subscribers as subscribers. Also, D subscribers are classified as neutral subscribers if they do not give special rephrases to A content, or if they do not mention and deliver them to other users.

Meanwhile, as a new term is created and a term not registered in advance is included in the content reaction information, the vocabulary emotion database 300 is provided with an emotion data storage unit 310 for analyzing spontaneous spelling errors and coined words, 320, a syllable extractor 330, and a lexical adder 340.

[Adding Emotion Vocabulary]

The addition of the emotional vocabulary according to the present invention will be described with reference to FIG. In the emotion data storage unit 310, the basic emotion vocabulary for positive and negative formed in a predetermined syllable unit is classified and stored. E.g. Vocabulary such as 'good, thank you, longing' is stored as a positive vocabulary, and vocabulary such as 'ephemeral, truthful, irritable' is classified and stored as a negative vocabulary. Here, the predetermined syllable unit is a matter which can be changed by the user such as two syllables and three syllables.

The syllable extracting unit 320 extracts at least one cut syllable by cutting a specific word input from the outside into predetermined syllable units. For example, if a particular word entered is' liked, hateful deb. The syllable extracting unit 320 extracts 'good', 'good', 'bad', and 'deb', respectively. We extract 'dog breeding' as 'petty dogs', 'petty dogs', 'dog irritation' as 'fake and irritated' and 'thanksgiving' as 'thanks and use'.

The vocabulary identification unit 330 identifies a specific emotional vocabulary corresponding to the cut syllable by comparing the extracted cut syllable with the basic emotional vocabulary. For example, the cut syllable of 'good' is 'good' and the 'good' is a vocabulary included in the positive vocabulary of the sentiment data storage unit 310. Therefore, the vocabulary identification unit 300 identifies a positive vocabulary of 'good' by comparing the vocabulary stored in the vocabulary storage unit 31 with the cut syllable 'good' of 'good'.

The vocabulary addition unit 340 classifies a specific word into a specific emotional vocabulary identified and stores the vocabulary. Referring to FIG. 2, it can be confirmed that the emotional vocabulary of the good is identified as' good ', and the emotional vocabulary of the dog petal is identified as' Therefore, the vocabulary addition unit 340 classifies and stores the favorable words as affirmative vocabulary, and stores the vocabulary words as the negative vocabulary.

[Extracting key forwarders]

Referring to FIG. 1, the sender extractor 400 extracts a main sender among the subscribers according to the order of delivery of the subscribers that have transferred the contents of the user to other users in the social network, and the mutual subscription relationship among the subscribers. 3, a description will be given of a method of extracting a main sender according to the present invention. Referring to the mutual subscription relationship in FIG. 3, B and C subscribe (follow) users A and C subscribe (follow) B users. Thus, A's subscribers (followers) become B and C, and B's subscribers (followers) become C. Referring to the first embodiment of the content delivery procedure of FIG. 3, B first delivers the content of A to another user, and then C delivers the content of A to another user. If B delivers the content of A first as in the first embodiment, C that subscribes to (A) and B (B) subscribes to the content of A through B and the content of A from B repeatedly. Therefore, B receives a content from A, and a weight of 1 is applied. C receives a content from A and B, and a weight of 0.5 is applied.

Meanwhile, referring to a second embodiment of the content delivery order of FIG. 3, C first delivers the content of A to another user, and then B delivers the content of A to another user. In the case of the second embodiment, C and B receive contents of A independently. Therefore, the weights of C and B are applied equally. On the other hand, the weights 1 and 0.5 are numbers arbitrarily used to denote the ratio, and do not include the meaning of numbers.

The main message extracting method according to the weight can be created by the following equation.

[Key deliverer extraction expression]

Figure 112016013942370-pat00001
Figure 112016013942370-pat00002

w ij : weight of e ij

e ij is the number of users i subscribes to (the top of the list) when i transfers the contents of j to the top of the j

[Key keyword recognition through key forwarders]

Referring to FIG. 1, the content extracting unit 500 extracts the valid content from the content uploaded by the user by removing the content without the content response information. For example, if the user uploads content for a specific artist, content for a specific actor, and content for a particular politician within the social network, and if there is content response information for a particular singer and a particular politician, Removes the content for a specific actor, and extracts the content for the specific singer and the content for the specific politician as the effective content. On the other hand, it refers to excluding content classification for extracting an effective content, rather than deleting content data for a specific singer.

