KR20140129833A - Computer-executable sensibility keyword classification method and computer-executable device performing the same - Google Patents

Abstract

A sensitivity keyword classification method is performed in a sensitivity keyword classification server. The sensitivity keyword classification method comprises a step for classifying a sensitivity keyword which represents the emotion and the opinion of a content user with regard to content, by a mood category comprising a mood state before the user uses the content, a situation category comprising a situation of enabling the user to use the content, a feeling category comprising the emotion and the opinion after the user uses the content, and a style category comprising the property of the content defined by a content manufacturer or one or more users; a step for selecting one or more sensitivity keywords included in the mood category, the situation category, the feeling category, and the style category; and a step for determining a user emotion based on the selected sensitivity keyword. The present invention is provided to classify the sensitivity keyword for the content by the mood category, the situation category, the feeling category, and the style category, thereby enabling the user to express various emotions.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer-executable emotional word classification method and a computer-

The present invention relates to a method of classifying emotional vocabularies and a computer executing apparatus for performing the same. More particularly, the present invention relates to a method of classifying emotional vocabulary that can express various kinds of emotions of a user and accurately measure contents, .

Emotion refers to the complex emotional state that occurs in the human body due to the sense or perception that responds to external physical stimuli obtained through the user's experience. Emotional ICT (Information Communication Technology) , And technology that processes necessary emotional information and provides necessary products and services.

Korean Patent Laid-Open Publication No. 10-1102823 (Dec. 29, 2011) grasps the situation of a user and deduces whether or not the user wants to listen to music. If the user wants to listen to music, he infer the mood of the user, And recommending music suitable for the mood, thereby improving the effectiveness of the recommendation.

Korean Patent Laid-Open No. 10-2008-0002187 (Apr. 14, 2008) provides a preference for a user-preferred content based on a content selected by a user, thereby reducing an error of providing a content that the user does not like, The present invention relates to a personalized emotional service system and a personalized emotional service system in accordance with changes in personal emotions and circumstances, which have an effect of enabling users to enjoy services more similar to their emotions, preferences, personalities and tastes.

However, these prior arts and conventional technologies use 7 to 12 limited emotional vocabulary, which makes it difficult to express diverse and complex emotions. In addition, the related arts have a problem in that only the emotional state of the user is described and the expression of the feeling of the contents is excluded and the user-customized service can not be provided properly.

Korean Patent Publication No. 10-1102823 Korean Patent Publication No. 10-2008-0002187

The present invention is to provide a method and system for emotional vocabulary classification capable of expressing various complex expressions of contents.

The present invention provides a method of classifying emotional vocabulary that can more accurately measure a user's emotions in a content retrieval text.

The present invention provides an emotional vocabulary classification method for recommending user-customized contents based on user's emotions.

Among the embodiments, a computer-executable emotional word classification method is performed in a computer-implemented apparatus. The computer-executable emotional vocabulary classification method is classified into a Sensitivity Keyword indicating the feeling and opinion of the content user about the content, a Mood Category including the mood state before the user experiences the content, A feeling category including emotions and opinions after the user has experienced the content, and attributes of the content defined by the content creator or at least one or more users (e.g., And a style category including a property (Property); Selecting at least one emotional vocabulary belonging to the presented mood category, context category, feel category and style category; And determining the user emotion based on the selected at least one emotional vocabulary.

In one embodiment, the step of selecting at least one emotional vocabulary belonging to the mood category, the context category, the feel category and the style category is extracted from the mood category, the context category, the feel category and the style category related phrases collected through social mining Providing each of the middle class category vocabularies representing the emotional vocabulary group as a first sub-emotional vocabulary of the corresponding category; And providing the emotional vocabularies associated with the selected first sub-emotional vocabulary and collected through the social mining to second sub-emotional vocabularies if a first sub-emotional vocabulary of one of the provided first sub- . ≪ / RTI >

In one embodiment, the step of determining user emotions for content based on the selected one or more emotional vocabulary comprises: determining emotional values for each of the selected emotional vocabulary; And generating an emotional vocabulary vector indicating user emotions based on the determined emotional values.

In one embodiment, determining an emotion value through social web mining for at least one content that includes the selected at least one emotional vocabulary; And generating an emotional vocabulary vector representing a user emotion for the content based on the determined emotion value and linking the content to a content-emotional mapping database.

In one embodiment, the step of linking the content to a content-emotional mapping database comprises determining a similarity between the generated emotional vocabulary vector and an emotional vocabulary vector of each of the content in the content-emotional mapping database, And receiving information on at least one emotional content corresponding to the selected emotional vocabulary in the mapping database.

