KR100998291B1 - Method and apparatus for structuralize keyword string and searching keyword string - Google Patents

Abstract

The present invention relates to a method and apparatus for structuring a keyword string using a keyword management server and detecting a structured keyword string. The present invention relates to a keyword management server after structuring and storing a keyword string to be structured. Provided are a method and an apparatus capable of detecting the structured keyword string in a sentence input to a server.

Keyword, structured, detected, character, structured data

Description

METHOD AND APPARATUS FOR STRUCTURALIZE KEYWORD STRING AND SEARCHING KEYWORD STRING}

The present invention relates to a method and apparatus for structuring and detecting a keyword string using a keyword management server. More particularly, the present invention relates to a keyword management server after structuring and storing a keyword string to be structured. A method and apparatus for quickly and accurately detecting the structured keyword string in an input sentence.

Various conventional techniques exist with regard to the method of detecting a particular keyword in a sentence, document or data. The keyword detection method may be used in various fields such as text communication between mobile communication networks and search engines in Internet portal sites. Looking at the prior art associated with such a keyword detection as follows.

1. Method and system for synonym extraction and document retrieval (application number 10-2000-44283, hereinafter referred to as prior art 1). Prior art 1 is an invention characterized by extracting similar words by quantifying each word into first and second feature vectors, storing it in a database, and comparing the searched document or string with a stored word vector. However, there is a problem that such string vectorization may be subjective when there is no objective vectorization criterion, and there is a problem that the vectorization itself takes a lot of time.

2. Information retrieval method and information retrieval apparatus (application number 10-2000-41311, hereinafter referred to as prior art 2). Prior art 2 aims at preprocessing of a natural language system for retrieval omission and noise reduction, but there is a problem in that a preprocessing procedure such as data storage means, text segmentation means and word segmentation text is complicated.

3. An information retrieval apparatus for searching for a character string matching a keyword (application number 10-1997-9211, hereinafter referred to as prior art 3). Prior art 3 is characterized in that it is indexed in the order of the occurrence of the character form, and converted into a pattern, so that a high-speed search is possible.

4. Natural language processing technology and its application based on parsing (application number 10-2000-58702, hereinafter referred to as prior art 4). Prior art 4 is a Korean natural language processing technique based on syntax analysis, and aims at natural language processing in the form of grasping a collective dependency relationship between words. However, the prior art 4 has a problem that the efficiency of the operation itself is inferior because it takes a considerable operation, such as handling similar words / parts of speech.

Since the retrieval method shown in the above-described prior arts is for natural language processing and pattern recognition, there have been problems in that the efficiency of the search itself is inferior and many operation processes are required in preprocessing for the search itself. There is a need for a keyword structuring method and a search method that can solve the problem.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a method and apparatus for structuring a keyword string input as a structured object from a keyword management server.

In addition, the present invention provides a more accurate and efficient keyword detection method and apparatus by directly detecting a keyword using code value information corresponding to each character when the keyword is detected in a predetermined sentence input to the keyword management server. For another purpose.

In addition, the present invention can directly access the keyword using the code value information corresponding to each character when the keyword is detected, it is possible to detect the keyword included in the data when using the text service or data communication using the mobile communication network It is another object to provide a method and apparatus that can be quickly and accurately.

According to an aspect of the present invention, there is provided a method of structuring a keyword string input as a structured object in a keyword management server, the method comprising: first of at least one character constituting the keyword string input as the structured object; A first step of recognizing a character; A second step of searching for keyword-related information set corresponding to the recognized first character in the first structured data including preset keyword-related information for each of the characters capable of constructing the keyword string, wherein the The keyword related information includes character correspondence code value information, character correspondence information, and address information of substructured data set for each of the above characters; Updating character correspondence information included in the retrieved keyword related information; And a position for storing the substructured data stored by the address information of the substructured data included in the keyword related information of the character corresponding to the character correspondence information updated in the previous step when the keyword string is composed of two or more characters. Step), wherein new substructured data is generated when address information of the substructured data does not exist, and location information of the generated new substructured data is stored in the address information of the substructured data; B) searching for the keyword related information set corresponding to the next order character of the character corresponding to the character corresponding information updated in the previous step in the identified substructured data, and updating the character corresponding information included in the retrieved keyword related information. Step to the last character of the keyword string It provides a repeat first keyword string structure comprises the step 4 to.

