KR101599692B1 - A method for visualizing vocabularies by utilizing pca method and the apparatus thereof - Google Patents

Abstract

The present invention relates to a method and an apparatus for visualizing a vocabulary using a PCA technique, and more particularly, to a method and apparatus for visualizing a vocabulary using a PCA technique, ) To extract the coordinates of each vocabulary included in the vocabulary list and to visualize each vocabulary with a list of extracted coordinates by using a graph, a figure, a picture, an image (including a moving picture), or a combination thereof . According to the present invention, the position of the corresponding vocabulary is given according to the meaning of each vocabulary included in a specific vocabulary list to help understand the meaning of the vocabulary spatially and understand it, and to recognize the difference with other vocabulary at a glance. Is a new method that can replace vocabulary dictionary and vocabulary learning of vocabulary, so that it is possible to visualize vocabulary learners in a foreign language and to learn vocabulary efficiently.

Description

TECHNICAL FIELD The present invention relates to a method of visualizing a vocabulary using a PCA technique,

The present invention relates to a method and an apparatus for visualizing a vocabulary using a PCA technique, and more particularly, to a method and apparatus for visualizing vocabulary using a PCA technique, Analysis, PCA technique) is applied to extract the coordinates of each vocabulary included in the vocabulary list, and each vocabulary is visualized as a graph, a figure, a picture, an image (including a moving picture), or a combination thereof by using the extracted coordinate list Method and apparatus thereof.

In general, learning method of a vocabulary for a specific language uses a dictionary in an off-line mode or a dictionary on an online basis to sense the meaning of each word. In this case, the semantics of each vocabulary are spatially listed, and the search is acquired according to the user's needs. In this way, when a specific vocabulary has many meanings, it takes a lot of time and effort to learn each individual meaning.

In the case of using only texts for learning vocabulary, there is a problem that the efficiency of learning is poor because it does not intuitively remind the actual meaning thereof.

Recently, there have been attempts to improve the efficiency of vocabulary learning in various ways by using ancillary means such as images. However, its effectiveness has been rarely proved, and especially it has a difficult problem to learn a similar vocabulary or an opposite word .

The present invention has been made to solve the above problems and it is an object of the present invention to extract the coordinates of each vocabulary for at least one or more vocabulary lists instead of simply listing the dictionary meanings of vocabularies like a general dictionary, And to provide a method and apparatus for visualizing a vocabulary using the PCA technique.

In addition, the present invention can be implemented in the form of graphs (e.g., distributed graphs), graphics (e.g., language maps), pictures, images (including moving pictures) And to provide a visualization method and apparatus for a vocabulary to be processed and displayed.

The present invention also aims to provide a position for each vocabulary according to its meaning so that the user can visually understand and understand the meaning of each vocabulary.

The present invention also provides a method and a device for visualizing a vocabulary utilizing a PCA technique for generating a map of each vocabulary and inducing a user to memorize a location based vocabulary.

It is another object of the present invention to provide a method and apparatus for visualizing a vocabulary utilizing a PCA technique, which helps a user distinguish differences between similar words as well as an opposite vocabulary.

According to one aspect of the present invention, there is provided an apparatus for visualizing a vocabulary utilizing a PCA technique, comprising: input means for inputting a vocabulary list of at least one vocabulary from a user; Vector generating means for generating a vector for each lexical item for the input lexical item list and vectorizing the generated vector; And coordinate extracting means for extracting coordinates for each vocabulary by applying the PCA technique to the vectorized result in the vector generating means. The apparatus for visualizing a vocabulary utilizing the PCA technique may further include a database for storing data including the meaning of each of the vocabularies and the frequency of use of each meaning for each vocabulary of the vocabulary list, A visualization means for generating, processing, and outputting spatially-visualized data using a graphic including a distributed graph, a graphic map including a language map, a picture, an image, or a combination thereof; Wherein the vector generating means vectorizes the meanings of the respective vocabularies as coordinate axes for each vocabulary in the vocabulary list and sets the frequency of use for each meaning as a coordinate value, ; Vector generating means; The coordinate extracting means, the database, the visualizing means, or a combination thereof may be a wired, wireless or wireless network consisting of a distributed network environment comprising a single server or a cloud computing structure, a local intranet, the Internet, a private network, And is connected to a network constituted by a combination.