Among the embodiments, a computer-executable emotional vocabulary classification method includes: extracting emotion related vocabulary from emotional text received from a content user; A mood category including the mood state before the user experiences the content, a situation category including the situation where the user can experience the content, a content category in which the user experiences the content, Matching with at least one emotional vocabulary belonging to a style category including a feeling category including later emotions and opinions and a property of a content creator or content defined by at least one or more users step; And determining the user emotion based on the matched at least one emotional vocabulary.

In one embodiment, matching the at least one emotional vocabulary comprises: determining whether the extracted vocabulary is identical to emotional vocabularies included in the respective ranges; And if they are not the same, judging similarity using an emotion lexical ontology (Ontology).

Among the embodiments, the computer executing apparatus may be configured to store a sensibility keyword indicating the feeling and opinion of the content user on the content, a mood category including the mood state before the user experiences the content, A feeling category including emotions and opinions after the user has experienced the contents, and a property category of contents defined by the content creator or at least one or more users, And a style category including a category (Property); An emotional vocabulary selection unit for selecting at least one emotional vocabulary belonging to the presented mood category, situation category, feeling category, and style category; And a user emotion determination unit for determining a user emotion based on the selected at least one emotion word.

In one embodiment, the user emotion determination unit determines an emotion value for each of the selected emotional vocabularies, and generates an emotional vocabulary vector representing the emotion of the user based on the determined emotion values.

In one embodiment, the computer-implemented device comprises a content-emotional mapping database comprising the at least one emotional vocabulary selected via social web mining for each of the at least one content; The computer-implemented apparatus includes an emotional vocabulary extractor for extracting emotional-related vocabulary from emotional text received from a content user; And an emotional vocabulary matching unit for matching the extracted emotion related vocabulary with at least one emotional vocabulary belonging to the mood category, the context category, the feel category, and the style category.

Among the embodiments, the recording medium on which the computer program executed in the computer executing apparatus is recorded may be configured to record the sensibility keyword indicating the feelings and opinions of the content user on the content, the mood state including the mood state before the user experiences the content A Mood Category, a Situation Category that includes a situation where a user can experience the content, a Feeling Category that includes feelings and opinions after the user has experienced the content, and a content creator or at least one And a style category including a property of a content defined by the user; Selecting at least one emotional vocabulary belonging to the presented mood category, context category, feel category and style category; And determining a user emotion based on the selected at least one emotional vocabulary.

The emotional vocabulary classification method and related techniques of the present invention can provide various and complex emotional expressions by providing a feeling vocabulary and a style vocabulary for a content and a user's mood vocabulary and a context vocabulary.

The emotional vocabulary classification method and related arts of the present invention can extract the emotion related vocabulary from the content search text and determine the emotional vocabulary to more accurately measure the emotion of the user.

The emotional vocabulary classification method and the related arts of the present invention can extract emotion related vocabulary from the emotional evaluation of the contents and recommend user customized contents by determining the emotional vocabulary.

1 is a block diagram illustrating an emotional vocabulary classification system.
2 is a block diagram illustrating the computer-implemented apparatus of FIG.
3 is a flowchart of a sensory vocabulary classification method performed in the computer executing apparatus shown in Fig.
4 is a flowchart illustrating a content recommendation method performed by the computer executing apparatus shown in FIG.
FIG. 5A is a diagram for explaining the mood categories among the major categories of the emotional vocabulary classification of the emotional vocabulary ontology.
FIG. 5B is a view for explaining the category of the sensible lexical category of the emotional vocabulary ontology among the major categories.
FIG. 5C is a view for explaining the feeling category among the major categories of the emotional vocabulary classification of the emotional vocabulary ontology.
FIG. 5D is a view for explaining the style category among the major categories of the emotional vocabulary classification of the emotional vocabulary ontology.
6A is a diagram for explaining the relationship between the representative emotional vocabulary and the emotional vocabulary group in the emotional vocabulary ontology.
6B is a view for explaining the relationship between the emotional vocabulary and the emotion related vocabulary group in the emotional vocabulary ontology.
FIG. 7A is an exemplary Web service providing emotional vocabulary in the computer-implemented apparatus shown in FIG. 1; FIG.
FIG. 7B is an exemplary mobile service providing emotional vocabulary in the computer-implemented apparatus of FIG.

The description of the present invention is merely an example for structural or functional explanation, and the scope of the present invention should not be construed as being limited by the embodiments described in the text. That is, the embodiments are to be construed as being variously embodied and having various forms, so that the scope of the present invention should be understood to include equivalents capable of realizing technical ideas. Also, the purpose or effect of the present invention should not be construed as limiting the scope of the present invention, since it does not mean that a specific embodiment should include all or only such effect.

Meanwhile, the meaning of the terms described in the present application should be understood as follows.

The terms "first "," second ", and the like are intended to distinguish one element from another, and the scope of the right should not be limited by these terms. For example, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component.

