KR20080048149A - Advertising system and method using category - Google Patents

Abstract

A system and a method for providing category AD(ADvertisement) are provided to maximize an AD effect by analyzing a category and a keyword related to the category of a document stored in a website, and matching the document with the AD of an advertiser, and link the new keyword with the AD of the advertiser by finding the new keyword from the documents having the keyword included in the category. A keyword extracting module(120) determines a document including a main keyword representing each category by analyzing a plurality of documents and extracts at least one word included in the document including the main keyword. A weight calculating module(130) calculates category weight of the extracted word according to a display frequency displayed with each document including the main keyword and a display concentration rate in the category. A category matching module(140) selects the word included in the category according to the category weight of the word, and generates a category set linking the word included in the category and the category weight of the word. An AD matching module(160) matches AD of an advertiser with the document including the word included in the category selected by the advertiser.

Description

Advertising System and method using category}

1 is a view showing an example for a conventional keyword advertising,

2 is a view for conceptually explaining a category advertisement method of the present invention;

3 is a block diagram of a category advertisement system according to an embodiment of the present invention;

4 is a diagram conceptually illustrating a category matching method when two or more categories are matched with respect to one word;

5 is a diagram illustrating a weight calculation method for words belonging to two or more categories; and

FIG. 6 is a diagram for explaining an example of differently assigning weights according to positions occupied by a word in a document.

* Description of the symbols for the main parts of the drawings *

100: category advertising system 110: analysis module

120: keyword extraction module 130: weight calculation module

131: exposure concentration calculation module 132: site weight calculation module

133: confidence weight calculation module 134: position weight calculation module

135: first weight calculation module 132: second weight calculation module

140: category matching module 141: association keyword determination module

142: category set generation module 150: database

160: advertising combination module

KR 10-2005-0058172

The present invention relates to a category advertisement system and a category advertisement matching method, and more particularly, to a category advertisement system for matching an advertisement with a category of interest of an advertiser, and a method, which deviates from the conventional advertisement method for matching keywords and advertisements.

Typically, when accessing content or documents provided on a website, advertisements are sometimes included in the documents or content provided by the website. An advertisement included in a document (or content) is generally displayed in correspondence with a word included in the document. In the Overture advertisement, a similar advertisement technique, when the searcher's search term matches the advertisement keyword selected by the advertiser, the advertiser's advertisement is matched to the search result corresponding to the search term and provided to the searcher. Overture advertising is a keyword selected by an advertiser, for example, when the advertiser sets the word "bag" and pays the cost (advertising cost) for the word, the advertiser's advertisement is combined with the search result for the "bag" and displayed in the search result. However, if the searcher performs a search with the search term "handbag", the Overture ad does not combine the advertiser's advertisement (keyword "bag") with the search results. This causes the overture advertisement to match the advertisement only for documents containing the advertiser's words. In addition, if a document containing the keyword "bag" adversely affects the advertisement, for example, as shown in Figure 1, the word bag 51 in the news article 50 that "the killer has abandoned the body in the bag" 51 ), The advertiser's advertisement (bag advertisement 52) is preferably not provided to match the news article. In this regard, Korean Patent Laid-Open Publication No. 10-2005-0058172 discloses an "online advertising system using a negative keyword and method thereof" that does not combine an advertisement with respect to content including "negative keyword" having a negative meaning. In Korean Patent Laid-Open No. 10-2005-0058172, this problem is solved by preventing the advertiser's advertisement from being combined with the negative content including the negative keyword. In addition, Patent Publication No. 10-2005-0058172 categorizes the content and the advertisement of the advertiser and allows the advertisement combination only when the categories match each other. As an example,

News articles that say "The Killer Abandoned the Body in a Bag" would be a category included in "Social / Cultural / News", and an advertiser's category would be a category belonging to "Shopping / Fashion Goods / Bags". You are blocking the association of ads for. That is, the category referred to in Patent Publication 10-2005-0058172 represents a category according to the classification criteria of the portal site. However, when the searcher's search term is "diet", the search term may appear in the news or the shopping mall, and the content and the advertiser's category may not match. In addition, it is difficult to match the search results and advertisements for content that includes new buzzwords on the diet, such as S-line and body.