Here, the character correspondence information may include whether a keyword string ending with the character exists and the number of keyword strings following the character.

Here, the substructured data may include keyword related information preset for each character.

Here, when the character is Korean, the character correspondence code value information may be set using KSC 5601 Korean code.

According to another aspect of the present invention, in a method for detecting a keyword string from a sentence input as a detection target in a keyword management server using structured data structured by the keyword string by the above-described method, the sentence input as the detection target Recognizing a first character among at least one or more characters constituting the first step; A second step of retrieving keyword related information set corresponding to the recognized first character from the first structured data including keyword related information preset for each of all the characters, wherein the keyword related information is set to all the characters. Including character correspondence code value information, character correspondence information, and address information of substructured data set for each of the two; A third step of detecting a detectable keyword string by analyzing character correspondence information included in the searched keyword related information, and determining whether to move to substructured data; A fourth step of terminating a keyword string detection process if it is determined by the third step not to move to substructured data; And if it is determined by the third step to move to the substructured data, move to the substructured data, and the next order of the character corresponding to the character correspondence information analyzed in the previous step among the input sentences in the moved substructured data. A) searching for keyword related information set according to the character; And analyzing the character correspondence information included in the searched keyword related information to detect a detectable keyword string, and repeating step b) of determining whether to move to substructured data until it is determined not to move to substructured data. It provides a keyword string detection method comprising a fifth step to

Here, when it is determined not to move to the substructured data by the fourth step or the fifth step, after recognizing the next order character of the character corresponding to the character correspondence information analyzed in the previous step, the second to fifth steps It may be characterized by performing a step.

Here, the character correspondence information may include whether a keyword string ending with the character exists and the number of keyword strings following the character.

Here, the substructured data may be configured to include preset keyword related information for each of all characters.

Here, when the character is Korean, the character correspondence code value information may be set using KSC 5601 Korean code.

According to still another aspect of the present invention, in the keyword string structuring apparatus, which is included in a keyword management server to structure a keyword string input as a structured object, the keyword string structured device constitutes a keyword string input as the structured object. Recognizing a first character among at least one or more characters, and corresponding to the first character recognized in the first structured data comprising the keyword-related information preset for each of the characters that can constitute a keyword string Search for the set keyword related information, wherein the keyword related information includes character corresponding code value information, character corresponding information, and address information of sub-structured data set for each of all the characters; included When the character correspondence information is updated and the keyword string is composed of two or more characters, the substructured data is stored by the address information of the substructured data included in the keyword related information of the character corresponding to the character correspondence information updated in the previous step. A) step of grasping the location information, wherein new sub-structured data is generated when the address information of the sub-structured data does not exist, and the location information of the generated new sub-structured data is converted into the address of the sub-structured data. Information stored in the information and keyword related information set corresponding to the next order character of the character corresponding to the character correspondence information updated in the previous step is searched in the identified substructured data, and the character correspondence included in the searched keyword related information is stored. B) updating the information, the keyword It provides a structured keyword string and wherein repeating the last character in Tring.