It is to be understood that when an element is referred to as being "connected" to another element, it may be directly connected to the other element, but there may be other elements in between. On the other hand, when an element is referred to as being "directly connected" to another element, it should be understood that there are no other elements in between. On the other hand, other expressions that describe the relationship between components, such as "between" and "between" or "neighboring to" and "directly adjacent to" should be interpreted as well.

It should be understood that the singular " include "or" have "are to be construed as including a stated feature, number, step, operation, component, It is to be understood that the combination is intended to specify that it does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

In each step, the identification code (e.g., a, b, c, etc.) is used for convenience of explanation, the identification code does not describe the order of each step, Unless otherwise stated, it may occur differently from the stated order. That is, each step may occur in the same order as described, may be performed substantially concurrently, or may be performed in reverse order.

The present invention can be embodied as computer readable code on a computer readable recording medium, and the computer readable recording medium includes all kinds of record servers for storing data that can be read by a computer system . Examples of the computer-readable recording medium include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage server, and the like, and also implemented in the form of a carrier wave (for example, transmission over the Internet) . In addition, the computer-readable recording medium may be distributed over network-connected computer systems so that computer readable codes can be stored and executed in a distributed manner.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined. Commonly used predefined terms should be interpreted to be consistent with the meanings in the context of the related art and can not be interpreted as having ideal or overly formal meaning unless explicitly defined in the present application.

1 is a block diagram illustrating an emotional word classification system 100. As shown in FIG.

Referring to FIG. 1, an emotional word classification system 100 includes a computer execution unit 110, a content providing server 120, and metadata 130. Here, the computer execution unit 110 and the metadata 130 are connected to the content providing server 120 via a network.

The computer execution device 110 may correspond to a computer executing device that accesses the content providing server 120 to search for a content, download the content, or receive a content streaming service. For example, ), A desktop, a laptop, a smartphone, or a tablet personal computer.

The computer execution unit 110 may extract emotion-related vocabularies from the metadata of the user and the contents to select the emotional vocabulary, and classify the contents according to the emotional vocabulary.

The content providing server 120 may receive the text input for retrieving the content from the computer execution device 110 and may measure the emotion of the user based on the content retrieval text. In one embodiment, the content providing server 120 may provide the computer executing apparatus 110 with an emotional vocabulary classification program implemented as a computer-readable recording medium. Here, the provision may include distribution of a computer program, and execution of a computer program in a virtual or cloud computing environment.

The metadata 130 corresponds to the data including the emotion-related vocabulary in relation to the emotional contents. Emotion related vocabularies correspond to vocabulary expressing emotions or opinions of users.

2 is a block diagram illustrating the computer execution device 110 in FIG.

2, the computer execution apparatus 110 includes an emotional vocabulary presentation unit 210, an emotional vocabulary selection unit 220, a user emotion determination unit 230, a transceiver unit 240, an emotional vocabulary extraction unit 250, A content-emotional mapping database 260, an emotional lexical matching unit 270, and a control unit 280. [

The emotional vocabulary presentation unit 210 classifies the sensibility keyword into categories such as a mood category, a situation category, a feeling category, and a style category. Hereafter, the emotional vocabulary and each category will be explained in detail.

The emotional vocabulary corresponds to a vocabulary expressing the feelings and opinions of the content users on the content. Here, the emotional vocabulary can be collected from the metadata 130 through social mining and can be periodically updated.

FIG. 5A is a diagram for explaining the mood categories among the major categories of the emotional vocabulary classification of the emotional vocabulary ontology.

Referring to FIG. 5A, the mood category corresponds to a category including mood states before the user experiences the content, that is, emotional vocabularies indicating mood states suitable for the user to experience the contents. More specifically, the emotional vocabulary belonging to the mood category may correspond to a vocabulary representing a mental state that continuously appears to the user without external stimulation.

FIG. 5B is a view for explaining the category of the sensible lexical category of the emotional vocabulary ontology among the major categories.

Referring to FIG. 5B, the category of the situation may be classified into the following categories: a situation where the user can experience the content, that is, an emotion indicating a time, a place and an occasion in which the user is judged suitable for experiencing the content It corresponds to the category including the vocabulary. More specifically, an emotional vocabulary belonging to the category of a situation may correspond to a surrounding condition that affects a user, or a vocabulary representing a total or environmental condition of a stimulus that stimulates a user.

FIG. 5C is a view for explaining the feeling category among the major categories of the emotional vocabulary classification of the emotional vocabulary ontology.

Referring to FIG. 5C, the feeling category corresponds to a category including emotional voices expressing emotions and opinions that the user felt after experiencing contents. More specifically, the emotional vocabulary belonging to the emotional category may correspond to a vocabulary representing a momentary response to the content derived from the user's experience, and may correspond to a user's opinion on the content.

FIG. 5D is a view for explaining the style category among the major categories of the emotional vocabulary classification of the emotional vocabulary ontology.

