KR20020023737A - A method and system for searching by using Fuzzy Relational Products - Google Patents

Abstract

PURPOSE: A search method using fuzzy relational products and a device thereof are provided to establish similarity relation about each classification and to multiplex the classification of documents or Internet sites. And the method and the device are provided to enable the extension of classification for an ambiguous search word and supply the commonness for the classification and the mobility of classified levels, also variously set values describing error limits to variously link classification systems. CONSTITUTION: A classification item and a search word are inputted from a user. The classification item is considered as a fuzzy set having the search word as an element, and a fuzzy degree is calculated depending on the frequency of the search word for each result included in the classification item. A fuzzy relational products is applied to a basic classification system based on the classification item and the fuzzy degree, to calculate a relation between classification items of the basis classification system with a degree between pre-defined true and false values. An extended similar classification system is calculated according to the calculated degree. A first search result with the search word for the classification item in the basic classification system is extracted from a database. A second search result with the search word for an extended similar classification item corresponding to the classification item in the extended similar classification system is extracted from the database. And the first and the second search result are displayed.

Description

A method and system for searching by using Fuzzy Relational Products}

The present invention relates to a search algorithm for improving search efficiency in an Internet search engine, and more particularly, to a search method and apparatus for forming a similar correlation between Internet search classifications using Fuzzy Relational Products.

It is a "degree of study" that deals with the degree of truth. In other words, it is the study of overcoming the limitations of the Boolean logic of binary (0 or 1), which is true or false. In fuzzy the logic of 0 and 1 is treated as a rather special case.

The idea of fuzzy logic was first developed in the 1960s by Dr. Lotfi Zadeh of UC Berkeley. He was studying the problem of computers understanding natural language, and realized that the absolute expressions of 0 and 1 could not adequately express rich vocabulary. His research, which began with "How pretty is my wife?", Is a computerization of human's natural expression and is also called "computation with word." In other words, when expressing 'pretty, not pretty' and 'pretty,' 0 and 1, it can be expressed as "a little, moderately, a lot ... pretty" about 0.2, 0.5, 0.8. 0.2, 0.5, and so on are referred to as membership grade data. It is also called soft computing because it is distinct from the absolute expressions of 0 and 1.

Natural language actually used in everyday life is not easy to be expressed in the absolute form of 0 or 1 (whether everything in this world can be represented in dichotomous form is a philosophical question worth studying, but actually we Many of the data you want to put on the computer are in the middle, not 0 or 1, even the result of the calculation). Fuzzy logic is an extreme case of true (or 'fact' or the state of a thing), including 0 and 1, but also the trueness of the various states in between.

Fuzzy logic is therefore a study of ambiguity, and the results include some error. Appropriate reasoning came up here, and the concept of alpha-cut is used to define an acceptable limit of error. Fuzzy logic is similar to human thinking process, and there are many philosophical elements, so it may be the logic of liberation that excludes extremities like Buddhism in Buddhism. In other words, it is closer to the Asian way of thinking.

Fuzzy logic seems to be closer to how our brain works. We gather data, make partial numbers of truth, go further, gather higher-order truths, and when they cross some boundaries, they produce the same results as motor-neural responses. Similar processes are used in artificial intelligence computers, neural networks, and expert systems.

The present invention relates to a search algorithm of a search engine, i.e. a search method. Comparing the conventional search methods of domestic and international famous search engines and their respective advantages and disadvantages are as follows.

Search engines can be broadly classified into categorical searches, index searches, and meta searches using multiple search engines at once. A classification search is a method of constructing and searching a directory database, and an index search is a method of indexing and searching a document. A classification search has a relatively high accuracy but a low reproducibility, and vice versa.