Accordingly, an object of the present invention is to analyze a document (or content) provided on a website to analyze a category and a related keyword associated with the category, and match the document containing the same with the advertisement to maximize the advertisement effect A system, and method are provided. Another object of the present invention is to provide a category advertisement system and method for additionally finding a new related keyword in a document including related keywords belonging to a category, and associating the new related keyword with an advertisement of an advertiser. In addition, another object of the present invention is to set a plurality of words in one category, the advertiser selects the category category advertising that can increase the advertiser's exposure to the advertising compared to the Overture advertising to buy and manage for each keyword A system, and method are provided. In addition, another object of the present invention is to provide a category advertising system, and method that can combine the advertisement with the online content and provide it to the user.

The above object is a keyword extraction module for analyzing a plurality of documents, determining a document including a main keyword representing each category, and extracting at least one word provided in the document including the main keyword. A weight calculation module for calculating a weight of the word for the category according to an exposure frequency exposed in each document including the main keyword and an exposure concentration in the category, and belonging to the category according to the weight of the word. A category matching module for generating a category set in which a word belonging to the category, a word belonging to the category and the weight of the word, and a word corresponding to a category selected by the advertiser are matched with the advertisement of the advertiser. And an advertisement matching module, wherein the keyword extraction module Analyzing the document containing the words provided in the category set by the advertising system is achieved by a category extracting the associated keywords that are associated with the category.

The concentration of exposure is preferably calculated by the ratio of the exposure frequency at which the word is exposed in one document and the frequency at which the word is exposed in the entire document.

The weight calculation module calculates an exposure frequency at which any one of the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword is exposed on a site, and calculates a site weight according to the result. The final weight of one of the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword is added or subtracted by the site weight.

The weight calculation module may increase or decrease the weight according to the reliability of the site, and the reliability of the site may be given according to a site order of frequently exposing a high weight word for the category.

The weight calculation module may assign a position weight to the word according to a position at which one of the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword is exposed in the document. The position may be any one of a title, a body, and a periphery of the document, and the weight for each position is preferably reduced in the order of the title, the body, and the periphery.

The category matching module preferably determines at least one of the words included in the category as an associated keyword according to the order of the weight of each word.

When the word is a unique word found only in the single category, the advertisement matching module preferably matches a document including the unique word with the advertisement of the single category.

Preferably, the method further includes a database including a set of categories in which each word is associated with the weight for each word.

The keyword extraction module may analyze a document including a related keyword included in the category set included in the database, and find a related keyword included in the first category from a document including the related keyword.

The advertisement matching module may vary a matching ratio between the advertisement and the document according to an order of payment amount of at least one advertiser who paid an advertisement fee for the category.

The keyword extracting module may extract the word by performing a morphological analysis on a sentence composed of a combination of words included in the document.

The above object is to analyze a plurality of documents, determine a document including a main keyword representing each category, and extracting at least one word provided in the document containing the main keyword, the Calculating a weight of the word for the category according to an exposure frequency exposed in each document including a main keyword and an exposure concentration in the category, and selecting a word belonging to the category according to the weight of the word And generating a category set in which words belonging to the category and weights of the words are linked, and matching the advertisement of the advertiser to a document including a word corresponding to a category selected by the advertiser.

The concentration of exposure is preferably calculated by the ratio of the exposure frequency at which the word is exposed in one document and the frequency at which the word is exposed in the entire document.

The calculating of the weight of the word may include calculating an exposure frequency at which any one of the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword is exposed to a site and according to the result. The site weight is calculated, and one of the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword is preferably final weighted by the site weight.

The calculating of the weight of the word may increase or decrease the weight according to the reliability of the site, and the reliability may be given according to a site order of frequently exposing a word having a high weight for the category.

The calculating of the weight of the word may include: selecting a position of the word according to a position where one of the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword is exposed in the document. The weight is given, and the position is any one of the title, the body, and the surroundings of the document, and the weight for each position is preferably reduced in the order of the title, the body, and the surroundings.

The method may further include determining at least one of the words included in the category as an associated keyword according to the order of the weights.

The extracting of the word may include analyzing a document including a related keyword included in the category set included in the database, and finding a related keyword included in the first category from a document including the related keyword. It is preferred to include the step.