According to still another aspect of the present invention, there is provided a keyword string detection apparatus which is included in a keyword management server and detects a keyword string from a sentence input as a detection target by using the structured data structured by the apparatus of claim 10. The apparatus for detecting a keyword string may include: a first structure configured to recognize a first character among at least one or more characters constituting a sentence input as the detection target, and to include keyword-related information preset for each character; Search for keyword-related information set corresponding to the recognized first character in the data, wherein the keyword-related information includes character-corresponding code value information, character-corresponding information, and address information of substructured data set for each of all the characters. Contains -Detecting keyword strings by analyzing character correspondence information included in the searched keyword related information, determining whether to move to substructured data, and determining not to move to substructured data. If it is determined to move to the substructured data, and move to the substructured data, and corresponds to the next order character of the character corresponding to the character correspondence information analyzed in the previous step of the input sentence in the moved substructured data A) retrieving the set keyword related information, and analyzing the character correspondence information included in the searched keyword related information to detect a detectable keyword string, and determining b) whether to move to substructured data. By not moving to structured data Provided is a keyword string detection apparatus characterized by repeating until determined.

According to the present invention, a method and apparatus capable of structuring a keyword string input as a structured object in a keyword management server can be provided.

In addition, according to the present invention, when detecting a keyword in a predetermined sentence input to the keyword management server, a more accurate and efficient keyword detection method and apparatus by directly detecting the keyword using the code value information corresponding to each character. Can provide.

In addition, according to the present invention, when the keyword is detected, direct access to the keyword is possible by using code value information corresponding to each character, and the keyword included in the data when using a text service or data communication using a mobile communication network. It is possible to provide a method and apparatus that can make detection quickly and accurately.

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

1 is a diagram schematically illustrating a process of structuring a keyword string using a keyword management server and detecting a structured keyword string according to an embodiment of the present invention.

Referring to FIG. 1, the keyword management server 10 according to the present embodiment first performs a process of structuring a keyword string to be detected. Thereafter, when a sentence to be detected is input to the keyword management server 10, the keyword management server 10 performs a process of detecting the structured keyword string in the input sentence.

The keyword management server 10 according to an embodiment of the present invention is for quickly and efficiently detecting a specific keyword string in a binary string or document data. The keyword management server 10 is a text communication using a wired or wireless communication network. It can be used for various purposes, such as a search engine or an application program using a character string on an Internet site. In a specific embodiment of the present invention, the keyword management server 10 will be assumed and described as a general server. However, this is only one embodiment, and various forms, for example, a program installed in a computer or a semiconductor chip installed in a mobile phone, may be used depending on the intended use. It may be present in the form of such that will be apparent to those skilled in the art.

Before describing keyword string structuring and keyword string detection in the keyword management server 10 in detail with reference to FIGS. 2 to 7, the definition of each term used in the embodiment of the present invention will be described first.

The keyword string means a word to be structured or detected, that is, a predetermined word having a meaning. For example, in Korean, words such as “thank you”, “love”, and “cell phone” may be keyword strings, and in English, words such as “apple” and “love” may exist as keyword strings.

The character means each syllable or character constituting the keyword string. For example, three characters of "hand", "de", and "phone" exist in the keyword string of "cell phone" described above, and "l" for "love". There are four characters "o", "v" and "e".

The first structured data is a term defined separately in the present invention and is data set to include keyword related information preset for all characters. The structured data includes keyword related information for all characters. For example, when using the KSC 5601 Hangul code, keyword related information for each of 2350 Hangul codes is included. A configuration example of the structured data will be described in detail later with reference to FIG. 5. The first structured data is the uppermost concept, and keyword-related information corresponding to each character included in the first structured data may be linked with a lower structured level corresponding thereto.

Substructured data refers to structured data that depends on a specific character. For example, there may be sub-structured data that depends on keyword related information corresponding to the character "sensation" included in the first structured data, that is, subordinate to "sensation". Expressed as substructured data of the character.

The keyword related information refers to information which is set in advance for each character and is characterized by including information related to each character. The keyword related information includes each character correspondence code value, character correspondence information, and address information of substructured data.