The representative search engine using the classification search method is Yahoo. The representative search engine using the index search is Alta Vista and the pure Korean search engine Simmani performs two methods in parallel. Here are some of the advantages and disadvantages of categorical search and index search:

Classification search Index search Characteristic Register and classify websites by taxonomy. When you enter a search term, find a matching word in the categorized taxonomy. The words in the web document are indexed as search terms. It searches for indexed search terms that match the search terms entered and lists the sites that contain the indexed words in order of priority. Advantages Compresses the document into the correct subject, so if the document to be searched can be classified as the correct subject, it is relatively efficient. If you're inaccurate about what topics the document you're looking for is inaccurate, enter the expected topic as a search term and find all the web documents that contain that search term. The search range is wide. Disadvantages If the document you are looking for is difficult to sort by any exact taxonomy, it becomes difficult to find the document. Producing too broad a search results in a search efficiency that is necessary because it is necessary to search documents that are not needed.

As such, the two search methods have complementary aspects. Currently, most search engines do both. This gives you the option to try something else if you aren't satisfied with the results from one method. This is a bit cumbersome and does not combine the advantages of the two methods.

In addition, in the conventional search method, there is a case that a waste of time selecting a desired site due to too many search results during index search or a desired site cannot be found in a classification other than the corresponding classification during the classification search.

Accordingly, an object of the present invention is to establish a relationship for similarity between classifications using fuzzy relational products in classification search, and to increase the reproducibility of classification search by multiplexing the classification of documents or Internet sites. It is to provide a method and apparatus for raising.

It is also an object of the present invention to provide a retrieval method and apparatus using a fuzzy relation method that facilitates retrieval by providing a classification extension for a search term that is ambiguous.

In addition, an object of the present invention is to provide a search method using a fuzzy relation method that can effectively manage the search classification system by providing the shareability of the classification and the fluidity of the classification level for the classification items that can belong to the sub-items of several classification items and To provide a device.

It is also an object of the present invention to provide a retrieval method and apparatus using a fuzzy relation method which can variously link a classification scheme by setting various values defining an error limit in fuzzy theory.

In addition, an object of the present invention is a fuzzy relationship method that solves the specificity of the upper and lower concepts between search classifications (for example, one search classification item becomes a lower concept and a higher concept for another search classification) through providing compatibility. It is to provide a search method and apparatus using the.

1 is a block diagram of a search system by forming a similar correlation between search classifications using a fuzzy relationship method according to an embodiment of the present invention.

Figure 2 is a functional block diagram for forming a similar correlation between search classifications using the fuzzy relationship method according to an embodiment of the present invention.

3 shows a list of results by an application operator in accordance with one preferred embodiment of the present invention.

4 is a diagram illustrating an application of a fuzzy relation method to a search classification word according to an exemplary embodiment of the present invention.

5a and 5b is a view showing the result according to the alpha cut (α-cut) in the application of the fuzzy relationship method of the search classifier according to an embodiment of the present invention.

6 is a diagram illustrating actual classification and extended classification according to a fuzzy relation method according to an embodiment of the present invention.

7 is a diagram illustrating a classification system of extended classification according to a fuzzy relation method according to an embodiment of the present invention.

8 is a flowchart illustrating a search process by forming a similar correlation between search classifications using a fuzzy relationship method according to an exemplary embodiment of the present invention.

<Explanation of symbols for main parts of the drawings>

101: search server 103: fuzzy search module

107: Database Server 109: Search Information Database

111: network 113: user device

115: manager device

According to an aspect of the present invention, in order to achieve the above objects, in the state provided with a basic classification system composed of a plurality of classification items, in providing a search according to the similar correlation between search classifications, the classification items from the user And receiving a search word, considering the classification item as a fuzzy set having the search word as an element, and fuzzy degree according to the frequency of the search word for each search result belonging to the classification item. degree) and apply the fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree to calculate the inclusion relation between the classification items of the basic classification system as a predefined value between true and false values. Calculate an extended similarity classification system according to the calculated degree, A first search result for the search term for the classification item in the classification system, and a second search result for the search term for the extended similar classification item corresponding to the category in the extended similar classification system from the database. In addition, a search method and apparatus and system corresponding to the method may be provided according to a similar correlation between search classifications that output the first search result and the second search result.