In the extracting of the word, the word may be extracted by performing a morphological analysis on a sentence composed of a combination of words included in the document.

The document is preferably combined with an advertisement corresponding to the category.

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

2 is a diagram for conceptually explaining a category advertisement method of the present invention.

In the illustrated category, the "diet" category shows related keywords such as body and S-line, and the category "car insurance" includes related keywords such as traffic accident, car center, and towing. A category is a collection of at least one related keyword, and the main keyword representing related keywords belonging to the category usually matches the name of the category. That is, the main keyword of the category "diet" is the same diet as the category name. According to this category structure, the advertiser selects a category to match with his advertisement by referring to the category and related keywords belonging to the category, and the advertiser's advertisement is combined with the document corresponding to the category. When a document associated with an advertisement is viewed on a web site, such as a portal site, the combined advertisement is exposed to a large number of users, and the user is exposed to the document that he or she views and the appropriate advertisement in the content.

The category advertisement method of the present invention classifies documents collected from a web site into respective categories, and combines advertisements of advertisers corresponding to the classified categories with the documents. The category advertisement method according to the present invention does not only find related keywords centered on the main keyword and matches the advertisement of the advertiser. The analysis module (not shown and described later) is used to determine a document including a main keyword (e.g., a web page, a web document, a web site, and other online content, which will be omitted below), and the related keyword is included in the document. After extracting, additional related keywords exposed together with the related keywords are searched based on the related keywords. Through the analysis method having such a circulation structure, newly generated related keywords are continuously found according to the language usage trend of the Internet, and a document including the found related keywords can be newly registered in a category. Accordingly, since a plurality of words associated with the category correspond to the category selected by the advertiser, the number of exposures of the advertisement may be increased and the advertisement effect may be increased. In addition, since the entire keywords belonging to the category are matched with the advertisement, the user may encounter an advertisement highly related to the content of the document in the document selected by the user.

3 is a block diagram of a category advertisement system according to an embodiment of the present invention.

The illustrated embodiment includes an analysis module 110, a keyword extraction module 120, a weight calculation module 130, a category matching module 140, a database 150, and an advertisement matching module 160.

The analysis module 110 analyzes a web document using a main keyword of a category, and determines a document including a main keyword among documents included in the web site. The analysis module 110 preferably targets documents provided in one or more predetermined sites (eg, portal sites). As described above with reference to FIG. 2, when the analysis module 110 analyzes the “diet” category, a document including “diet” is selected based on a main keyword (diet) representing “diet” as an analysis target. Figure out. The keyword extraction module 120 applies a morpheme analysis method to each document analyzed by the analysis module 110 to extract a word included in the document. The word extracted by the keyword extraction module 120 is provided to the weight calculation module 130, the weight calculation module 130 is a frequency, exposure concentration, exposure rate of each word extracted from the keyword extraction module 120, the document And a weight for the extracted word by referring to the number of documents to which each word is exposed.

Preferably, the weight calculation module 130 includes an exposure concentration calculation module 131, a site weight calculation module 132, a trust weight calculation module 133, a position weight calculation module 134, and a first weight calculation module 135. ), A second weight calculation module 136.

The exposure concentration calculation module 131 calculates the weight of the exposed word based on the frequency of exposure of the words found together in the document including the main keyword and the number of the exposed documents. The exposure concentration calculation module 131 may increase the number of exposures while decreasing the number of documents exposed for one category, and decrease the opposite case. This is due to an increase in the frequency in the exposed documents, while the number of documents exposed is small in the case of words frequently exposed only in a specific field. On the contrary, that is, when the number of documents exposed and the frequency of exposure in each document are high, the corresponding word may be determined as a general word. The frequency of word exposure in a limited document is called exposure concentration, and the higher the exposure concentration, the higher the weight. The weight whose magnitude increases in proportion to the increase in the exposure concentration is calculated by the formula shown in Table 1 below.

Weight = 5log (TF / (iDF + 1)), where TF is the number of documents to be analyzed and iDF is the number of documents in which words are found.

If the analysis module 110 analyzes 1000 문서 documents for the first category, the word “S line” is found in 1 문서 documents among the total 1000 문서 documents for analysis, and averages 5 in one document. Suppose it is used twice, the weight of the word "S line" is

5 log (1000 萬 / (1 萬 +1)) = 2.49 (weights do not have special units, and the number "5" represents the frequency of exposure per document.)