Character corresponding code value information Means a code value corresponding to each character. For example, when using the KSC 5601 Korean code, code value information corresponding to "A" is represented by "0XB0A0". Each character has code value information corresponding to each character, and the code value information functions as a memory address for searching for the corresponding character in the structured data. 2 to 7 described later, although the above-described KSC 5601 Korean code is used as the character corresponding code value information, it should be noted that the code value information may be various conventional code value information.

The character correspondence information includes detailed information about each character. For example, the character correspondence information may be configured to include whether there is a keyword string ending with the character and the number of keyword strings following the character.

The address information of the substructured data refers to structured data that is dependent on each keyword related information. The address information of the substructured data is stored in the location information of the substructured data and can be moved to the substructured data.

2 is a block diagram showing a detailed configuration of a keyword management server according to an embodiment of the present invention.

Referring to FIG. 2, the keyword management server 10 according to the present embodiment includes a keyword string structure device 11 and a keyword string detection device 12. The keyword string structuring device 11 and the keyword string detection device 12 are connected to each other to share information.

The keyword string structuring apparatus 11 is a means for structuring a keyword string input as an object of structure. The keyword string structuring apparatus first recognizes a first character among at least one or more characters constituting the keyword string input as the structured object, and sets keyword-related information preset for each of the characters that may constitute the keyword string. The keyword related information set corresponding to the recognized first character is searched for in the first structured data configured to include. Here, the keyword related information may include character correspondence code value information, character correspondence information, and address information of substructured data for each of the characters.

Next, when the keyword string is composed of only one character, the character correspondence information included in the searched keyword related information is updated. When the keyword string is composed of two or more characters, the character correspondence included in the searched keyword related information is updated. A) step of recognizing the location information in which the substructured data is stored by updating the information, and the address information of the substructured data included in the keyword related information of the character corresponding to the character correspondence information updated in the previous step; If the address information of the structured data does not exist, create new sub-structured data, store the position information of the generated new sub-structured data in the address information of the sub-structured data, and correspond to the character updated in the previous step. The next character of the character corresponding to the information. The keyword string is structured by retrieving correspondingly related keyword related information from the identified substructured data and updating the character correspondence information included in the retrieved keyword related information to the last character of the keyword string. do.

The keyword string detection device 12 is a means for detecting a keyword string structured by the keyword string structuring device 11 in a predetermined sentence input as a detection object. The keyword string detecting apparatus firstly recognizes a first character among at least one or more characters constituting a predetermined sentence input as the detection target, and includes a first keyword-related information for each of all characters. The keyword related information set corresponding to the recognized first character is retrieved from the structured data. Here, the keyword related information may include character correspondence code value information and character correspondence information for each character.

Next, character correspondence information included in the searched keyword related information is analyzed to detect a detectable keyword string and determine whether to move to substructured data. Here, if it is determined not to move to the substructured data, the keyword string detection process is terminated, and if it is determined to move to the substructured data, it moves to the substructured data, and the previous step among the input sentences in the moved substructured data. A) searching for keyword related information set corresponding to the next order character of the character corresponding to the character corresponding information analyzed in step a); And analyzing the character correspondence information included in the searched keyword related information to detect a detectable keyword string, and repeating step b) of determining whether to move to substructured data until it is determined not to move to substructured data. By doing so, keyword strings are detected.

3 is a flowchart illustrating a process of structuring a keyword string input as a structured object in a keyword management server according to an embodiment of the present invention.

Referring to FIG. 3, first, a keyword string to be structured is input to the keyword management server 10 (S100). The keyword string may be composed of at least one character, and the keyword string may be configured in various languages such as Korean and English. When the keyword string to be structured is input, the keyword management server 10 recognizes the first character of the input keyword string (S110). Next, the keyword management server 10 sets keyword related information set corresponding to the first character recognized in the first structured data configured to include preset keyword related information for each of the characters that may constitute the keyword string. Search for (S120). In this case, the keyword related information may be configured to include character correspondence code value information, character correspondence information, and address information of substructured data for each character.