Equation corresponding to the fuzzy relation method is ( ◁ ) Where k is an index for a search term, j, m is an index for a taxonomy, R is a column for the taxonomy, and the search term is a row. A matrix with fuzzy values between 0 and 1 indicating the relationship between Is a matrix where the taxon is Row and the search term is Column. Is the total number of queries. Is the degree to which the search term k is included in the taxonomy j. Is the degree to which the search term k is included in the classification word m, and → may be an application operator.

Top and bottom concepts may be applied to the basic classification system according to the inclusion relationship between a plurality of classification systems.

The extended similarity classification scheme may be changed corresponding to the membership function data and alpha-cut applied to the fuzzy relation method.

In order to achieve the above object, according to another aspect of the present invention, in a state provided with a basic classification system consisting of a plurality of classification items, in providing a search according to the similar correlation between the search classification, the user searches for a search word from the user Receives a list of at least one classification item corresponding to the search term and a search result corresponding to the classification item list from a database, and selects a category item having the highest fuzzy degree according to the frequency of the search term from the search result; And applying a fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree to calculate the degree of inclusion between the classification items of the basic classification system as a predefined value between true and false values. Calculate an extended similarity classification system according to the calculated degree, Extracting an extended similar classification item corresponding to the classification item using a chapter similar classification system, extracting a search result for the search word for the classification item and the extended similar classification item from a database, and outputting the search result It is possible to provide a search method according to similar correlation between search classifications, an apparatus and a system corresponding to the method.

Equation corresponding to the fuzzy relation method is ( ◁ ) Where k is the index for the search term, j and m are the indexes for similarly matched taxonomy, and R is the search term as the column and the search term as the row. Is a matrix with a fuzzy value between 0 and 1 indicating the relationship between each classification and a search term. Is a matrix where the taxon is Row and the search term is Column. Is the total number of queries. Is the degree to which the search term k is included in the taxonomy j. Is the degree to which the search term k is included in the classification word m, and → may be an application operator.

In the basic classification system, a concept of upper and lower levels may be applied according to an inclusion relationship between a plurality of classification systems.

The extended similarity classification item may be changed to correspond to the membership function data and alpha-cut applied to the fuzzy relation method.

According to another aspect of the present invention to achieve the above object, in providing a search according to the similar correlation between search classifications through a network, with a basic classification system composed of a plurality of classification items, Receive a classification item and a search word from a user, regard the classification item as a fuzzy set having the search word as a component, calculate a fuzzy degree according to the frequency of the search word for each search result belonging to the classification item, and Based on a classification item and the fuzzy degree, a fuzzy relation method is applied to the basic classification system to calculate the inclusion relationship between the classification items of the basic classification system as a degree between a predefined true value and a false value, and the calculated degree Calculate an extended similarity classification scheme according to the basic classification from the database Extracting a first search result for the search term for the classification item from a system, extracting a second search result for the search term for an extended similar classification item corresponding to the classification item from the extended similar classification system from a database; The present invention may provide a search method according to a similar correlation between search classifications for transmitting the first search result and the second search result to the user, and an apparatus and system corresponding to the method.

According to another aspect of the present invention to achieve the above object, in providing a search according to the similar correlation between search classifications through a network, with a basic classification system composed of a plurality of classification items, Receiving a search word from a user, extracting a list of at least one category item corresponding to the search word and a search result corresponding to each of the category item list from a database, and a fuzzy degree according to the frequency of the search word A high classification item is determined, and a fuzzy relation method is applied to the basic classification system based on the classification item and the fuzzy degree to convert the inclusion relationship between the classification items of the basic classification system into a predefined value between true and false values. Calculate the extended similarity classification system according to the calculated degree. Extracts an extended similar classification item corresponding to the classification item using the extended similar classification system, extracts a search result for the search word for the classification item and the extended similar classification item from a database, and the user A search method, an apparatus and a system corresponding to the method may be provided according to a similar correlation between search classifications for transmitting the search results.

Next, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a search system by forming a similar correlation between search classifications using a fuzzy relationship method according to an exemplary embodiment of the present invention.

Referring to FIG. 1, the search server 101 includes a fuzzy search module 103, and the fuzzy search module 103 provides a search based on a similar correlation between search classifications using a fuzzy relation method. Here, the searching method according to the fuzzy relation method will be described in detail with reference to FIGS. 3 to 8.