According to another aspect of the present invention, there is provided a method of visualizing a vocabulary using a PCA technique, the method comprising: inputting a vocabulary list of at least one vocabulary from a user; A vectorization step of generating vectors for each lexical item for the input vocabulary list and vectorizing the vectors; And a coordinate extracting step of extracting coordinates for each vocabulary by applying the PCA technique to the vectorized result in the vector generating means. In addition, the method of visualizing a vocabulary utilizing the PCA technique may further include storing data including the meaning of each vocabulary and the frequency of use of each meaning for each vocabulary of the vocabulary list, A visualization step of generating, processing, and outputting spatially-visualized data using a graphic including a distributed graph, a graphic map including a language map, a picture, an image, or a combination thereof, Wherein the vector generating step comprises vectorizing the vocabulary list by using the meanings of the vocabularies as coordinate axes for each vocabulary in the vocabulary list and setting the frequency of use for each meaning as a coordinate value, Extracting a coordinate by extracting a coordinate of the vector, the step of extracting coordinates includes a step of obtaining an average vector of vectorized vocabularies; Subtracting an average vector from each vector of the lexicals to obtain a center vector; Obtaining a matrix constituted by the center vectors and a transpose matrix of the matrix, and obtaining a covariance matrix by multiplying the matrix by a transpose matrix; Obtaining eigenvalues and eigenvectors of the covariance matrix; Arranging eigenvectors in order of largest eigenvalues, selecting a predetermined number of eigenvectors from the eigenvectors to generate a eigenspace; Generating an intrinsic matrix composed of the selected eigenvectors and obtaining a transpose matrix of the intrinsic matrix; And obtaining new coordinates in a unique space of each vocabulary by multiplying the transpose matrix of the intrinsic matrix by a center vector of the input vocabularies and outputting new coordinates of the vocabularies projected in the unique space Further comprising: an input step; Vector generation step; The coordinate extracting step, the storing step, the visualizing step, or a combination thereof may be a wired, wireless or wireless network consisting of a distributed network environment including a cloud computing structure, a local intranet, the Internet, a dedicated network, And a network connected to the network.

According to the present invention, the position of a corresponding vocabulary is given according to the meaning of each vocabulary included in a specific vocabulary list, so that the meaning of the vocabulary can be spatially visualized and understood, and the difference with other vocabulary can be recognized at a glance. Based on this, it is possible to map the vocabulary and induce the user to memorize the location based vocabulary. It can also help to distinguish the differences between similar words as well as antonyms. This is a new method that can replace existing vocabulary dictionary and vocabulary learning, and it is effective to enable vocabulary to be efficiently learned by visualizing the vocabulary learners in a foreign language spatially.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1A is a diagram illustrating a configuration of an apparatus for visualizing a vocabulary utilizing a PCA technique according to an embodiment of the present invention; FIG.
FIG. 1B is a diagram illustrating a configuration of a vocabulary visualization apparatus using a PCA technique according to another embodiment of the present invention. FIG.
2 is a flowchart illustrating an operation of a vocabulary visualization apparatus using a PCA technique according to an embodiment of the present invention.
FIG. 3 is a flowchart illustrating a method for generating new coordinates of vocabularies from a vocabulary list input by a visualization apparatus for a vocabulary utilizing a PCA technique according to an exemplary embodiment of the present invention.
4 is a conceptual diagram for obtaining new coordinates in a unique space of each vocabulary by multiplying a transpose matrix of an original matrix according to an embodiment of the present invention by a center vector of vocabulary input first.

Hereinafter, a method of visualizing a vocabulary utilizing the PCA technique and an embodiment of the apparatus will be described with reference to the accompanying drawings.

FIG. 1A is a block diagram illustrating an apparatus for visualizing a vocabulary using a PCA technique according to an embodiment of the present invention. As shown in FIG. 1A, a client 100, a service server 200, a PCA server 300, and a DB 400. [0040]

Here, the client 100 receives the vocabulary list from the user and receives and displays the spatialized visualized data of the vocabulary in the service server through a graph, an image, a picture, an image (including a moving picture), or a combination thereof.