In the next process, the keyword management server 10 updates the character correspondence information included in the searched keyword related information (S130). The character correspondence information may include whether there is a keyword string ending with the character and the number of keyword strings following the character. Updating the character correspondence information in step S130 eventually means updating various kinds of information included in the character correspondence information. After updating the character correspondence information included in the keyword related information, the keyword management server 10 divides the process according to whether or not the number of characters of the keyword string to be structured is two or more (S140). When the keyword string consists of only one character, the keyword string structure is terminated by updating the character correspondence information included in the searched keyword related information.

When the keyword string is composed of two or more characters, the character correspondence information included in the searched keyword related information is updated, and next, the substructured data included in the keyword related information of the character corresponding to the character correspondence information updated in the previous step is updated. A) step of identifying the location information where the substructured data is stored by the address information, wherein new substructured data is generated when the address information of the substructured data does not exist, and Storing the location information in the address information of the substructured data; and searching for the keyword related information set corresponding to the next order character of the character corresponding to the character correspondence information updated in the previous step in the identified substructured data. Characters included in searched keyword related information Response to the step b) for updating the information repeatedly performed up to the last character of the keyword string (S150 to S180). That is, each character of the keyword string input as the structured object is structured in the structured data corresponding thereto in order. After completing the above-described process, the structure of the keyword string input to the keyword management server 10 ends.

4 is a flowchart illustrating a process of detecting a keyword string structured by the method described with reference to FIG. 3 in a sentence input as a detection target in a keyword management server according to an embodiment of the present invention.

Referring to FIG. 4, first, a predetermined sentence to be detected is input to the keyword management server (S200). After inputting the sentence, the keyword management server 10 recognizes the first character among at least one or more characters constituting the predetermined sentence (S210). Next, the keyword related information set corresponding to the recognized first character is searched for in the first structured data configured to include the keyword related information preset for each character (S220). Here, the keyword related information is preferably configured to include character correspondence code value information, character correspondence information, and address information of substructured data for each character. The keyword management server 10 that has searched for keyword related information analyzes character correspondence information included in the searched keyword related information to detect a detectable keyword string.

Since the character correspondence information includes information on whether there is a keyword string ending with the character, the keyword string ending with the character can be detected by referring to the information. Next, it is determined whether to move to the substructured data (S240). Here, the sub-structured data is preferably configured to include keyword-related information preset for each of the characters as in the first structured data, and the sub-structured data exists so as to be dependent on the keyword-related information preset for each character. It is preferable. In addition, it is preferable that the address information of the substructured data included in the keyword related information for each character is stored as a pointer PTR including position information for searching for the substructured data depending on the character. At this time, the reference position information (reference code value information) of the substructured data is stored in the pointer.

If it is determined that the character correspondence information is not moved to the substructured data, the keyword string detection process is terminated. Here, the case where it may be determined not to move to the substructured data may be, for example, the case where it is determined that there is no keyword string that can no longer lead to the corresponding character. If it is determined that the character correspondence information is moved to the substructured data, the character correspondence moves to the substructured data, and the next character of the character corresponding to the character correspondence information analyzed in the previous step among the input sentences in the moved substructured data is moved. A) searching for keyword related information set according to the order character; And analyzing the character correspondence information included in the searched keyword related information to detect a detectable keyword string, and repeating step b) of determining whether to move to substructured data until it is determined not to move to substructured data. It performs by (S250 and S260).

If it is determined through the above process not to move to the substructured data, the keyword string detection process is terminated. At this time, the detection process may be terminated without analyzing all the characters included in the predetermined sentence input as the detection target. Therefore, if it is determined by the above process not to move to the substructured data, the next sequence character of the character used for determining whether to move to the substructured data is newly recognized among the predetermined sentences, and then detected from step S220. The process resumes. By repeating this process, keyword string detection may be performed on all characters included in the input predetermined sentence.

5 is a diagram illustrating an example of a structured data structure.