The database server 107 coupled with the search server 101 has a search information database 109, which includes a search term table 121, a classification item table 123, and a fuzzy degree table 125. ), And the like.

The search word table 121 stores a search word to be searched through the search server 101, and the classification item table 123 stores a classification item of a basic classification system. In the fuzzy degree table 125, a classification item is regarded as a fuzzy set having a search word as an element, and a fuzzy degree calculated according to the frequency of the search word for each search result belonging to each classification item ( Fuzzy degree is stored.

The manager 119 and the user 115 access the search server 101 through the network 111 by using the manager device 117 and the user device 113 having a web browser capable of internet access to use the search service. Can be.

Fuzzy theory has been proposed as a substitution logic (a logic value between 0 and 1) to reflect the uncertain real logic of the crisp logic value defined by Zadeh as [0, 1]. This uncertainty is due not only to the variability of the logic subject and environment, but also to the unreality of the crisp logic values of 0 and 1 of the logic itself. Fuzzy theory is intended to try to apply real logic to real life by destroying the concept of crisp of logic and having a value between 0 and 1.

This fuzzy theory was developed by zadeh after seeing the light of the world, and many studies were developed especially by mathematicians, and it was applied to Japan in real life. Fuzzy theory was introduced as a way to increase the sharpness when shooting moving objects from cameras and has since spread to artificial intelligence washing machines, refrigerators and air conditioners. The heat is not as cool now, but there is no doubt that it is a key part of AI algorithms.

Fuzzy theory may also be applied to relational or object-oriented databases. It can be used to set the uncertainty of a relationship in a relational database, and it can be used to set the ambiguity of a class hierarchy structure.

Perhaps the most widely used field of fuzzy theory is in establishing relations between objects. It is usually described as a containment relationship and the fuzzy set, Relational Products, explains this relationship well.

Each classification is regarded as a fuzzy set of search terms as a component in a preset basic classification scheme. The fuzzy degree is determined according to the frequency of search terms among the documents or site summary words belonging to the classification. Based on this, the fuzzy relation method between the classification and the classification is applied to express the inclusion relation as between 0 and 1.

The search method in this system can be found in the classification system display window when the classification of the document to be searched is known, and it can be extended to the similar classification when there is nothing to be found in the classification. If the classification is ambiguous, enter a search term and select a category having the highest degree of classification among the categories associated with the search term. If it does not find it in the classification, it extends to the similar classification as well.

2 is a functional block diagram for forming a similar correlation between search classifications using a fuzzy relationship method according to an exemplary embodiment of the present invention.

Referring to FIG. 2, the functions of the fuzzy search module 103 are described. The search condition 201 input or received from the user 115 is transmitted to the database processing and fuzzy search unit 207. Here, the search condition may be a category item and a search word. The database processing and fuzzy retrieval unit 207 performs a retrieval process of a search word for a classification item from the retrieval information database 109 and outputs a retrieval result 203.

The database processing and fuzzy search unit 207 extracts the search term information, the classification item information and the fuzzy degree information between each search word and the classification item stored in the search information database 109 and applies the classification item to the fuzzy relation method to expand the similarity. Calculate the classification item. Then, a search process is performed on each of the extended similar classification items, and a search result 203 is output.

3 is a view showing a result list by an application operator according to an exemplary embodiment of the present invention.

Referring to FIG. 3, in the general concept of set, inclusion of set A of component x is included. Is represented by a value of 0 or 1 in the Crisp logic. But in fuzzy set A Has a logical value between 0 and 1.

The crisp implication operator has the following values:

0 One 0 0 One One One One

By applying the concepts of fuzzy set, fuzzy power set, and crisp implication operator, With the domain We need a binary function that produces a result in the range of. These functions are called implication operators and can be expressed in various forms for each individual environment. There are hundreds of implication operators known to date. The representative ones are as follows.

KD. The Kleene-Diense systems and operator

Implication operator is a function that sets the logical value for fuzzy implication. From this we can express the two sets of inclusion phase relationships for all components of the universal set as values between 0 and 1, as shown in Table 301 (Degree of membership).