Referring to FIG. 5D, the style category corresponds to a category including emotional vocabularies representing a property of content defined by a content creator or at least one or more users. More specifically, an emotional vocabulary belonging to a style category may correspond to a vocabulary representing a general form of content containing a similar feature, and may be a vocabulary that is not explicitly, but implicitly recognizable, .

Referring to Figs. 5A to 5D, each of the major category categories for the emotional vocabulary classification includes one or more sub category categories.

For example, mood categories include sub-categories [when depressed], [when they are laughing], [when they are laughing], [when they are alive].

5A to 5D, each of the secondary classification categories includes emotional vocabularies that are related to the corresponding secondary classification category among the sensory vocabularies extracted from the mood category, the condition category, the feeling category, and the style category related phrases collected through social mining do.

For example, the category [category of mood] of the mood category [when depressed] is classified into emotional vocabulary such as [when depressed], [when sad], [when annoyed], [when anger], [when lonely] .

6A is a diagram for explaining the relationship between the representative emotional vocabulary and the emotional vocabulary group in the emotional vocabulary ontology.

Referring to FIG. 6A, each representative emotional vocabulary represents an emotional vocabulary group composed of one or more emotional vocals having the same or similar meaning.

Here, the representative emotional vocabularies may correspond to emotional vocabularies extracted from the emotion related phrases collected through social mining from the metadata associated with the contents. In addition, if the number of extracted emotional vocabularies exceeds a certain criterion, the extracted emotional vocabularies can be clustered, and the second sub-emotional vocabulary can correspond to the emotional vocabulary representing each of the emotional vocabulary groups generated through clustering have.

For example, in the case of [emotional vocabulary] [loneliness], which belongs to the category [category of mood], which is a category of mood categories, the words of the same or similar meaning [lonely], [when solitude], [when lonely] , 'Emotional vocabulary', 'emotional vocabulary', and 'emotional vocabulary'.

In another example, if 30 emotional vocabularies such as [addictive], [attractive], [falling], [lovely], [lovely], [beautiful] You can create groups of emotional vocabulary including [Attractive] and [Disappearing] and others including [Lovely] and [Lovely], and you can create [addictive], [lovely] As a representative emotional vocabulary. Through this, it is possible to provide emotional vocabularies that are diverse and uncomplicated.

In one embodiment, the first sub-emotional vocabulary may correspond to the vocabulary having the highest appearance frequency in the classified emotional vocabulary set by clustering the emotional vocabularies collected through social mining. In other words, the first sub-emotional vocabulary is characterized in that the emotional vocabularies collected by the social mining are clustered to correspond to the emotional vocabulary having the highest appearance frequency in the classified emotional vocabulary set. Also, when the number of emotional voices collected by social mining exceeds a certain range, the second sub-emotional vocabulary may be determined by the method described above.

For example, if emotional vocabulary such as [unique], [lovely], [bright], and [refreshing] are collected and the frequency of occurrence is 100, 50, 200, And the [brightest] with the highest appearance frequency can be selected as the first sub-emotional vocabulary.

6B is a view for explaining the relationship between the emotional vocabulary and the emotion related vocabulary group in the emotional vocabulary ontology.

Referring to FIG. 6B, each emotional vocabulary represents a sensitivity related vocabulary group consisting of one or more emotional related vocabulary expressing the same or similar emotions. Here, emotion-related vocabularies correspond to words, phrases, or phrases expressing emotion in emotion related phrases collected through social mining from metadata associated with contents.

For example, in the case of [emotional vocabulary belonging to the mood category, [lonely when], expressions of the same or similar emotions such as [lonely time], [loneliness ride], [loneliness], [when lonely] It can be defined as representing emotional lexical group consisting of extended emotion related vocabulary such as [good to hear when lonely], [lonely when to listen], [to listen when lonely] as well as simple emotional related vocabulary.

Another example is the use of phrases such as [addictive], [addicted], [addictive], etc. in phrases such as [addictive song], [addicted song], [addictive music] Emotion related vocabularies can be extracted, and emotional vocabularies such as [addictive] can be derived in consideration of the semantic identity, similarity and representativeness of the above emotional related vocabularies.

In one embodiment, the emotional vocabulary presentation unit 210 provides a mood category, a mood category, a feeling category, and a style category, which are the major category categories, and, when one of the major classification categories is selected from the provided major classification categories, Category vocabularies as first sub-emotional vocabularies, and when one of the provided first sub-emotional vocabularies is selected, the emotional vocabulary associated with the selected first sub-emotional vocabulary and collected through the social mining, Vocabulary can be provided.

FIG. 7A is an exemplary Web service providing emotional vocabulary in the computer-implemented apparatus shown in FIG. 1; FIG.

