KR101151965B1

KR101151965B1 - An n-Dimensional Vector Tag and Its Uitlization for Internet Contents

Info

Publication number: KR101151965B1
Application number: KR1020070100358A
Authority: KR
Inventors: 정재윤; 김종근
Original assignee: 김종근; 정재윤
Priority date: 2007-10-05
Filing date: 2007-10-05
Publication date: 2012-06-01
Also published as: KR20090035215A

Abstract

According to the present invention, when a user uses content on the Internet and classifies it and sets a bookmark in a web space, a multidimensional vector is used as an intuitive display method for a tag given by a user. And this has the purpose to make the search engine more efficient by using it as metadata. The idea is to provide high performance for information browsing by multiplying the multi-dimensional vector used to define the tag to express the weight of the information, thereby increasing efficiency in classifying, managing, and recycling the information.

The tag definition using the multidimensional vector created in the present invention is used as a starting point of a position vector representing a weight of information as a search intent and a keyword, which is a classification criterion of contents, and information for constructing a position vector in a virtual vector space. It consists of a vector component based on a rank, an information time (information generation time), and a reference value (weight of another person).

The tagging using the multi-dimensional vector of the present invention expresses users' evaluation of the Internet content by an intuitive method, and can infer the value of information by using the tag without directly accessing the content. In addition, it can also express the evaluation according to the purpose of using the information.

Multi-dimensional vector, tags, internet content, browsing

Description

An n-Dimensional Vector Tag and Its Uitlization for Internet Contents}

The present invention relates to the idea of using multidimensional vectors to intuitively express the weight of information when tagging Internet content.

Categorizing information into categories is a difficult task. Moreover, in a world where information on the web is exploding, classifying content on the Internet is usually not a difficult task. Internet search engines continue to study the technology to determine the value of information based on content classification in order to meet the needs of users who are dissatisfied with deriving the wrong information by not removing unnecessary information.

The currently widely used and applied technology in this field is social bookmarking. This is the principle of registering as a favorite when you find interesting sites on the Internet. By analyzing the favorites of millions of individuals, you can see in real time which Internet content is receiving attention. In general, the value of information depends on search engine users or visitors, such as page views, links, referrals, page rank, etc. Make an effort to identify it.

On the other hand, as a technology to complement the mechanical search method, like FolkSonomi, it uses a tag to give information or a word to give a different value to the document or words. This technology has been used as a flag, label, etc. in the past, it is applied and automated in various fields than before, and it is easy to share through the Internet. However, with a simple tag, it is difficult to determine how much the value included in the Internet content corresponds to the search intention, and if a large number of keywords are defined in the tag, complicated comparison operations may be required to reduce performance. .

Accordingly, the present invention was devised to improve such a point, and the search engine intuitively expresses the weight of information using a vector in the tag in performing a semantic search based on an individual tagging behavior. The purpose of this is to suggest an idea to be provided with information that matches search intention.

The present invention for achieving the above object, in order to express the information weight of the multi-dimensional vector method to the tag to enable the classification, management and recycling of information in the user's search behavior for Internet content,

When defining the tag of the information, as the starting point of the position vector for intuitively expressing the weight while specifying the index and the keyword of the classification,

In order to form a multi-dimensional virtual vector space of the position vector using the keyword as a starting point, the X axis is represented by a rank of information, the Y axis is represented by information time (information generation time), and the Z axis is represented by a reference value (weight of another person).

The vector size calculated by each component of the location vector may express a degree of agreement with the search intention. In this case, Internet content including a tag using the same or similar keyword is classified into the same or similar categories.

The present invention is an intuitive information weight expression using a multidimensional vector to a tag given when an individual classifies the content of the Internet and sets a bookmark, and helps search agents browse information closer to the search intention.

Multi-dimensional vectors used in tags help to increase efficiency in classifying, managing, and recycling information along with keywords, and can overcome the limitations of information weight expression of existing bookmarks and forksonomies, and are useful for using semantic search. Provide an environment.

In addition, the implicit weights included in the multi-dimensional vector of the present invention support ranking operations for information search results targeting a myriad of Internet contents, so that not only easy to share excellent contents but also more quickly and accurately match the search intentions. You can browse.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings. The present embodiment uses a three-dimensional vector that defines three vector components, but can be expanded to n-dimensional by adding the vector components as necessary.

1 is a block diagram showing a tag definition format using a vector.

In general, search engines produce a huge number of search results by semantic search according to the information searcher's needs. Therefore, it is necessary to select the results that match the searcher's search intention and apply the ranking to the search results. To do this, when defining the tag of the information, the keyword 101 indicating the classification index and the search intention is designated and searched. The suitability for intention is expressed by the vector components 102-104.