4 is a diagram illustrating an application of a fuzzy relation method to a search classification word according to an exemplary embodiment of the present invention.

Referring to FIG. 4, a fuzzy relation method will be described first to describe a fuzzy relation method in a search classification word.

The fuzzy relationship method was first developed by Bandler and Kohout and was applied to establish the relationship between each patient by observing psychological symptoms of Parkinson's patients at London Hospital to address the real-world possibilities.

The equation calculated by the fuzzy relation method is as follows.

( ◁ ) Equation 1)

In general, search taxonomy is associated as a containment relationship of parent and child concepts. In the Crisp environment, subcategories of subcategories exist in some classifications, such as the library's book classification. Therefore, each of the other subclassifiers included in the subcategories of the subordinate concept includes all of the superclassifiers. However, depending on the subjective view of the actual search subject, the relationship between the top and bottom may be somewhat different. From the point of view of the public, the Internet can be thought of as the area of web pages searched through web browsers. However, from an expert point of view, the term Internet refers to communication protocols such as FTP, Gopher, NEWS, and Email, as well as the World Wide Web.

In order to deal with the ambiguity, uncertainty, and variability of inclusion relations, the above-described fuzzy relational products are implemented to implement a flexible classification scheme that varies according to the search environment and the search subject.

The first thing to consider is to establish a relationship that exists between the taxonomy and the search terms. In the existing search engine, there is no search word correlated with the classification other than the classification word itself. In the present invention, each category is associated with a certain number of search terms as a degree having a value between 0 and 1.

A search classification method using Equation 1) in the fuzzy relation method will be described below.

home)

1) K1, K2, K3, K4, K5: Keyword

2) C1, C2, C3, C4, C5: Fuzzy set using the search word as an element

3) R: A matrix representing the relationship between each classification and the search word, with the classification word as the column and the search word as the row, and having a fuzzy value between 0 and 1.

4) : Matrix with categorical terms as rows and search terms as columns

The following equation is calculated through the assumption and equation 1).

( ◁ ) Equation 2)

: Total number of queries

: The degree to which the search term k is included in the classification term j

: The degree to which the search term k is included in the classifier m

→: implication operator

Equation 2) represents the degree to which the m-th classification set includes search terms of the j-th classification set.

Yes) ( ◁ ) , ( ◁ )

Then, the relationship between the classifications may be set as 401, 403 and 405 by the fuzzy relationship method.

Here, alpha-cut is a reference value for converting the final result into a crisp value. In other words, when alpha cut is 1 (407), the relational products below 0.8 are 0 and above 0.8 are 1, and when alpha cut is 0.8 (409), the relational products below 0.8 are 0 and above 0.8 Becomes 1

5A and 5B are diagrams illustrating results according to an alpha-cut in applying a fuzzy relation method of a search classification word according to an exemplary embodiment of the present invention.

Referring to FIGS. 5A and 5B, the configuration diagram showing the relation between each search classification word as the value of the final relational products obtained as shown in FIG. 4 is equal to 501 when the alpha cut is 1, and the alpha cut is 1. If it is 0.8, it is equal to 503.

Looking at the inclusion relationship between search categories when the alpha cut is 1 (501), the C1 classification item is a subclass of all classification items, and C3 and C5 are subclasses of C2 and C4, respectively.

Looking at the inclusion relationship between search categories when the alpha cut is 0.8 (503), C4 is located at the highest category, and C1 and C3 are located at the lowest category. It can be seen that the alphacut is extended while including all of the classification relations when 1.

Here, we can find the following specificities that differ from the general search classifications, as seen at 501 and 503.

1) Sharedness of classification: At 501, C1 is composed of C2, C3, C4, and C5 subcategories simultaneously. Instead of the exclusive concept of general search taxonomy, the concept of shared, multi-inheritance is introduced. In other words, sites classified by the search classifier C1 share C2, C3, C4, and C5 as upper search classifiers.