In addition, the service server 200 transmits the list of lexical items to the PCA server 300, assigns positions to the vocabularies according to the list of coordinates received from the PCA server, and provides the positional information of the vocabulary to a graph, , Images (including moving images), or a combination thereof (e.g., distributed graphs, language maps, etc.).

The PCA server 300 extracts the vocabulary list received from the service server 200 and the meaning and frequency of use of each vocabulary in the vocabulary list from the DB 400, And generates coordinates to output to the service server. More specifically, the PCA server includes a vocabulary-based vector generation unit 310 for generating a vector on the basis of the meaning and frequency of use of a specific vocabulary extracted from the DB 400, and a coordinate generating unit 310 for generating coordinates based on the vocabulary- And a database 400 for storing data on the meaning and frequency of use of the vocabulary for each language used in the PCA server and providing the data according to the request of the PCA server can do.

Although the service server 200 and the PCA server 300 are separately described in the present invention, the service server 200 and the PCA server 300 may be separately configured in one server, A distributed parallel server may be constructed.

1B is a block diagram illustrating a configuration of a vocabulary visualization apparatus using a PCA technique according to another embodiment of the present invention. As shown in FIG. 1B, a client 100, a PCA server 300, and a DB 400, May be connected to a network consisting of a wired, wireless, or a combination thereof, consisting of a distributed network environment including a cloud computing architecture, a local intranet network, the Internet, a private network, or a combination thereof. Here, the function of the service server can be included in the PCA server. In FIG. 1A, the service server and the PCA server can be integrated with each other, and for convenience of explanation, the service server and the PCA server are shown separately I want to leave.

Accordingly, the DB 400 may be configured in the PCA server, that is, locally, as shown in FIG. 1A, in a storage device via a wired or wireless network, or in a cloud computing environment. In this case, when a server and a client are connected to a cloud computing environment, the server performs all processes necessary for extracting and visualizing a vocabulary coordinate list by all computing, that is, a PCA technique. The client is a so- The processing required for the computing of the server depends on the server.

In addition, the client 100 can display any type of device including a PC, a tablet, a notebook, a netbook, a smart phone, a PDA, or a combination thereof regardless of the resolution of the screen by using the N-screen . In order to support the N-screen, the server may detect the resolution of the client device between the client and the server (including the service server and the PCA server), output the result according to the resolution, The resolution information of the signal can be analyzed and output according to the display resolution.

In order to provide a N-screen function independent of the resolution and to provide a web application independent of the type of the client-side device and the OS, the PCA server and the client device can transmit and receive the mutual information through the web application by HTML5 have.

More specifically, the service server 200 shown in FIG. 1A can display a graph, a figure, a picture, an image (including a moving picture) desired by the user from a list of coordinates transmitted from a PCA server and a role of delivering a vocabulary list simply from a client to a PCA server Or a combination thereof (for example, a distributed graph or a language map), and may be configured to be performed by an application program on a separate web server, a PCA server, or a client. The application program may also have a message standard or interface capable of communicating with the PCA server by wire / wireless. The client may be a general PC or a dedicated terminal, and may also include a smart phone, a tablet PC, It is possible to provide a device independent web application so that the application programs can be independently executed on the various types of heterogeneous terminals. That is, it means that it is possible to construct the web application program in the HTML5 standard so as to be able to be operated in any device or OS environment.

1A and 1B are one example for realizing the technical features of the present invention. Therefore, the PCA server and the service server can operate on one system or operate as a distributed or concentrated system on a network In addition, each server may operate in a single process without specifying its functional boundary. For example, according to the intention of a system designer, a list of vocabulary items is input from a user and a coordinate of a vocabulary is calculated It is enough to include the extraction process.