Referring to FIG. 7A, the emotional vocabulary presentation unit 210 provides a general category category of [feel], [mood], [context], and [style], and when the [feel] category is selected, If the first sub-emotional vocabulary such as [bright] is selected, the [exciting], [sub-emotional], and [ Pleasant] can be provided as the second sub-emotional vocabularies. Here, the subdivision category [all] can be defined as including all emotion vocabularies belonging to the corresponding major category.

FIG. 7B is an exemplary mobile service providing emotional vocabulary in the computer-implemented apparatus of FIG.

Referring to FIG. 7B, the emotional vocabulary presentation unit 210 may classify middle classification category vocabularies such as [bright], [dynamic], [calm], and [dark] And when a first sub-emotional vocabulary such as [bright] is selected, emotional vocabulary such as [exciting], [amusing] or the like can be provided as second sub-emotional vocabularies.

The emotional lexical selecting unit 220 selects at least one emotional vocabulary belonging to the mood category, the context category, the feeling category, and the style category.

In one embodiment, the emotional vocabulary selecting unit 220 can select an emotional vocabulary according to a user's selection request. Alternatively, the emotional vocabulary selecting unit 220 may select the emotional vocabulary most frequently selected at the same time, or may select the emotional vocabulary at random.

In one embodiment, the emotional vocabulary selection unit 220 may select the emotional vocabulary to evaluate the emotional state of the user before using the content. In order to measure the emotional change of the user after the user uses the content, .

The emotional vocabulary selecting unit 220 may match the emotional vocabulary extracted from the received emotional text with the emotional vocabulary. More specifically, the emotional vocabulary selecting unit 220 can match the emotional vocabulary with the emotional vocabulary by checking the corresponding relationship between the emotional vocabulary and the emotional vocabulary based on the sensible vocabulary list stored in the sensible vocabulary ontology. In addition, the correspondence relation between the emotional vocabulary and the representative emotional vocabulary can be confirmed, and the emotional related vocabulary can be matched with the representative emotional vocabulary.

For example, if the extracted emotion-related vocabulary contains [good to be heard when lonely] and [emotionally good to listen] and emotional vocabulary [when lonely] have a correspondence relationship, Can be matched to [when lonely]. Also, when the emotional vocabulary [lonely time] has a correspondence with the representative emotional vocabulary [lonely time], the emotional vocabulary [good when listening] can be matched to the representative emotional vocabulary [lonely time].

The emotional vocabulary selecting unit 220 can select the emotional vocabulary when the matching between the emotional vocabulary and the emotional vocabulary is performed.

The user emotion determination unit 230 determines a user emotion representing the feeling and opinion of the content user on the content based on the emotional vocabulary selected in the mood category, the context category, the feeling category, and the style category.

In one embodiment, the user emotion determination unit 230 may determine the emotion of the user based on at least one emotional vocabulary selected within a mood category, a context category, a feeling category, and a style category. That is, the user emotion determination unit 230 may determine the user emotion by combining at least some of the emotion words selected in each category.

For example, the user emotion determination unit 230 may determine each emotion word belonging to a mood category, a situation category, a feeling category, and a style category as user emotion, or a combination of two or more emotion categories, And emotional vocabularies belonging to the style category.

In one embodiment, the user emotion determination unit 230 may generate a sensory vocabulary vector by giving a weight to the selected emotional vocabulary. Here, the emotional vocabulary vector may indicate the degree of emotion of the content user with respect to the content. In addition, the emotional lexical vectors of the individual users may be combined to indicate the degree of emotion of the content user group with respect to the content.

The emotional lexical vector can be implemented in the form of an n-dimensional vector corresponding to one dimension, each of the emotional vocabularies belonging to at least one category of the mood category, the context category, the feeling category, and the style category according to the following formula .

SKV (x, y): A function that expresses the degree of emotion of the content user (or the content user group) y with respect to the content x as an n-dimensional vector.

SKW (i, x, y): a function indicating the weight for the emotional vocabulary i of the content user (or the content user group) y for the content x

Here, n corresponds to the size of the entire emotional vocabulary, and may not include the content user y if the content user group corresponds to the entire user (for example, SKV (x), SKW (i, x) ). Likewise, it may not include content x (for example, SKV (y), SKW (i, y)) when expressing emotion for any content other than emotion for specific content x.

For example, for content a, content user b selects emotional vocabulary [exciting] and [when depressed], content user c selects emotional vocabulary [exciting], [autumn], [ballad] (A, b) and SKV (a, c) of the content users b and c for the content a according to Equation (1), the user emotion determination unit 230 determines the weight Can be generated as (1, 1, 0, 0), (1, 0, 1, 1) Here, the total number of emotional vocabulary is four, and it is presumed that they are listed in order of feeling ([exciting]), mood ([when depressed]), situation ([autumn]) and style category ([ballad]).

The user emotion determination unit 230 may generate the emotional vocabulary vector of the user group representing the degree of emotion of the content user group for the content by combining the emotional vocabulary vectors of the individual users.