2) Liquidity of classification level: At 501, C1 has liquidity that can be located in several classification layers. If the search path is C2-> C3-> C1, it belongs to the third classification hierarchy, but if the search path is C2-> C1, it belongs to the second classification hierarchy.

3) Two or more representations of one semantic: at 503, another specificity is found. As shown in the figure, C1 and C3 are merely expressed differently as search classifications with the same meaning. In other words, C1 and C3 have the same classification.

4) Classification compatibility: Another peculiarity shown in 503 is that C2 and C5 are both subordinate and mutually superior. However, C2 and C5 are not representations with the same semantics as the relationship between C1 and C3. C5 is a subclass of C1, C3, but C2 is not. This situation is similarly explained in the relationship between C1, C3 and C5.

5) Subjectivity of the classification system: As shown in 501 and 503, the classification system can vary depending on the alphacut value. Depending on the search environment and the search subject, alphacut values can be linked.

6 is a diagram illustrating actual classification and extended classification according to a fuzzy relation method according to an exemplary embodiment of the present invention.

7 is a diagram illustrating a classification system of extended classification according to a fuzzy relation method according to an embodiment of the present invention.

6 and 7, the inclusion relationship between the search classifications A, B, C, D, E, F, and G is indicated by a solid line at 601, and the extension classification relationship according to the fuzzy relationship method is dotted. Let's say

That is, the extended classification relationship according to the fuzzy relationship method between A, B, C, D, E, F, and G may be expressed as 701.

8 is a flowchart illustrating a search process by forming a similar correlation between search classifications using a fuzzy relationship method according to an exemplary embodiment of the present invention.

Referring to FIG. 8, when a classification item and a search word are received from the user 115 (S801), a lower classification item corresponding to the classification item is extracted from the search information database 109 (S803). In operation S805, similar (expanded) classification items for each of the classification items and the lower classification items are detected according to the fuzzy relation method. When the search classification item and the extended similar classification item are detected in this way, a search process for the search word is performed for each classification item (S807).

Specifically, the search order through the search system supporting the fuzzy relation method is as follows.

From the user interface screen in the tree view form, expand the search tree until the desired search classification is displayed or enter the desired search classification word in the search word input field and hit the return key. The search tree will then automatically develop the classification process.

Each time you select a category from the tree, a summary of the sites that belong directly to that category may be displayed on a separate screen. Each time you select a summary site, you switch to that site and can automatically modify statistical data about how many times you visited or selected the site.

If the site you want isn't in that category or a subcategory, select the Expand to similar category option. The extended similarity classification item according to the fuzzy relation method may be detected with respect to the extended classification selection of the user 115 to process a search in the corresponding extended similarity classification item.

The present invention is not limited to the above embodiments, and many variations are possible by those skilled in the art within the spirit of the present invention.

As described above, according to the present invention, by establishing a relation for similarity between classifications by using the Fuzzy Relational Products in classification search, and by increasing the reproducibility of classification search by multiplexing the classification of documents or Internet sites It is possible to provide a method and apparatus for increasing the search efficiency.

In addition, according to the present invention, it is possible to provide a search method and apparatus using a fuzzy relation method that facilitates a search by providing a classification extension for a keyword whose classification is ambiguous.

In addition, according to the present invention, a search method and apparatus using a fuzzy relation method that can effectively manage the search classification system by providing the shareability of the classification and the fluidity of the classification level for the classification items that can belong to the lower items of the various classification items Can be provided.

In addition, according to the present invention, it is possible to provide a retrieval method and apparatus using a fuzzy relation method that can variously link the classification scheme by setting various values defining the error limit in the fuzzy theory.

In addition, according to the present invention, there is provided a fuzzy relation method for solving specificity (eg, one search classification item becomes a lower concept and a higher concept with respect to another search classification) by providing a compatibility between search classifications. It is possible to provide a search method and apparatus using the same.

In addition, according to the present invention, it is possible to provide a search method and apparatus using a fuzzy relation method which is used as a basic model of an intelligent classification method as well as a search engine.

In addition, according to the present invention, by combining a search method that combines the advantages of the classification search and the index search with an excellent graphical user interface (GUI) environment in the form of a tree view to provide a user convenience A search method and an apparatus using a fuzzy relation method can be provided.