As described above, the PCA server 300 shown in FIG. 1A and FIG. 1B receives a list of vocabulary from a client through a service server through a service server and receives a vocabulary list from the DB 400, Lt; / RTI > At this time, the parts of the vocabulary taken from the DB 400 are removed and the dimension is set as one dimension, and the vocabulary requested by the user is vectorized using the frequency of use of each vocabulary as a specific value. In this case, when a portion where the meaning overlaps with respect to a plurality of vocabularies is set as one dimension, the frequency of use of the meaning with respect to each vocabulary may be different. Therefore, when the frequency of use of the specific meaning according to each vocabulary is expressed as a value, A value is generated and a representative value must be set in order to set the plurality of different values to a value for one meaning (dimension). In the present invention, the representative value is taken as an average value for a plurality of use frequency values, In some cases, the maximum value or the minimum value may be set as the representative value, or separately applied to the lexical coordinate list while maintaining the frequency of use of the individual vocabulary.

Each of the vectorized vocabularies is composed of a predetermined number (for example, three in the case of three dimensions) of eigenvectors according to dimensions (for example, two dimensional, three dimensional, Creates a new unique space, creates new coordinates of each vocabulary projected in the unique space, and returns the list to the service server.

FIG. 2 is a flowchart for explaining an operation of a lexical coordinate extracting apparatus using the PCA technique according to the present invention. When a PCA server receives a lexical item list from a client or a service server (S101) (S102), and a vocabulary-specific vector is generated using the frequency-of-use data of the semantics of each vocabulary obtained from the database (S103) The vocabularies are generated using a PCA technique to create a unique space composed of eigenvectors having a predetermined dimension (which can be two-dimensional, three-dimensional or multi-dimensional), and new coordinates of each vocabulary projected in the unique space are generated S104), and outputs the coordinates for each input vocabulary (S105).

3 is a flowchart for explaining a method for generating new coordinates of vocabularies from a list of vocabulary input from a client or a service server and outputting the new coordinates in a vocabulary coordinate extraction apparatus using the PCA technique according to the present invention. (S201). Then, the system obtains the meaning of each vocabulary from the database and each frequency of use data used in the meaning (S202). If the meaning of each vocabulary is duplicated, (S203). Then, a mean vector of the re-vectorized vocabularies is obtained and a center vector obtained by subtracting the average vector from a vector of vocabularies is obtained (S204). The center vectors A matrix composed of the center vectors and a matrix composed of the center vectors are obtained. The covariance matrix is obtained by multiplying the transpose matrix of this matrix (S206), and the eigenvector and eigenvalue of the covariance matrix are obtained again (S207). The eigenvalues are sorted in the order of largest eigenvalues, and a predetermined number of eigenvectors are selected from the eigenvectors to generate a eigenspace (S208). A unique matrix composed of the selected eigenvectors is generated, a transpose matrix of the eigenmatrix is obtained S209). Next, the new matrix is multiplied by the center matrix of the first input vocabulary by the transpose matrix of the original matrix (S210), and new coordinates of the vocabulary projected in the unique space are output (S211).

The operation of the PCA server 300 will be described in more detail as follows.

1. (Step S201) The PCA server receives a vocabulary list from a client or a service server.

2. (Step S202) The DB 400 obtains the meaning of each vocabulary and frequency data used in the meaning of the vocabulary in the DB 400 as a list of inputted vocabularies.

3. (Step S203) The redundancy is removed from the meaning of each vocabulary retrieved from the database and the frequency data used as the meaning thereof (the redundant one is treated as one), and the vocabularies are vectorized using the meaning as the dimension and the frequency of use. Considering the case where the number of vocabulary is n, the vectorization of each vocabulary using the meaning of the vocabulary as a dimension and the frequency of use as a value can be generalized as follows.

In other words, when a vector is expressed as a vector,

와 같이 표현된다.

.

A simpler explanation is as follows. For example, if v1 has the meaning a1, a2, a3, a4, and v2 has the meanings a3, a4, a5, and a6, Here, the meaning a3 and a4 overlap with v1 and v2. Thus, v1 and v2 have a total set of six a1, a2, a3, a4, a5, and a6. Here, a3 and a4 of v1 and v2 are the same (ie, have the same meaning), but the frequency of each meaning is different. Now, the frequencies of the meanings a1, a2, a3, a4, a5, and a6 of v1 and v2 are expressed as follows for v1 and v2, respectively.