Referring to FIG. 5, the structured data is a concept including both the first structured data and the substructured data shown in the above description, and the structured data is configured to include keyword related information preset for each of all the characters. Here, the keyword related information is configured to include character correspondence code value information, character correspondence information, and address information of substructured data. Here, the character correspondence information is configured to include whether there is a keyword string again ending with the character and the number of keyword strings following the character.

The character corresponding code value information means that a corresponding code value is determined for each character that may exist. In FIG. 5, the KSC 5601 Korean code is used, and code value information corresponding to each of the 2350 Korean characters is matched. Character-corresponding code value information may be expressed by various methods as well as expressed in FIG. 5.

The character correspondence information refers to information related to the corresponding character, and means a part that can be updated when the keyword string is structured. As described above, the character correspondence information includes whether there is a keyword string ending with the character and the number of keyword strings following the character.

Referring to the character "ga" in FIG. 5, it can be seen that the character correspondence code value information is "0XB0A1", and "0" and "1" are described in order as the character correspondence information corresponding to the character "ga". The first information means that there are no keyword strings ending with "A" (0), and the second information means that the number of keyword strings following "A" is one. The third information may refer to a pointer PTR including location information of sub-structured data ending in keyword related information corresponding to “a” as address information of sub-structured data. The character correspondence information before being structured, that is, the character correspondence information in the initialization state is set to "N" and "0" in order. In this initialization state, when constructing a keyword string starting with "ga", for example, "scissors", the character correspondence information of "ga" is updated to "0" and "1" in order. That is, since the keyword string to be structured is "scissors", the keyword string ending in "ga" is "0", and the number of keyword strings following "ga" becomes "1".

6 is a diagram illustrating an example of keyword string structure.

Referring to FIG. 6, it can be seen that a keyword string “Dragonfly” is structured. When the keyword management server receives an input for structuring a keyword string of "dragonfly", first, the keyword-related information corresponding to "sleep" is searched for in the first structured data. In this case, searching for keyword related information corresponding to “sleep” in the first structured data may be searched using code value information of each character. That is, when the code value difference between "0XBOA0" and "sleep" code value information "0XC0E1" corresponding to the reference code value information described with reference to FIG. 4 is used, the keyword related information corresponding to "sleep" from the position of the reference code value information is used. Where can be inferred.

After retrieving keyword related information corresponding to "sleep" from the first structured data, the character correspondence information included in the keyword related information is updated. In other words, it updates information that the number of keyword strings following "sleep" has increased. Next, since "dragonfly" is composed of two or more characters, a process of searching for keyword-related information corresponding to "child" is continued by moving to substructured data that is dependent on keyword-related information corresponding to "sleeping." In this case, it is assumed that substructured data that is dependent on each keyword related information is generated in advance. After retrieving the keyword related information corresponding to "child", the character correspondence information included in the keyword related information, that is, information that the number of keyword strings following the "child" has been updated is updated. Subsequently, a process of searching for keyword related information corresponding to “L” by moving to substructured data that is dependent on keyword related information corresponding to “child” is continued. After retrieving the keyword related information corresponding to "Lee" and updating the character correspondence information included in the keyword related information, "Lee" is the last character of the keyword string, so the keyword string structuring ends. That is, it indicates that there is a keyword string ending with "Lee". Through this process, the keyword string structure for the input "dragonfly" is achieved.

7 is a diagram illustrating an example of keyword string detection.

Referring to FIG. 7, when a sentence “Thank you very much” is input, a process of detecting a keyword string “thank you” corresponding to a structured keyword string may be found in the input sentence. First, "Thank you very much", which is a sentence to be detected, is input to the keyword management server. The keyword management server that senses the sentence input recognizes "true", which is the first character of the sentence, and searches for keyword related information corresponding to "true" in the first structured data. Next, character detection information included in keyword related information corresponding to "true" is analyzed to detect a detectable keyword string and determine whether to move to substructured data. Referring to the character correspondence information of "true" shown in Figure 7, it can be seen that there is no keyword string ending with "true", and that the number of keyword strings following "true" is zero. Therefore, since there is no keyword string following "true", the keyword management server decides not to move to the substructured data.