Claims (16)

In the method having a basic classification system composed of a plurality of classification items, providing a search by forming a similar correlation between search classifications, Receiving a category and a search word from a user; Considering the classification item as a fuzzy set having the search word as an element, and calculating a fuzzy degree according to the frequency of the search word for each search result belonging to the classification item. ; Calculating a inclusion relationship between predefined true and false values by applying a fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree; Calculating an extended similarity classification system according to the calculated degree; Extracting a first search result for the search word for the classification item from the database in the basic classification system; Extracting a second search result for the search word for the extended similar classification item corresponding to the classification item in the extended similar classification system from a database; And Outputting the first search result and the second search result Search method according to the similar correlation between the search classification comprising a. In claim 1, Equation corresponding to the fuzzy relation method ( ◁ ) This, Where k is the index for the search term, j and m are classification fuzzy sets using the search terms as elements, R is a matrix with fuzzy values between 0 and 1 representing the relationship between each classification and the search terms, with the taxonomy column as the column, and the search term as the row. Is a matrix where the taxon is Row and the search term is Column. Is the total number of queries. Is the degree to which the search term k is included in the taxonomy j. Is the degree to which the query k is included in the classifier m. → is the application operator Searching method according to the similarity correlation between search classifications. In claim 1, The basic classification scheme The concept of parenting being applied according to the inclusion relationship between multiple classification systems Searching method according to the similarity correlation between search classifications. In claim 1, The extended similar classification system Change corresponding to the membership function data and alpha-cut applied to the fuzzy relation method Searching method according to the similarity correlation between search classifications. In the method having a basic classification system composed of a plurality of classification items, providing a search by forming a similar correlation between search classifications, Receiving a search word from a user; Extracting at least one category item list corresponding to the search term and a search result corresponding to the category item list from a database; Selecting a classification item having the highest fuzzy degree according to the frequency of the search word in the search result; Calculating a degree of inclusion between a classification item of a basic classification system as a degree between a predefined true value and a false value by applying a fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree. ; Calculating an extended similarity classification system according to the calculated degree; Extracting an extended similar classification item corresponding to the classified item by using the extended similar classification system; Extracting a search result for the search word for the classification item and the extended similarity classification item from a database; And Outputting the search results Search method according to the similar correlation between the search classification comprising a. In claim 5, Equation corresponding to the fuzzy relation method ( ◁ ) This, Where k is the index for the search term, j and m are classification fuzzy sets using the search terms as elements, R is a matrix with fuzzy values between 0 and 1 representing the relationship between each classification and the search terms, with the taxonomy column as the column, and the search term as the row. Is a matrix where the taxon is Row and the search term is Column. Is the total number of queries. Is the degree to which the search term k is included in the taxonomy j. Is the degree to which the query k is included in the classifier m. → is the application operator Searching method according to the similarity correlation between search classifications. A search method according to the formation of a similar correlation between search classifications. In claim 5, The basic classification scheme The concept of parenting being applied according to the inclusion relationship between multiple classification systems Searching method according to the similarity correlation between search classifications. In claim 5, The extended similar classification category Change corresponding to the membership function data and alpha-cut applied to the fuzzy relation method Searching method according to the similarity correlation between search classifications. In the method having a basic classification system composed of a plurality of classification items, providing a search according to the similar correlation between search classifications through a network, Receiving a category and a search word from a user; Considering the classification item as a fuzzy set having the search word as a component, and calculating a fuzzy degree according to the frequency of the search word for each search result belonging to the classification item; Calculating a inclusion relationship between predefined true and false values by applying a fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree; Calculating an extended similarity classification system according to the calculated degree; Extracting a first search result for the search word for the classification item from the database in the basic classification system; Extracting a second search result for the search word for the extended similar classification item corresponding to the classification item in the extended similar classification system from a database; And Transmitting the first search result and the second search result to the user. Search method according to the similar correlation between the search classification comprising a. In the method having a basic classification system composed of a plurality of classification items, providing a search according to the similar correlation between search classifications through a network, Receiving a search word from a user; Extracting at least one category item list corresponding to the search term and a search result corresponding to each of the category item list from a database; Determining a classification