In this case, a15 = a16 = 0 and a21 = a22 = 0. Also, a13 ≠ a23 and a14 ≠ a24.

4. In step S204, an average vector of the vectorized vocabularies is obtained. Also, if the mean vector of the vectorized vocabularies is generalized,

와 같이 나타낼 수 있다. 좀 더 간략화하여 설명하면, v1과 v2의 평균벡터({g}(=AVG))는 다음과 같이 표현할 수 있다. 여기서 평균벡터는,

As shown in Fig. To explain it more simply, the average vector ({g} (= AVG)) of v1 and v2 can be expressed as follows. Here,

As shown in Fig.

Next, the mean vector of the vectorized vocabularies is obtained, and the center vector is obtained by subtracting the mean vector from the vector of the vectorized vocabularies. Expressing this as an equation,

와 같이 표현할 수 있다.

Can be expressed as

5. (Step S205) One matrix may be constructed by center vectors of the vectorized vocabulary. That is, if a matrix is constructed with the center vectors,

와 같이 표현할 수 있다. 또한 벡터화한 어휘들의 중심벡터들로 구성한 행렬의 전치행렬을 구하면,

Can be expressed as In addition, if the transpose matrix of the matrix composed of vector vectors of center vectors is obtained,

와 같이 표현할 수 있다.

Can be expressed as

 6. (Step S206) Next, the covariance matrix is obtained by multiplying the matrix composed of the vector vectors of the lexical vectors and the transpose matrix of the matrix. That is, the covariance matrix,

와 같이 나타낼 수 있다.

As shown in Fig.

7. (Step S207) Then, the eigenvectors and eigenvalues of the covariance matrix generated by the matrix multiplication are obtained by the following procedure.

8. (Step S208) Next, the eigenvectors are sorted in order of largest eigenvalues, and three or less eigenvectors are selected from the eigenvectors to generate a unique space. That is, the eigenvectors are sorted in order of largest eigenvalues to select a predetermined number of eigenvectors (e.g., three if three dimensions are displayed), and the remaining eigenvectors are discarded. Thus, the selected eigenvectors form the axis of the new eigen space.

9. (Step S209) A unique matrix composed of the selected eigenvectors is created, and a transpose matrix of the eigenmatrix is obtained. That is, the permutation matrix and the permutation matrix of the permutation matrix,

와 같이 나타낼 수 있다.

As shown in Fig.

10. In step S210, the new matrix is obtained by multiplying the transpose matrix of this eigenmatrix by the center vector of the first inputted vocabulary. This relationship is shown in FIG.

In other words,

및

같은 3개의 좌표를 얻는다.

And

Gets the same three coordinates.

11. (Step S211) The new coordinates of the projected vocabularies in the unique space are outputted and returned, and the work routine of the PCA server 300 is terminated.

The new coordinates of the returned vocabularies will be generated and generated as spatially visualized data in the form of graphs, figures, drawings, images (including video) or a combination thereof (eg, distributed graphs or language maps) Processed and provided to the user.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, I will understand the point. Accordingly, the technical scope of the present invention should be defined by the following claims.

100: client 200: service server
300: PCA server 310: List of vocabulary (vocabulary-specific vector generation)
320: Coordinate list (coordinate creation) 400: Database (DB)

Claims (13)