If the analysis module 110 analyzes 1000 문서 documents for the first category, the word “weight” is found in 300 문서 documents among the total 1000 문서 documents to be analyzed, and averages 10 times in one document. Assuming that the weight of the word "weight" is

10 log (1000 萬 / (300 萬 +1)) = 2.30

As described above, the weight calculation method according to the present invention does not give a high weight simply because the number of exposures in a plurality of documents is high. The present invention allows words that are intensively exposed in a limited document to have a higher weight in a category.

The site weight calculation module 132 increases or decreases the weight of the word according to the source of the word included in the document. If an important site (hereinafter referred to as A site), which is a main source of documents collected for a specific category (hereinafter referred to as A category), belongs to category A, the site weight calculation module 132 may collect If the documents contain the same words as those collected at Site A, additional weight is given to the words contained in the document. The weight given to the word preferably has a form of increasing or decreasing the weight previously calculated by the exposure concentration calculation module 131. For example, among the collected words, a word collected at the site A may be given a weight of +2, and a word collected at another site may be given a weight of +1.

The position weight calculation module 134 assigns weights to words differently according to the positions of the words collected in each document. The position of a word may be any one of a title, a body, and a periphery, and the position weight calculation module 134 assigns weights in the order of title> body> periphery. This will be described with reference to FIG. 6 together.

FIG. 6 is a diagram for explaining an example of differently assigning weights according to positions occupied by a word in a document. In the illustrated document, "model Kang Na-yeon" is described in the title portion 61, and "diet" (62, 64) are shown in the body and the periphery (65). As shown, there is a main keyword "diet" in the document, but the main purpose is to promote the model Kang Na-yeon shown in the title 61, diet is only included in the incidental content. Therefore, the position weight calculation module 134 gives the maximum value when the position in the document of the collected word is the title portion, gives the next weight to the body, and gives the minimum weight to the periphery. The weight given to the word is preferably in the form of adding or subtracting from the weight calculated by the exposure concentration calculation module 131. The trust weight calculation module 133 varies the weight calculated by the exposure concentration calculation module 131 according to the reliability of the site. The trustworthiness of a site is determined by whether the site is operated in a certain area (eg, a diet) continuously for a certain period of time.

The first weight calculation module 135 and the second weight calculation module 136 determine which category the collected words belong to when one word is found and collected in two or more categories.

If a word is found in more than one category, the weight for the word is determined by the exposure concentration calculation module 131, the site weight calculation module 132, the confidence weight calculation module 133, and the position weight calculation module 134. The final weight may be calculated by finally collecting the values calculated in, and the final weight for the word may be calculated by determining which weight is higher. Since this has been described in detail above, a detailed description thereof will be omitted. On the other hand, when a word is found in more than one category, it is a matter of course that a category for a word may be set by referring to the frequency of the single word is exposed in both categories. For example, if the words exposed in the most documents among the documents analyzed by each category in the analysis module 110 are in order of body, S-line, and obesity, 9, 8, 7 for body, S-line, and obesity, respectively. Can be weighted. If the words "S-line" and "obesity" have the same percentage of exposure in each document (for example, the same exposure in 7 of 10 documents), then the top weights are assigned to the words with the most exposures in each document. Applicable This will be described with reference to FIG. 4 together. 4 conceptually illustrates a category matching method for a case where two or more categories are matched with respect to one word. The word “weight scale” shown may be exposed to both the diet category (hereinafter referred to as the first category for ease of explanation and understanding) and the fitness equipment category (hereinafter referred to as the second category for ease of explanation and understanding). have. If the word "weight" is 3 in the first category and the weight in the second category is 10, it may be determined that the category of the word "weight" is the second category. Meanwhile, the weight calculation module 130 calculates a weight for each document having two or more categories (when one document includes words having two or more categories) and two or more categories. If the word "weight balance" is present among the keywords extracted for the category "diet" in the keyword extraction module 120 and the word "weight scale" also exists in the category "exercise apparatus", the weight calculation module 130 is shown in FIG. As shown, the category " diet " and the category " exercise device " are set to the first category and the second category, and the first weight value calculating module respectively sets the first category weight and the second category weight for the word “weight scale”. And the result of the calculation, by the 135 and the second weight calculation module 136, and providing the result to the category matching module 140. Of course, after calculating the weight for each category of words having two or more categories through the exposure concentration calculation module 131, the site weight calculation module 132, the trust weight calculation module 133, and the position weight calculation module 134, By comparing them, the words of a category having a large weight may be selected and provided to the category matching module 140. If there is a unique word found only in the first category, the weight calculation module 130 calculates the weight of the unique word only for the first category, and the category matching module 140 generates the unique word. Match only one category.