For example, when defining a content user group d composed of content users b and c, the user emotion deciding unit 230 converts the emotional vocabulary vector SKV (a, d) of the content user group d with respect to the content a to SKV (2, 1, 1, 1) through SKV (a, b) and SKV (a, c) SKV (a, d) can be generated. On the other hand, when considering the intersection of SKV (a, b) and SKV (a, c), the user emotion determination unit 230 determines SKV (a, d) having a value of (2, 0, 0, Can be generated.

The user emotion determination unit 230 may apply the TF-IDF (Term Frequency-Inverse Document Frequency) scheme used in fields such as information retrieval and text mining to adjust the weight of the emotional vocabulary vector of the content user group for the content have.

TF-IDF is a weight used in information retrieval and text mining, which is a statistical value that indicates how important a particular word is in a particular document in a document set of documents.

Here, the document corresponds to a document including an emotional vocabulary in relation to the content, and may correspond to, for example, a comment or a content evaluation.

TF (word frequency, term frequency) is a value that indicates how often a particular word appears in the document. The higher the value, the more important it is in the document. However, if a particular word is frequently used in a document set, it can be determined that the importance is low because a particular word is common. This value is called DF (document frequency) document frequence). TF-IDF is the product of TF and IDF.

For example, if an emotional vocabulary [exciting] appears 10,000 times in emotion related metadata, and [exciting] is included in a total of 1,000 documents or contents, the weight for emotional vocabulary [exciting] is 10,000 / 1,000 = 10 .

The transceiver 240 receives emotional text and content information or transmits emotional vocabulary and content information determined by the user emotion deciding unit 230 to the outside.

More specifically, the transmission / reception unit 240 can receive sentence text indicating the emotion of the user from the user. Here, the text corresponds to the text of the finished form including the word, or may correspond to the vocabulary itself. The content retrieval text may also include emotion related vocabularies such as the user's mood, the user's situation, the feel of the content, and the style of the content.

Emotional texts can include a [rainy day] situation and a feeling of [sad] content, such as [a sad song that wants to hear a rainy day] Can be expressed in the form.

In one embodiment, the transceiver 240 may restrict the size of the emotional text input from the user to a specific size or less, and limit the format of the emotional text to be input. For example, the transmitting and receiving unit 240 may restrict the size of the emotional text to be input to 10 characters or less, and may restrict the input of nouns, thereby disallowing adjectives such as [rain]. By restricting the formatting, it is possible to facilitate the determination of the thesaurus associated with the emotional lexical decision.

The transceiver 240 can transmit the emotion vocabulary determined by the user emotion deciding unit 230 to the outside, and can receive emotional contents related thereto.

The transmitting and receiving unit 240 can transmit the emotion evaluation of the user or the emotional vocabulary to the external to the emotional contents. For example, the computer execution device 110 transmits the determined emotional vocabulary to the content providing server 120, receives the corresponding emotional content, and transmits the emotional vocabulary representing the user's emotional change to the content after using the emotional content, To the providing server 120.

The emotional vocabulary extraction unit 250 extracts the emotion related vocabulary from the emotional text received from the content user.

The emotional vocabulary extractor 250 may extract emotion related vocabulary using K-structure, PMI-IR, semantic web, emotional vocabulary ontology, and the like.

In one embodiment, the emotional vocabulary extraction unit 250 may extract a vocabulary per unit vocabulary except for a content name, an artist name, a genre, and the like in the received emotional text.

For example, if the received emotional text is [a sad song that wants to listen to a rainy day], the emotional vocabulary extractor 250 removes [desired to hear], [song] related to the genre of the content, You can extract emotion related vocabulary such as [sad].

The content-emotional mapping database 260 stores an emotional vocabulary ontology including a sensible vocabulary list and a sensibility-related vocabulary list. The emotional vocabulary list and the sensibility related vocabulary list included in the emotional vocabulary ontology may include the contents of FIGS. 5A to 6B as described above.

The content-emotional mapping database 260 may store a content list including general information (e.g., genre, name, producer, etc.) of content and emotional vocabulary vectors.

The content-emotional mapping database 260 may include a content-emotional mapping table that stores emotional vocabulary, and mapping information between emotional vocabulary vectors and emotional contents.

The content-emotional mapping table may include emotional vocabulary corresponding to a specific emotional vocabulary vector and emotional content information. In recommending emotional contents to a user, the emotional vocabulary may be classified into emotional vocabulary, Can be used to search and provide corresponding content based on a vector.

The emotional vocabulary matching unit 270 matches the emotional vocabulary with the sensible content.

In one embodiment, the emotional vocabulary matching unit 270 may match the emotional vocabulary selected with the emotion vocabulary based on the content-emotional mapping table. The emotional vocabulary matching unit can search for and match contents having the same value as the calculated emotional vocabulary vector of the emotional vocabulary.