Input means for inputting a vocabulary list of at least one vocabulary from a user;
A database for storing and managing data related to a vocabulary including a plurality of vocabularies, meaning of each vocabulary, and frequency of use for each meaning;
Data from each of the vocabulary lists of the vocabulary list is retrieved from the database, including the meaning of each vocabulary and the frequency of use of each meaning, thereby generating vectors for each vocabulary and vectorizing the vocabulary. A vector generating means for setting vectors used as the coordinate values of the coordinate axes and generating vector vectors for the input vocabulary lists and vectorizing the vocabulary vectors;
Coordinate extracting means for extracting new coordinates for each vocabulary by applying a PCA technique to the vectorized result in the vector generating means; And
A coordinate list of the vocabulary list extracted by the coordinate extracting means is generated and processed as a spatially visualized data using a graphic including a decentralized graph, a graphic including a language map, a picture, an image, a moving picture or a combination thereof And a visualization means for outputting the visualization result to the visualization means. The method according to claim 1,
The frequency of use of a specific meaning for each vocabulary may be different when the part of the vector generation means in which the meaning for a plurality of vocabularies is overlapped is set as one coordinate axis, A coordinate value for a plurality of different usages is generated for each meaning, and if the coordinate value for the plurality of different usages is set as a coordinate value for one coordinate axis for a specific meaning, Wherein the representative value is defined as one of a mean value, a maximum value, and a minimum value of a plurality of usage frequencies. delete delete The method according to claim 1,
The input means; Vector generating means; Coordinate extraction means; Database; Visualization means, or a combination thereof, may be composed of a wired, wireless, or a combination thereof, configured in one server or in a distributed network environment including a cloud computing architecture, a local intranet, the Internet, a private network, Wherein the computer-readable medium is connected to a network. A method for visualizing a vocabulary using a visualization device of a vocabulary utilizing a PCA technique,
An input step of the vocabulary visualization device receiving a vocabulary list of at least one vocabulary from a user;
Storing and managing data related to a vocabulary including a plurality of vocabularies, meaning of each vocabulary, and frequency of use for each meaning;
The visualization apparatus of the lexical item retrieves data including the meaning of each vocabulary and the frequency of use of each meaning for each vocabulary in the vocabulary list from the database to generate vector for each vocabulary and vectorize the vocabulary, A vector for generating a vector for each lexical item for the input lexical item list by setting the frequency of use of the meaning as a coordinate value of the corresponding coordinate axis, Generating step;
A coordinate extracting step of extracting new coordinates for each vocabulary by applying the PCA technique to the vectorized result of the vocabulary visualization device in the vector generation step; And
The visualization apparatus of the lexical portion may perform a spatial visualization using a graphic including a decentralized graph desired by the user, a graphic including a language map, a picture, an image, a moving picture, or a combination thereof, And a visualization step of generating and processing one data and outputting the generated data. The method of claim 6,
In the vector generation step, when a part where the meaning of a plurality of vocabularies overlap is set as one coordinate axis, the frequency of use of a specific meaning for each vocabulary may be different. Therefore, A coordinate value for a plurality of different usages is generated for each meaning, and if the coordinate value for the plurality of different usages is set as a coordinate value for one coordinate axis for a specific meaning, Wherein the representative value is determined to be one of a mean value, a maximum value, and a minimum value of a plurality of use frequencies, and a method of visualizing a vocabulary utilizing the PCA technique. delete delete The method of claim 6,
In the coordinate extraction step of extracting coordinates by applying the PCA technique,
Obtaining an average vector of vectorized vocabularies, subtracting an average vector from each vector of the vocabularies to obtain a center vector;
Obtaining a matrix constituted by the center vectors and a transpose matrix of the matrix, and obtaining a covariance matrix by multiplying the matrix by a transpose matrix;
Obtaining eigenvalues and eigenvectors of the covariance matrix, sorting the eigenvectors in order of increasing eigenvalues, and selecting a predetermined number of eigenvectors to generate eigenvalues;
Generating an intrinsic matrix composed of the selected eigenvectors and obtaining a transpose matrix of the intrinsic matrix; And
And obtaining new coordinates in a unique space of each vocabulary by multiplying the transposed matrix of the intrinsic matrix by the center vector of the input vocabularies. The method of claim 10,
And outputting new coordinates of the respective vocabularies projected in the unique space by using the PCA technique. The method of claim 6,
The input step; Vector generation step; A coordinate extraction step; A storage step; Visualization steps, or a combination thereof, may be implemented in a network consisting of a wired, wireless, or a combination thereof configured in a single server or in a distributed network environment including a cloud computing architecture, a local intranet, the Internet, a private network, And a visualization method of a vocabulary utilizing the PCA technique. A recording medium storing a computer executable program for executing a method of visualizing a vocabulary utilizing the PCA technique according to any one of claims 6, 7, 12,

Images