The category matching module 140 matches each word with a category prepared in advance by referring to the weight of each category word provided by the weight calculation module 130. In addition, the category matching module 140 includes a main keyword of each category, generates a category set based on a high-weighted related keyword in the category, and stores it in the database 150. The category set is data in which the associated keyword and the associated keyword weights are combined into one. If the related keywords belonging to the category "diet" are body, S-line, and obesity, and the weights are 9, 8, and 7, respectively, 9, S-line and 8, obesity and 7 combined information means. Since the related keyword is a word distributed in a document including the main keyword, it is most often used together with the main keyword, and is most likely a word having a similar meaning to that of the main keyword (eg, diet). Therefore, the higher the semantic similarity with the main keyword, the higher the frequency of exposure together in the document including the main keyword. Accordingly, the category set stored in the database 150 is provided to the analysis module 110 again, and the analysis module 110 discovers a new word related to the related keyword through the related keyword combined with the category set. Through this process, it is possible to continuously find related keywords derived from the original meanings over time. Here, when the word found through the main keyword is a unique word, the category matching module 140 assigns a maximum weight to the unique word. Accordingly, the keyword extraction module 120 extracts a related keyword based on the unique keyword, assigns a weight to the extracted related keyword, and determines whether to belong to a category according to the size of the weight. That is, the analysis module 110 includes a unique word instead of a main keyword in the web site, extracts an associated keyword belonging to a category, and the category matching module 140 forms a category based on the unique word. The unique word belongs to the main keyword, but the related keyword belonging to the category may be configured around the unique word. On the other hand, when the unique word is one of a plurality of related keywords belonging to one category (eg, a diet category), the category is not configured based on the unique word, and the unique word may have the maximum weight in the category.

The category set generation module 142 generates a category set by combining the weights calculated for the words and the categories of the words and stores them in the database 150. The advertisement matching module 160 is connected to portal sites, search sites, and other websites, and combines the documents (or digital contents) provided on the sites with advertisements or combines advertisements with newly generated documents. To provide.

In addition, the present invention may be combined with a search engine to combine an advertisement with a search result. In this case, the method of combining the ad with the document

1) receiving a search result from a website server having a search function and adding feedback to the provided search result.

2) A method for receiving a search term from a website server having a search function and providing corresponding advertisement contents to the website server.

3) A method of integrating a category advertising system according to the present invention with a website server having a search function, wherein the website server combines advertisements with the search results.

4) ASP lease of a part of the search result provided to a searcher by a website server having a search function, and insert advertisement content (banner advertisement, text advertisement, and URL indicator related to the advertisement) in the lease area. .

5) If a search term is received from a website having a search function and a corresponding category is searched, and if a search result exists, the website information of the advertisement content for the search term (for example, URL information of a server where the advertisement content is stored) is searched. The server provides the server, and the website server receives the advertisement from the directory information provided by the category advertisement system and combines the advertisement and the search result.

In addition, as technology advances, advertising content and search results can be combined in various ways.

As described above, the present invention extracts the related keywords belonging to the category including the main keyword, and combines the extracted related keywords and the main keyword to the category selected by the advertiser to maximize the advertisement effect of the advertiser. In addition, the present invention includes a related keyword, by extracting a new word associated with the category can continuously update the word associated with the advertiser's advertisement. In addition, the present invention may increase the number of impressions of the advertisement and increase the advertisement effect since a plurality of words associated with the category correspond to the category selected by the advertiser. It can also combine advertising closely with online content and provide it to the user.