The emotional vocabulary matching unit 270 can determine the similarity between the emotional vocabulary vectors and match the emotional vocabulary with the contents.

In one embodiment, the emotional vocabulary matching unit 270 includes an emotional vocabulary corresponding to at least one of the emotion category, the style category, the mood category, and the context category through the emotional vocabulary ontology for each of the at least one content. Content-emotional mapping database. That is, the emotional vocabulary matching unit 270 may search the content corresponding to the selected emotional vocabulary by connecting the selected emotional vocabulary to the content-emotional mapping database. Here, the content search can be performed based on the similarity between emotional vocabulary vectors.

In one embodiment, the emotional lexical matching unit 270 can determine the similarity between the emotional vocabulary vector generated by the user emotion deciding unit 230 and the emotional vocabulary vectors of the contents in the content-emotional mapping database. Here, the degree of similarity may be calculated by a similarity calculation method such as an Euclidean distance between emotional vocabulary vectors, a cosine coefficient, and the like, and the cosine coefficient may be determined by the following equation.

θ: Angle between emotional vocabulary vectors

sim (x, j): cosine coefficient similarity between emotional lexical vector SKV (x) for content x and emotional lexical vector SKV (j)

SKV (x): an emotional vocabulary vector indicating the degree of emotion of the entire user with respect to the content x

SKV (j): an emotional vocabulary vector representing the degree of emotion of user j for arbitrary contents

For example, if the emotional vocabulary vectors of contents a and b are (3,0,4,0) and (4,0,3,0) respectively, the similarity angle between emotional vocabulary vectors is cos ^-1 ((3 * 4 + 4 * 3) / (5 * 5)) = 18.2. The emotional lexical matching unit 270 can judge that the contents a and b are similar according to a predetermined specific criterion (for example, when the similarity degree is within 20 degrees).

The control unit 280 includes a content emotional lexicon selecting unit 210, a user emotional lexical selecting unit 220, a user emotion deciding unit 230, a transceiver 240, an emotional lexical extractor 250, The emotional lexical matching unit 270 and the emotional lexical matching unit 270. [

FIG. 3 is a flowchart for explaining an emotional vocabulary classification method performed by the computer execution apparatus 110 shown in FIG.

Referring to FIG. 3, the emotional vocabulary presentation unit 210 classifies and presents the emotional vocabulary into a feeling category, a situation category, a feeling category, and a style category (S310).

The emotional vocabulary selecting unit 220 selects at least one emotional vocabulary belonging to each of the presented categories (S320).

In one embodiment, the emotional vocabulary selection unit 220 provides the mood category, the context category, the feeling category, and the style category in a large category category, and when one of the large category categories is selected, the small category category belonging to the selected large category category Providing vocabularies in a first sub-emotional vocabulary, and when a first sub-emotional vocabulary of one of the provided first sub-emotional vocabularies is selected, the emotional vocabularies associated with the selected first sub- It can be provided with emotional vocabularies.

The user emotion determination unit 230 determines the user emotion based on the selected at least one emotion word (S330).

In one embodiment, the user emotion determination unit 230 may determine the user emotion by combining at least a part belonging to a mood category, a situation category, a feeling category, and a style category.

In one embodiment, the user emotion determination unit 230 may generate a sensory vocabulary vector by giving a weight to the selected emotional vocabulary.

4 is a flowchart illustrating a content recommendation method using the emotional vocabulary classification method performed by the computer execution apparatus 110 shown in FIG.

Referring to FIG. 4, the transceiver 240 receives emotional text indicating a user emotion from the user (S410). For example, the transceiver 240 may receive emotional text for retrieving content suitable for the user's emotions from the user.

The transceiver 240 may prompt the input of emotional text to receive emotional text from a user.

The emotional vocabulary extraction unit 250 extracts emotion related vocabulary from the sensible text received through the transceiver unit 240 (S420).

The sensibility vocabulary selection unit 220 matches the sensibility related vocabulary extracted by the vocabulary extraction unit 240 with the sensible vocabulary (S430).

The user emotion determination unit 230 determines the user emotion based on the matched emotional vocabulary (S440).

In one embodiment, the user emotion determination unit 230 may generate a sensory vocabulary vector by weighting each of the selected emotional vocabulary.

The emotional content providing server 120 may connect to a content-emotional mapping database including content information for recommending a content corresponding to the determined emotional vocabulary to a user.

In one embodiment, the emotional vocabulary matching unit 270 may link the determined emotional vocabulary to a content-emotional mapping database containing at least one emotional vocabulary vector determined through social web mining for each of the at least one content (S450 ).

In one embodiment, the emotional lexical matching unit 270 can determine the similarity between the emotional vocabulary vector generated by the user emotion deciding unit 230 and the emotional vocabulary vectors of the emotional contents in the content-emotional mapping database.