The keyword extracting unit 600 analyzes the text of the effective content to extract a valid keyword having a specific meaning. The keyword extracting unit 600 analyzes the type of Korean and removes a stop word. For example, if the text of the effective content is 'Korea is one of the Asian countries', the keywords containing specific meanings such as Korean and Asian countries are extracted through morphological analysis, and other negatives are deleted. On the other hand, the valid keyword can be set by the user. The above example assumes that the valid keyword is set to two, and in all the examples below, the case where there are two valid keywords is assumed.

The keyword recognition unit 700 recognizes a main keyword referred to another user by a main deliverer at a predetermined ratio or more of the valid keywords. A description will now be given of a method for recognizing a main keyword by a main messenger through FIG.

Referring to FIG. 4, it can be seen that the effective contents are from No.1 to No.18. Here, the valid content number is the list information of the valid content extracted through the content extracting unit 500. Referring to FIG. 4, it can be confirmed that the effective keyword for each effective content is extracted by the keyword extracting unit 600. The effective keywords of effective content No. 12 are sympathizers and weekends, and the effective keywords of effective content No. 16 are the prosecution and the blue house. The number of main forwarders mentioning the effective keyword of effective content No. 12 is 03/20, and the number of main forwarders referring to the effective keyword of effective content No. 16 is 11/20. If the ratio of the number of the main deliverers mentioned is 50%, the keyword recognition unit 700 recognizes the effective content No. 16 (11/20 persons, 55%) as the main content. Accordingly, in FIG. 4, the effective keywords No. 5, 11, 15, 16, 17, and 18 having the number of main deliverers of 50% or more are recognized as the main keywords.

The main keyword extraction method by the main messenger can be created by the following formula.

[Keyword extraction expression]

Figure 112016013942370-pat00003

t: words, d: content, N: the total number of users tweet, IM t: number of users referred to your word + t of the leading forwarders, Im all : The total number of users + main forwarders, extracting a value greater than the threshold value as the main content (0.05 <θ). Two words with the largest value of w (t) are extracted, and two words are the main keywords. On the other hand, the main keyword can be changed by the user.

[Clustering contents]

Referring to FIG. 1, the content classifier 800 sets a content having the highest number of effective contents delivered to other users as a central content. In addition, the number of effective keywords of the central content is applied to the predetermined clustering method to classify the effective contents of the user into sections.

A plurality of contents are clustered (clustering) in order to classify contents. In the clustering method, the content classifying unit 800 determines the number of K clusters (clusters). A content having the highest number of delivered (retrieved) to other users among the valid content of the user is set as the central content, and a center value is designated as the number of effective keywords of the central content. For each of the contents (d), the distance to the K center values (c) is obtained and belongs to the closest community. Here, the distance on each of the contents and K center values is

Figure 112016013942370-pat00004
(i = 1, 2, .., n, j = 1, 2, .., K, n is the total number of documents, and k is the number of center values). Through the contents belonging to each cluster, an average from the content to the center value is obtained, and the center value of each cluster is newly designated. Repeat until the newly specified center value is equal to the previous center value and the same time classification is completed. The effective content classified in this way becomes one section. For example, if the classified section (effective content) is 'country' or 'person', then the uploaded content for the 'a' country and the 'i' person is composed of one section.

[Effective keyword tendency analysis]