Claims (20)

A keyword extraction module configured to analyze a plurality of documents, determine a document including a main keyword representing each category, and extract at least one word provided in the document including the main keyword; A weight calculation module for calculating a weight of the word for the category according to an exposure frequency exposed in each document including the main keyword and an exposure concentration in the category; A category matching module for selecting a word belonging to the category according to the weight of the word, and generating a category set in which the word belonging to the category and the weight of the word are linked; And And an advertisement matching module for matching the advertisement of the advertiser to a document including a word corresponding to a category selected by an advertiser. The keyword extraction module, And a related keyword associated with the category is extracted by analyzing documents including the words provided in the category set. The method of claim 1, The exposure concentration is And a ratio of an exposure frequency at which the word is exposed in one document and a frequency at which the word is exposed in the entire document. The method of claim 1, The weight calculation module, Calculating an exposure frequency at which one of the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword is exposed on a site, and calculating a site weight according to the result; The one of the main keyword, the keyword classified into the same category as the main keyword, and the related keyword associated with the keyword is the final weight is added or subtracted by the site weight. The method of claim 3, The weight calculation module, Increase or decrease the weight according to the reliability of the site, The reliability of the site, The category advertising system, characterized in that given in the order of sites that frequently expose the high-weight words for the category. The method of claim 1, The weight calculation module, A position weight for the word is assigned according to a position in which the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword are exposed in the document, The position is, Any one of the title, body, and surroundings of the document, The weight for each location is a category advertising system, characterized in that the weight in decreasing order of the title, body, and surroundings. The method of claim 1, The category matching module, And at least one of the words included in the category is determined as an associated keyword according to the order of the weight of each word. The method of claim 1, The advertisement matching module, If the word is a unique word found only in the single category, And a document including the unique word matches the advertisement of the single category. The method of claim 1, And a database including a set of categories in which the words and the weights for the words are associated with each other. The method of claim 8, The keyword extraction module, Analyzing a document including related keywords included in the set of categories provided in the database, And a related keyword included in the first category from a document including the related keyword. The method of claim 1, The advertisement matching module, And a matching ratio between the advertisement and the document in accordance with the order of payment amount of at least one advertiser who paid the advertisement fee for the category. The method of claim 1, The keyword extraction module, The category advertisement system, characterized in that for extracting the word by morphological analysis for a sentence consisting of a combination of words included in the document. Analyzing a plurality of documents to determine a document including a main keyword representing each category, and extracting at least one word provided in the document including the main keyword; Calculating a weight of the word for the category according to an exposure frequency exposed in each document including the main keyword and an exposure concentration in the category; Selecting a word belonging to the category according to the weight of the word, and generating a category set in which the word belonging to the category and the weight of the word are linked; And And matching the advertisement of the advertiser to a document including a word corresponding to a category selected by the advertiser. The method of claim 13, The exposure concentration is And a ratio of the frequency of exposure of the word in one document and the frequency of the word in the entire document. The method of claim 13, Computing the weight of the word, Calculating an exposure frequency at which one of the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword is exposed on a site, and calculating a site weight according to the result; The one of the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword, the final weight is added or subtracted by the site weight. The method of claim 15, Computing the weight of the word, Increase or decrease the weight according to the reliability of the site, The reliability is, The category advertisement matching method of claim 1, wherein the category advertisement matching method is given according to a site order of frequently exposing a high weight word for the category. The method of claim 13, Computing the weight of the word, A position weight for the word is assigned according to a position in which the main keyword, a keyword classified into the same category as the main keyword, and an associated keyword associated with the keyword are exposed in the document, The position is, Any one of the title, body, and surroundings of the document, The weight for each location is a category advertising system, characterized in that the weight in decreasing order of the title, body, and surroundings. The method of claim 13, And determining at least one of the words included in the category as an associated keyword according to the magnitude order of the weights. The method of claim 18, Extracting the word, Analyzing a document including an associated keyword included in the set of categories provided in the database; And And searching for a related keyword included in the first category from a document including the related keyword. The method of claim 13, Extracting the word, Category advertising matching method, characterized in that for extracting the word by morphological analysis for the sentence consisting of a combination of words included in the document. The method of claim 12, The document, The category advertisement matching method, characterized in that coupled with the advertisement corresponding to the category.

Images