item having the highest fuzzy degree according to the frequency of the search word in the search result; Calculating a inclusion relationship between predefined true and false values by applying a fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree; Calculating an extended similarity classification system according to the calculated degree; Extracting an extended similar classification item corresponding to the classified item by using the extended similar classification system; Extracting a search result for the search word for the classification item and the extended similarity classification item from a database; And Sending the search results to the user Search method according to the similar correlation between the search classification comprising a. In the apparatus having a basic classification system composed of a plurality of classification items, to provide a search for forming a similar correlation between search classifications, Means for receiving a category and a search word from a user; A means for considering the classification item as a fuzzy set having the search word as an element, and calculating a fuzzy degree according to the frequency of the search word for each search result belonging to the classification item. ; Means for calculating a inclusion relationship between a predefined true value and a false value by applying a fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree; Means for calculating an extended similarity classification scheme according to the calculated degree; Means for extracting a first search result for the search term for the classification item from the database from the database; Means for extracting a second search result for the search term from the database for an extended similarity classification item corresponding to the classification item in the extended similarity classification system; And Means for outputting the first search result and the second search result Search apparatus according to the forming a similar correlation between search classification having a. In the apparatus having a basic classification system composed of a plurality of classification items, to provide a search for forming a similar correlation between search classifications, Means for receiving a search word from a user; Means for extracting at least one category item list corresponding to the search term and a search result corresponding to the category item list from a database; Means for selecting a classification item having the highest fuzzy degree according to the frequency of the search word in the search result; Means for calculating a containment relationship between classification items of a basic classification system as a degree between a predefined true value and a false value by applying a fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree. ; Means for calculating an extended similarity classification scheme according to the calculated degree; Means for extracting an extended similar classification item corresponding to the classification item using the extended similar classification system; Means for extracting search results for the search term for the classification item and the extended similarity classification item from a database; And Means for outputting the search results Search apparatus according to the forming a similar correlation between search classification having a. In the apparatus having a basic classification system composed of a plurality of classification items, to provide a search for forming a similar correlation between search classifications through a network, Means for receiving a classification item and a search word from a user; Means for considering the classification item as a fuzzy set having the search word as a component and calculating a fuzzy degree according to the frequency of the search word for each search result belonging to the classification item; Means for calculating a inclusion relationship between a predefined true value and a false value by applying a fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree; Means for calculating an extended similarity classification scheme according to the calculated degree; Means for extracting a first search result for the search term for the classification item from the database from the database; Means for extracting a second search result for the search term from the database for an extended similarity classification item corresponding to the classification item in the extended similarity classification system; And Means for sending the first search result and the second search result to the user Search apparatus according to the forming a similar correlation between search classification having a. In the apparatus having a basic classification system composed of a plurality of classification items, to provide a search for forming a similar correlation between search classifications through a network, Means for receiving a search term from a user; Means for extracting at least one category item list corresponding to the search term and a search result corresponding to each of the category item list from a database; Means for determining a classification item having the highest fuzzy degree according to the frequency of the search word in the search result; Means for calculating a inclusion relationship between a predefined true value and a false value by applying a fuzzy relation method to the basic classification system based on the classification item and the fuzzy degree; Means for calculating an extended similarity classification scheme according to the calculated degree; Means for extracting an extended similar classification item corresponding to the classification item using the extended similar classification system; Means for extracting search results for the search term for the classification item and the extended similarity classification item from a database; And Means for sending the search results to the user Search apparatus according to the forming a similar correlation between search classification having a. A program of instructions, which can be executed by a digital processing apparatus, is tangibly implemented in order to perform a searching method according to a similar correlation between the search classifications according to at least one of claims 1 to 8. The recording medium which can be read by. A memory in which a program is stored; A processor coupled to the memory to execute the program; The processor is a search system according to the similar correlation between the search classifications to execute the search method according to the similarity between the search classifications according to at least one of claims 9 to 10 by the program.