Referring to FIG. 1, the keyword analyzing unit 900 extracts keyword response information of valid keywords corresponding to effective contents classified by sections. Also, the extracted keyword reaction information is compared with the vocabulary emotion database 300 to analyze the subscriber propensity of the effective keyword. Referring to FIG. 5, the effective contents of user A are clustered into effective keywords of [Prosecutor, Blue House], [Extreme Wandering Wave, North Korea], [Human City Academy, Lecture], [Empathy, Weekend], [Election, have. The keyword analyzing unit 900 analyzes the tendency of the subscribers (positive, negative, neutral, and non-responding) for each effective keyword. Referring to FIG. 5, the keyword analyzing unit 900 extracts keyword reaction information of the grouped effective keywords [Prosecutor, Blue House] of the user A. The keyword response information includes the subscriber's comment, transmission, comment and comment on the effective content uploaded by the user. In the case of FIG. 4, 50/100 of the subscribers are positive, 15/100 are negative, 20/100 are neutral, and 15/100 are non-responding to effective content including valid keywords of the prosecution and the Blue House. The positive and negative tendencies are analyzed through comparison with the vocabulary emotion database 300. For example, a subscriber D would like to thank the prosecution and the Blue House for the effective content corresponding to [the prosecution, the Blue House]. " If you leave a ripple, you extract the keyword 'thanks' and classify it as a positive tendency. Also, when a subscriber D sends effective content corresponding to [Prosecutor, Blue House] of A within the social network to another user, it is classified as positive. Neutral cases are cases where there is no positive or negative vocabulary. On the other hand, non-response may be included as neutral.

By thus classifying the tendency of the subscribers for the effective keyword through the keyword analyzing unit 900, the user can confirm the section (area) of interest by the subscribers. In the case of FIG. 5, it can be confirmed that the non-response rate of the effective keyword A [human city academy, lecture] is 55%, which is an area of interest of subscribers. On the other hand, social and political fields such as the prosecution, the election, and North Korea are areas of concern. B's effective rate of [food, fast-paced] is 60%, which is the area outside of the subscriber's interest, and IT fields such as smartphone and communication are the areas of interest.

In addition, for example, even if the number of unsubscribed subscribers of A users is higher than that of subscriber subscribers, contents such as the prosecution and the Blue House can confirm that the subscribers have positive opinions. In this way, by understanding positive / negative / neutral opinions on the Korean keyword, it is possible to confirm whether or not the subscribers express a positive opinion on the valid keyword, regardless of support or non-support.

100: data extracting unit 200: subscriber analyzing unit
300: Vocabulary emotion database 310: Emotion data storage unit
320: syllable extracting unit 330: vocabulary identification unit
340: vocabulary addition unit 400:
500: Content extracting unit 600: Keyword extracting unit
700: Keyword recognition unit 800: Content classification unit
900: Keyword analysis section

Claims (5)

A data extracting unit for extracting content response information for a content uploaded by the user of at least one subscriber who subscribes to a content of a user using a social network;
A subscriber analyzer for analyzing the subscriber propensity of the user by comparing the extracted content response information with a vocabulary emotion database;
A sender extracting unit for extracting a main sender of the subscribers according to the order of delivery of the subscribers for transferring the contents of the user to other users in the social network and the mutual subscription relationship among the subscribers;
A content extracting unit for extracting an effective content by removing a content having no content response information among the content uploaded by the user; And
Setting a center value for a number of effective keywords of the center content as a center content and calculating a distance between each content and a center value of the set number of the crowd; Assigning each of the contents to the nearest community, reassigning the center value of each community through an average from each content belonging to the community to the center value, And a content classifying unit for classifying the effective content repeatedly until the content is the same as the content of the subscriber. The method according to claim 1,
The vocabulary emotion database
An emotion data storage unit in which basic emotion vocabularies for positive and negative formed in predetermined syllable units are classified and stored;
A syllable extraction unit for extracting at least one cut syllable by cutting a specific word input from the outside into the predetermined syllable unit;
A vocabulary identification unit for identifying a specific emotional vocabulary corresponding to the cut syllable by comparing the extracted cut syllable with the basic emotional vocabulary; And
And a vocabulary adding unit for classifying the specific words into the identified specific emotional vocabulary and storing the sorted vocabulary into the identified specific emotional vocabulary. 3. The method of claim 2,
A keyword extracting unit for analyzing the text of the effective content to extract a valid keyword having a specific meaning; And
And a keyword recognition unit for recognizing a main keyword transmitted to another user by a predetermined ratio or more of the valid keywords by the main communicator, according to the content uploaded to the social network.
delete The method of claim 3,
Further comprising a keyword analyzing unit for extracting keyword response information of valid keywords corresponding to effective contents classified by sections and comparing the extracted keyword response information with the vocabulary emotion database to analyze the subscriber propensity of the effective keyword A module for analyzing the tendency of subscribers according to contents uploaded to a social network.
KR1020160016083A 2016-02-12 2016-02-12 Module for analyzing of subscriber's tendency by uploaded contents to social network KR101733911B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020160016083A KR101733911B1 (en) 2016-02-12 2016-02-12 Module for analyzing of subscriber's tendency by uploaded contents to social network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020160016083A KR101733911B1 (en) 2016-02-12 2016-02-12 Module for analyzing of subscriber's tendency by uploaded contents to social network