After the detection process for "true" in the sentence is completed, the next character "feel" is recognized and the detection process proceeds again. That is, the keyword-related information corresponding to "feel" is searched again in the first structured data, the character-related information included in the searched keyword-related information is analyzed to detect a detectable keyword string, and whether to move to the sub-structured data. Determine. Referring to the character correspondence information of "persimmon" shown in Figure 7, it can be seen that there is no keyword string ending with "persimmon", and the number of keyword strings following "persimmon" is one. Thus, no keyword string detection is made, but it is determined to move to the substructured data.

When moving to the substructured data of "persimmon" by the above process, the keyword related information corresponding to the "sa" corresponding to the next character of "persimmon" is searched among the sentences inputted from the moved substructured data. The character correspondence information included in the retrieved keyword related information is analyzed to detect a detectable keyword string and determine whether to move to substructured data. Referring to the character correspondence information of "sa" shown in Fig. 7, it can be seen that there is a keyword string ending with "sa", and there is no keyword string following "sa". Thus, the keyword string detection process is terminated because it is determined that the keyword string " audit " is detected and does not move to substructured data.

In the same principle as described above, the characters "sum" and "nee" after the character of "four" in the input sentence. The same process is repeated for "da". In FIG. 7, since the structured keyword string is set to "audit" one, the keyword string detected eventually becomes "audit".

In the above, the configuration of the present invention with reference to the preferred embodiment of the present invention, but the present invention is not limited to the above embodiment, of course, it is within the scope of the present invention as grasped by the appended claims and drawings Various modifications, modifications, and improvements are possible in the art to those skilled in the art.

1 is a diagram schematically illustrating a process of structuring a keyword string using a keyword management server and detecting a structured keyword string according to an embodiment of the present invention.

2 is a block diagram showing a detailed configuration of a keyword management server according to an embodiment of the present invention.

3 is a flowchart illustrating a process of structuring a keyword string input as a structured object in a keyword management server according to an embodiment of the present invention.

4 is a flowchart illustrating a process of detecting a keyword string structured by the method described with reference to FIG. 3 in a sentence input as a detection target in a keyword management server according to an embodiment of the present invention.

5 is a diagram illustrating an example of a structured data structure.

6 is a diagram illustrating an example of keyword string structure.

7 is a diagram illustrating an example of keyword string detection.

Claims (11)