The keyword 101 is a key word representing a job or a field to be searched based on a profile of a user who uses a search service, and is an index for classifying searched Internet content. The keyword 101 is a classification index and is a starting point of a vector in the vector space constituted by the vector components 102 to 104.

The vector components 102 to 104 form a multidimensional vector on a scale representing the weight of the information. The vector component is composed of components corresponding to an information rank 102, an information time 103, and a reference value 104. The multidimensional vector composed of each component represents a weight of the corresponding content. The information rank 102 is a rank value such as a user's recommendation, a citation, and the like, and the information time 103 is time information displayed when the information is generated. In addition, the reference value 104 refers to a weight previously assigned to the Internet content by another person, and has a kind of normalization function.

The tag definition format shown in Fig. 1 is defined by a specific grammatical framework according to a markup language or a semantic language that employs the keyword 101 and the vector components 102 to 104. For example, in the case of using the HTML, the content can be inserted into the content as " METAA NA = MeeTe r "

FIG. 2 is a conceptual diagram illustrating a vector space used as the base of the multidimensional vector and the multidimensional vector used in the tag of FIG. 1.

The multidimensional vector used in the tag of FIG. 1 is a vector having the same concept as the multidimensional vector 202 of FIG. 2, and the multidimensional vector 202 of FIG. 2 is based on the vector space 203. The vector space 203 is composed of the vector components 102 to 104 constituting the vector, that is, the information rank 102, the information time 103, and the reference value 104. Of the vector components 102 to 104 constituting the vector space 203, the information rank 102 has an X axis, the information time 103 has a Y axis, and the reference value 104 has a Z axis. Accordingly, the multidimensional vector 202, which is a position vector to express the weight of the Internet content, has the keyword 101 from which the content is searched as the starting point 201 of FIG. 2, and the vector components 102 to 104. One of the points of the vector space 203 according to the end point. When the vector components 102 to 104 constituting the vector space 203 are extended to n, they can be expressed as n-dimensional vectors.

The starting point 201 constitutes a query for searching the Internet content as the keyword 101, and represents a search intention for searching the corresponding content. Accordingly, the vector size of the multidimensional vector 202 is an index indicating how close the information is to the search intention. In addition, the function of the indicator indicated by the keyword 101 can be utilized as a categorization function of the fork sonomi.

The utilization of the tag of FIG. 1 includes keyword-based group classification and information weight comparison in the same classification. Group classification can classify information which has in common with respect to the Internet content centering on the said keyword 101. FIG. This can be done by measuring the similarity of existing search engines,

The present invention devises a method for comparing information weights in a classification.

If it is determined that the content belongs to the same category by the keyword 101, the weights in the classification are compared to the vector size to measure the ranking. Referring to the multidimensional vector 202 as an example, any one point a _{1 of the} X axis indicating the information rank 102 among the vector components 102 to 104; Any point a ₂ on the Y axis indicating the information time 103; And the multidimensional vector 202 having any point a ₃ of the Z axis representing the reference value 104 is

. Therefore, the size of the multidimensional vector 202 is

Since the multi-dimensional vector 202 is labeled for each content, the comparison between them is a vector size calculated by the vector components 102 to 104 (

) Can be compared.

On the other hand, the reference value 104, which is referred to as the weight of another person, is used by accumulating the vector size. That is, one point on the Z axis represents a reference value, and is a value in which weights of others are accumulated. Since the tag ( i ) given by others belongs to the whole tag,

The cumulative value V is assumed to be Xi , Yi , and Zi .

It can be represented as. However, the accumulated weight V is used by dividing the total number of accumulated tags. This allows us to introduce objectivity into the information weights that are subjectively given.

) Has the advantage of implementing some sort of normalization.

1 is a block diagram showing a tag definition format using a vector.

FIG. 2 is a conceptual diagram illustrating a vector space according to a multidimensional vector and a vector component used in the tag of FIG. 1.

Claims

In classifying and searching for internet contents, in defining tags to enable search results and intuitive information weight expression in accordance with the search intention,

The tag includes a keyword including a purpose and a field to be searched and implying a search intention;

A method for deriving and utilizing a multi-dimensional vector in a tag definition for Internet content, characterized by a format defined by n vector components of the multi-dimensional vector for labeling information weights based on the keyword.

The method of claim 1,

A component corresponding to a reference value referred to by the weight of another among the n vector components is employed,

A vector size is calculated by the n vector components including the component corresponding to the reference value,

A method of introducing and utilizing a multi-dimensional vector in the definition of a tag for Internet content, characterized in that the function of normalizing the personal information rank of the current content user in measuring the ranking of the information as a result of the sum.