Publications (1)

Publication Number Publication Date
KR101733911B1 true KR101733911B1 (en) 2017-05-24

Family

ID=59051171

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020160016083A KR101733911B1 (en) 2016-02-12 2016-02-12 Module for analyzing of subscriber's tendency by uploaded contents to social network

Country Status (1)

Country Link
KR (1) KR101733911B1 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102183140B1 (en) 2020-05-20 2020-11-25 이창욱 Apparatus and method of calculating user's re-subscribing probability based on big data
CN112988973A (en) * 2021-03-25 2021-06-18 上海柏观数据科技有限公司 Talent emotional tendency detection method based on emotional word matching
WO2022119203A1 (en) * 2020-12-03 2022-06-09 주식회사 큐티티 System for suppressing comment-based distorted behaviors

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101491628B1 (en) * 2013-07-30 2015-02-12 성균관대학교산학협력단 Method, apparatus and system for extracting keyword affecting for mood change of the public using blog

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101491628B1 (en) * 2013-07-30 2015-02-12 성균관대학교산학협력단 Method, apparatus and system for extracting keyword affecting for mood change of the public using blog

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102183140B1 (en) 2020-05-20 2020-11-25 이창욱 Apparatus and method of calculating user's re-subscribing probability based on big data
WO2022119203A1 (en) * 2020-12-03 2022-06-09 주식회사 큐티티 System for suppressing comment-based distorted behaviors
CN112988973A (en) * 2021-03-25 2021-06-18 上海柏观数据科技有限公司 Talent emotional tendency detection method based on emotional word matching

Similar Documents

Publication Publication Date Title
Ottoni et al. Of pins and tweets: Investigating how users behave across image-and text-based social networks
Mehta et al. A review on sentiment analysis methodologies, practices and applications
Larsson et al. Triumph of the underdogs? Comparing Twitter use by political actors during two Norwegian election campaigns
Pennacchiotti et al. A machine learning approach to twitter user classification
Volkova et al. Inferring user political preferences from streaming communications
Bastos et al. Gatekeeping Twitter: message diffusion in political hashtags
Volkova et al. Inferring perceived demographics from user emotional tone and user-environment emotional contrast
Kim et al. Do you feel what i feel? social aspects of emotions in twitter conversations
Alshamsi et al. Sentiment analysis in English texts
US8667169B2 (en) System and method for providing argument maps based on activity in a network environment
US20140108388A1 (en) Method and system for sorting, searching and presenting micro-blogs
JP2012530982A (en) Message delivery method, system and architecture in a network for automatically increasing the signal-to-noise ratio of user interest
Hürlimann et al. A twitter sentiment gold standard for the brexit referendum
Panizo-LLedot et al. Describing alt-right communities and their discourse on twitter during the 2018 us mid-term elections
Sakaki et al. Twitter user gender inference using combined analysis of text and image processing
KR101733911B1 (en) Module for analyzing of subscriber&#39;s tendency by uploaded contents to social network
Okazaki et al. How to mine brand Tweets: Procedural guidelines and pretest
WO2014048479A1 (en) A system and method for the automatic creation or augmentation of an electronically rendered publication document
Ketcham et al. The forensic algorithm on Facebook using natural language processing
Long et al. Churn analysis of online social network users using data mining techniques
Lo et al. Identifying the high-value social audience from Twitter through text-mining methods
Skorniakov et al. Make Social Networks Clean Again: Graph Embedding and Stacking Classifiers for Bot Detection.
Kumari et al. Sentiment analysis of tweets
Dwianto et al. The Sentiments Analysis of Donald Trump and Jokowi's Twitters on Covid-19 Policy Dissemination.
Bernardi et al. Social media analytics for nonprofit marketing:# Downsyndrome on Twitter and Instagram

Legal Events

Date Code Title Description
E701 Decision to grant or registration of patent right
GRNT Written decision to grant