In the method for structuring a keyword string input as a structured object in the keyword management server, A first step of recognizing a first character among at least one or more characters constituting the keyword string input as the structured object; A second step of searching for keyword-related information set corresponding to the recognized first character in the first structured data including preset keyword-related information for each of the characters capable of constructing the keyword string, wherein the The keyword related information includes character correspondence code value information, character correspondence information, and address information of substructured data set for each of the above characters; A third step of updating character correspondence information included in the searched keyword related information; And If the keyword string consists of two or more characters, A) step of identifying the location information where the substructured data is stored by the address information of the substructured data included in the keyword related information of the character corresponding to the character correspondence information updated in the previous step, wherein the address information of the substructured data If is not present, create new substructured data, store the location information of the generated new substructured data in the address information of the substructured data, and the character corresponding to the character correspondence information updated in the previous step. Retrieving the keyword related information set corresponding to the next order character in the identified substructured data and updating the character corresponding information included in the retrieved keyword related information to the last character of the keyword string. 4 steps Keyword string structuring method comprising a. The method of claim 1, And the character correspondence information includes whether a keyword string ending with the character exists and the number of keyword strings following the character. The method of claim 1, And the sub-structured data includes keyword-related information preset for each of all the characters. The method of claim 1, If the character is Korean, the character correspondence code value information is set using the KSC 5601 Hangul code. A method for detecting a keyword string from a sentence input as a detection target in a keyword management server using structured data in which the keyword string is structured by the method of at least one of claims 1 to 4, A first step of recognizing a first character among at least one or more characters constituting a sentence input as the detection target; A second step of retrieving keyword related information set corresponding to the recognized first character from the first structured data including keyword related information preset for each of all the characters, wherein the keyword related information is set to all the characters. Including character correspondence code value information, character correspondence information, and address information of substructured data set for each of the two; A third step of detecting a detectable keyword string by analyzing character correspondence information included in the searched keyword related information, and determining whether to move to substructured data; A fourth step of terminating a keyword string detection process if it is determined by the third step not to move to substructured data; And If it is determined by the third step to move to the substructured data, A) moving to substructured data and retrieving keyword related information set corresponding to a next order character of a character corresponding to character correspondence information analyzed in a previous step among the input sentences in the moved substructured data; And B) detecting a detectable keyword string by analyzing character correspondence information included in the searched keyword related information, and determining whether to move to substructured data; Step 5 repeating until it is determined not to move to substructured data Keyword string detection method comprising a. The method of claim 5, If it is determined by the fourth step or the fifth step not to move to the substructured data, after recognizing the next order character of the character corresponding to the character correspondence information analyzed in the previous step, the second to fifth steps are performed. Keyword string detection method characterized in that for performing. The method of claim 5, And the character correspondence information includes whether a keyword string ending with the character exists and the number of keyword strings following the character. The method of claim 5, And the sub-structured data includes keyword-related information preset for each of all the characters. The method of claim 5, And the character corresponding code value information is set using KSC 5601 Korean code. In the keyword string structuring apparatus, which is included in the keyword management server to structure a keyword string input as a structured object, The keyword string structuring apparatus recognizes a first character among at least one or more characters constituting the keyword string input as the structured object, and pre-set keyword-related information for each of all characters constituting the keyword string. Search for keyword-related information set corresponding to the recognized first character in the first structured data, wherein the keyword-related information includes character-corresponding code value information, character-corresponding information, and character set for each of all the characters; Including address information of sub-structured data; updating character correspondence information included in the searched keyword related information; and if the keyword string is composed of two or more characters, character corresponding to the character correspondence information updated in the previous step. Keyword related A) step of identifying the location information where the substructured data is stored by the address information of the substructured data included in the beam, wherein new substructured data is generated if the address information of the substructured data does not exist; Storing the position information of the generated new substructured data in the address information of the substructured data; and the keyword related information set corresponding to the next order character of the character corresponding to the character correspondence information updated in the previous step. And searching b) in substructured data and updating character correspondence information included in the retrieved keyword related information to the last character of the keyword string. A keyword string detection device included in a keyword management server and detecting a keyword string from a sentence input as a detection target by using the structured data structured by the apparatus of claim 10, wherein the keyword string is detected. The keyword string detection apparatus recognizes a first character among at least one or more characters constituting a sentence input as the detection target, and includes first keyword-related information configured for each of all the characters. Retrieves keyword related information set corresponding to the recognized first character, wherein the keyword related information includes character correspondence code value information, character correspondence information, and address information of substructured data set for each of all the characters. Includes,-Detects a keyword string that can be detected by analyzing character correspondence information included in the searched keyword related information, determines whether to move to substructured data, and determines that the keyword string does not move to substructured data. Terminate the detection process If it is determined to move to the substructured data, it moves to the substructured data and is set corresponding to the next order character of the character corresponding to the character correspondence information analyzed in the previous step among the input sentences in the moved substructured data. A) searching for keyword related information and character correspondence information included in the searched keyword related information to detect a detectable keyword string, and b) determining whether to move to substructured data. Iterate until it is determined not to move to the keyword string detection apparatus.

Images