-
公开(公告)号：KR1020100132763A
公开(公告)日：2010-12-20
申请号：KR1020090051530
申请日：2009-06-10
Applicant: 성균관대학교산학협력단
IPC: G06F17/00
CPC classification number: G06F17/3071
Abstract: PURPOSE: A relevant tag group generating apparatus and a method thereof are provided to generate a relevant tag group having high reliability by evaluating the correlation among tags based on the characteristics of blogs. CONSTITUTION: A tag combiner(120) generates tag combination from tags, and a correlation calculating unit(130) calculates the correlation between two tags constituting tag combination. A clustering unit(140) clusters the tags based on the correlation. The correlation calculating unit comprises: a page correlation calculating unit which calculates the correlation between the two tags in a page; and a correlation adding unit which calculates the correlation through an adding operation.
Abstract translation: 目的：提供相关的标签组生成装置及其方法，以通过基于博客的特征评估标签之间的相关性来生成具有高可靠性的相关标签组。 构成：标签组合器（120）从标签生成标签组合，相关计算单元（130）计算构成标签组合的两个标签之间的相关性。 聚类单元（140）基于相关性来聚类标签。 相关计算单元包括：页面相关计算单元，其计算页面中的两个标签之间的相关性; 以及相关添加单元，其通过加法运算来计算相关性。
-
公开(公告)号：KR1020090087633A
公开(公告)日：2009-08-18
申请号：KR1020080012995
申请日：2008-02-13
Applicant: 성균관대학교산학협력단
IPC: G06T17/00
CPC classification number: G06T7/70 , G06K9/6223 , G06K9/6297 , G06T7/277 , G06T2207/30196
Abstract: A method and a device for generating an object location trace model and a method and a system for recognizing a pattern of an object location trace are provided to group object location information data per a specific time, thereby effectively modeling the information data. In at least one object location trace information belonging to an object location trace pattern, object location information data including information related to an object location of a specific time is generated(S200). The location information data corresponds coordinates including time. A coordinate within a preset distance range between the coordinates is grouped into one to generate at least one group(S300). The predetermined object location trace pattern corresponds to at least one group(S400). An object location trace model is generated by using a hidden Markov model in the groups(S500).
Abstract translation: 提供了用于生成对象位置跟踪模型的方法和装置，以及用于识别对象位置跟踪的模式的方法和系统，以在特定时间对对象位置信息数据进行分组，从而有效地建模信息数据。 在属于对象位置跟踪模式的至少一个对象位置跟踪信息中，生成包括与特定时间的对象位置有关的信息的对象位置信息数据（S200）。 位置信息数据对应包括时间的坐标。 在坐标之间的预设距离范围内的坐标被分组为一个以产生至少一个组（S300）。 预定对象位置跟踪图案对应于至少一个组（S400）。 通过使用组中的隐马尔可夫模型来生成对象位置跟踪模型（S500）。
-
公开(公告)号：KR102227939B1
公开(公告)日：2021-03-15
申请号：KR1020190038587
申请日：2019-04-02
Applicant: 성균관대학교산학협력단
Abstract: 본발명은단어의문맥정보와형태론적정보를고려한단어임베딩방법및 장치에관한것으로, 본발명의일 실시예에따른단어임베딩방법은, 학습시킬문장에서미등록단어(OOV: Out Of Vocabulary)를미지의토큰(unknown token)으로대체하여문장을가공하는단계, 상기가공된문장에서상기미등록단어를제외한타겟단어의문자(Character)를학습대상인문맥문자모델(Context Character Model)의입력으로입력하는단계, 상기문장에서타겟단어의주변단어에대한주변문맥벡터를조합하여상기문맥문자모델의초기상태로설정하는단계; 및상기문맥문자모델로부터산출된순방향은닉상태(Forward hidden state) 및역방향은닉상태(Backward hidden state)를연결하여생성된상기타겟단어의예측임베딩(Predicted embedding)과상기타겟단어의실제임베딩(Real embedding) 간의오류가최소가되도록, 상기문맥문자모델을학습하는단계를포함한다.
-
-
公开(公告)号：KR101881647B1
公开(公告)日：2018-08-24
申请号：KR1020160112693
申请日：2016-09-01
Applicant: 성균관대학교산학협력단
Abstract: 본발명은온라인게임에서의위치기반그룹분류를통한봇 탐지방법, 봇탐지장치및 컴퓨터프로그램을제공한다. 상기방법은봇 탐지를위한서버에서, 수집된게임데이터를이용하여상기온라인게임상에서의사용자들의위치에기반하여사용자들을군집화하는단계와상기봇 탐지를위한서버에서, 상기군집화하는단계에서산출된사용자그룹들을봇 그룹과사람그룹으로분류하는단계를포함한다.
-
Patent Agency Ranking
公开(公告)号：KR101710010B1
公开(公告)日：2017-02-24
申请号：KR1020150189996
申请日：2015-12-30
Applicant: 성균관대학교산학협력단
Abstract: 본발명은문서의상대적특징을반영한문서요약방법및 문서요약시스템을제공한다. 상기방법은대상문서와배경문서들을그래프로모델링하는단계, 문장간 유사도에기반하여문장들을클러스터링하는단계와상기대상문서와배경문서들간 공통특징들과고유특징들중 적어도하나를포함하여문장들을랭킹(ranking)하는단계를포함한다.