In one embodiment, the emotional vocabulary matching unit 270 may arrange content in order from contents having the highest similarity.

The transmitting and receiving unit 240 can provide the contents sorted by the emotional lexical matching unit 270 to the user.

In one embodiment, the transceiver 240 can receive the content selected by the user from the content providing server 120 and provide it to the user.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the present invention as defined by the following claims It can be understood that

100: Emotional Vocabulary Classification System
110:
120: Content providing server
130: Metadata
210: Emotional vocabulary presentation unit
220: emotional vocabulary selection part
230: User emotion determination unit
240: Transmitting /
250: emotional vocabulary extraction unit
260: Content-emotional mapping database
270: emotional vocabulary matching unit
280:

Claims (11)

A Sensitivity Keyword indicating the feeling and opinion of the content user about the content, a mood category including the mood state before the user experiences the content, and a situation where the user can experience the content A category including a feeling category including feelings and opinions after the user has experienced the content, and a style category including a property of the content creator or content defined by at least one or more users Style Category);
Selecting at least one emotional vocabulary belonging to the presented mood category, context category, feel category and style category; And
And determining user emotion based on the selected at least one emotional vocabulary.
The method according to claim 1, wherein classifying and presenting the emotional vocabulary into the mood category, the situation category, the feeling category,
Providing each of the category category vocabulary representing the emotional vocabulary group extracted from the sentence category, the context category, the feel category, and the style category related phrase collected through the social mining as the first sub-emotional vocabulary of the corresponding category; And
If one of the provided first sub-emotional vocabularies is selected, providing emotional vocabularies associated with the selected first sub-emotional vocabulary and collected through the social mining to second sub-emotional vocabularies Including computer executable emotional vocabulary classification methods.
The method of claim 1, wherein determining the user emotion for the content based on the selected one or more emotional vocabularies
Determining emotional values for each of the selected emotional vocabulary; And
And generating an emotional vocabulary vector representing user emotions based on the determined emotional values.
The method according to claim 1,
Determining an emotion value through social web mining for at least one content that includes the selected at least one emotional vocabulary; And
Further comprising generating an emotional vocabulary vector representing user emotions for the content based on the determined emotional value and linking the content to a content-emotional mapping database.
5. The method of claim 4, wherein the step of coupling the content to a content-
Determining a similarity between the generated emotional vocabulary vector and an emotional vocabulary vector of each of the contents in the content-emotional mapping database; And
Further comprising receiving information on at least one emotional content corresponding to the selected emotional vocabulary in the content-emotional mapping database.
Extracting emotion-related vocabulary from emotional text received from a content user;
A mood category including the mood state before the user experiences the content, a situation category including the situation where the user can experience the content, a content category in which the user experiences the content, Matching with at least one emotional vocabulary belonging to a style category including a feeling category including later emotions and opinions and a property of a content creator or content defined by at least one or more users step; And
And determining user emotion based on the matched at least one emotional vocabulary.
7. The method of claim 6, wherein matching the at least one emotional vocabulary comprises:
Determining whether the extracted vocabulary is the same as the emotional vocabularies included in the respective ranges; And
And if it is not the same, judging similarity using an emotional vocabulary ontology (Ontology).
A Sensitivity Keyword indicating the feeling and opinion of the content user about the content, a mood category including the mood state before the user experiences the content, and a situation where the user can experience the content A category including a feeling category including feelings and opinions after the user has experienced the content, and a style category including a property of the content creator or content defined by at least one or more users Style Category);
An emotional vocabulary selection unit for selecting at least one emotional vocabulary belonging to the presented mood category, situation category, feeling category, and style category; And
And a user emotion determination unit for determining a user emotion based on the selected at least one emotional vocabulary.
9. The apparatus of claim 8, wherein the user emotion determination unit
Determines emotion values for each of the selected emotional vocabulary, and generates an emotional vocabulary vector indicating the emotion of the user based on the determined emotional values.
9. The method of claim 8,
A content-emotional mapping database containing at least one emotional vocabulary selected via social web mining for each of the at least one content;
An emotional vocabulary extractor for extracting emotional related vocabulary from emotional text received from a content user; And
And an emotional lexical matching unit for matching the extracted emotion related vocabulary with at least one emotional vocabulary belonging to the mood category, the context category, the feel category, and the style category.
A Sensitivity Keyword indicating the feeling and opinion of the content user about the content, a mood category including the mood state before the user experiences the content, and a situation where the user can experience the content A category including a feeling category including feelings and opinions after the user has experienced the content, and a style category including a property of the content creator or content defined by at least one or more users Style Category);
Selecting at least one emotional vocabulary belonging to the presented mood category, context category, feel category and style category; And
And determining a user emotion based on the selected at least one emotional vocabulary.

Images