-
公开(公告)号：KR101491628B1
公开(公告)日：2015-02-12
申请号：KR1020130090408
申请日：2013-07-30
Applicant: 성균관대학교산학협력단
CPC classification number: G06F17/30616 , G06F17/2735 , G06F17/30861
Abstract: A method for extracting keywords affecting the emotional change of the public from the blog comprises: an emotion index calculating step of calculating an emotion index using an emotional vocabulary dictionary based on vocabularies included in each comment in the blog; an emotional change section extracting step of extracting an emotional change section of the blog using the calculated emotion index; and a keyword extracting step of collecting comments included in the emotional change section and extracting main keywords included in the emotional change section from the collected comments. Therefore, when keywords affecting the emotional change of the public are extracted, the method helps to infer reasons associated with the emotional change of the public in a specific company or social issue to easily identify the needs of the public, thereby creating economic benefits later.
Abstract translation: 一种从博客中提取影响公众情感变化的关键词的方法，包括：使用基于博客中的每个评论中的词汇的情感词汇词典来计算情绪指数的情绪指标计算步骤; 情绪变化部分提取步骤，使用所计算的情感指数提取博客的情绪变化部分; 以及关键词提取步骤，收集包含在所述情绪变化部分中的评论，并从收集的评论中提取包括在所述情绪变化部分中的主要关键词。 因此，当提取影响公众情感变化的关键词时，该方法有助于推测与特定公司或社会问题上的公众情绪变化相关联的原因，以便于识别公众的需求，从而在以后创造经济效益。
-
公开(公告)号：KR101351719B1
公开(公告)日：2014-01-16
申请号：KR1020120125661
申请日：2012-11-07
Applicant: 성균관대학교산학협력단
IPC: H04N21/258
CPC classification number: H04N21/25891 , H04N21/251
Abstract: A method for calculating program consumption intensity and a program recommendation method using the same are disclosed. The method for calculating program consumption intensity comprises: a program classification step for dividing a series program (P) produced into a total M (M is a natural number greater than or equal to two) series into N (N is a natural number of 2 to M inclusive) stages and granting a weighting to each stage; a watching confirmation step for confirming m (m is a natural number less than or equal to M) series information which the user has watched in the series program; a weighting calculation step for determining whether each series information confirmed in the watching confirmation step is in any one of the N stages and obtaining each weighting (W(P)); and a computation step for computing the consumption intensity (Cscr(P)) of the user for the program using a consumption intensity equation, wherein the consumption intensity equation is defined as Cscr(P)=∑W(P)×N(P), ∑W(P) is defined as the sum of weightings for each of m series, and N(P) is defined as m/M. Thus, the program consumption intensity of the user can be calculated, thereby recommending an appropriate program. [Reference numerals] (AA) Start; (BB) End; (S110) Program classification step; (S120) Watching confirmation step; (S130) Weighting calculation step; (S140) Computation step
Abstract translation: 公开了一种计算程序消耗强度的方法和使用其的程序推荐方法。 用于计算节目消耗强度的方法包括：节目分类步骤，用于将产生的串联节目（P）分为总数M（M是大于或等于2的自然数）串联成N（N是自然数2 到M），并授予每个阶段的权重; 观看确认步骤，用于确认用户在串联节目中观看的m（m是小于或等于M的自然数）序列信息; 加权计算步骤，用于确定在观看确认步骤中确认的每个系列信息是否处于N个阶段中的任一个并获得每个加权（W（P））; 以及计算步骤，用于使用消耗强度方程计算用户对于程序的消费强度（Cscr（P）），其中消耗强度方程被定义为Cscr（P）=ΣW（P）×N（P） ，ΣW（P）被定义为m系列中每一个的权重之和，N（P）被定义为m / M。 因此，可以计算用户的程序消耗强度，从而推荐适当的程序。 （附图标记）（AA）开始; （BB）结束; （S110）程序分类步骤; （S120）注意确认步骤; （S130）加权计算步骤; （S140）计算步骤
-
公开(公告)号：KR101318843B1
公开(公告)日：2013-10-17
申请号：KR1020110087110
申请日：2011-08-30
Applicant: 성균관대학교산학협력단
Abstract: 본 발명은 시간 정보를 활용하여 블로그의 카테고리를 분류하는 방법에 있어서, 상기 블로그 내에서 사용되는 단어들을 상기 블로그 내의 문서로부터 추출하는 단어 추출 단계; 상기 추출된 단어들이 상기 블로그 내의 문서 내에서 출현한 빈도를 토대로 상기 단어 출현의 시간적인 분포를 나타내는 종합 시간 정보 값을 계산하는 계산 단계; 및 상기 단어들의 시간 정보에 따라 상기 블로그의 카테고리를 분류하는 분류 단계를 포함하는 시간 정보를 활용한 블로그 카테고리 분류 방법을 제공한다.
-
公开(公告)号：KR101236897B1
公开(公告)日：2013-02-25
申请号：KR1020090051533
申请日：2009-06-10
Applicant: 성균관대학교산학협력단
IPC: G01C21/30 , G01C21/34 , G08G1/0968 , H04W4/02
Abstract: 목적지 예측 장치는 이동 중인 대상 객체에 대한 이동 경로 정보를 생성하는 경로 정보 생성부, 복수의 그룹들 중 대상 객체의 현재 위치에 상응하는 적어도 하나의 그룹-그룹은 적어도 하나의 이동 경로 정보를 포함함-을 선택하는 그룹 선택부, 선택된 그룹에 포함된 이동 경로 정보와 대상 객체의 이동 경로 정보와의 경로 유사도를 측정하는 유사도 측정부, 및 측정된 경로 유사도에 기초하여 대상 객체의 목적지를 예측하는 목적지 예측부를 포함한다.
-
-
-
-
-
-
-
-
-
Search Results
100
records
(took 0.026 seconds)
