KR101221959B1 - An Integrated Region-Related Information Searching System applying of Map Interface and Knowledge Processing - Google Patents

Abstract

Integrated search method of local information using map interface and knowledge processing to improve user convenience and provide cultural tourism information by connecting various local information included in web documents and displaying them on a map This is disclosed.
The integrated search method for local information using the disclosed map interface and knowledge processing includes a first step of extracting a representative keyword representing a region and collecting and grouping various documents about the region from the web based on the extracted representative keyword; ; Extracting a unique region keyword representing the region to generate related knowledge between the keywords; And a third step of comprehensively providing local information through a map-based interface so as to intuitively grasp relevant information of the extracted unique region keyword based on geographic information.

Description

An integrated region-related information searching system applying of map interface and knowledge processing

The present invention relates to integrated search of local information, and more particularly, by connecting various local information included in a web document and displaying it on a map, it is possible to facilitate user's information acquisition and efficiently provide cultural tourism information. The present invention relates to a method of integrated retrieval of local information using a map interface and knowledge processing.

With the development of internet technology and the formation of a user-oriented web environment, various information and services can be easily and conveniently provided through the web. Unlike the existing web, the current web (web 2.0) is formed around the collaboration, sharing, and participation of users, and there is a much larger amount of information than the previous web, where information producers unilaterally created and provided information. Was done.

With the change of the internet environment, the large portals such as Google and NAVER provide various map-based services, and the user's interest in the real space is leading to the acquisition of local information on the web. Indeed, local information on restaurants, real estate, jobs, traffic, festivals and events are generated in various forms such as blogs, encyclopedias, news, and images through service providers and users, and the number is exploding. In addition, such local information is currently provided in conjunction with maps around large portals.

When the user wants to acquire local information through the current web, the process goes to the process shown in FIG. Users typically generate search keywords and enter them into portal sites such as NAVER or Google. At this time, the portal site provides the user with a list of numerous web documents related to the search word on the search result page. The user should read through the web documents through the links provided in the list provided and identify and synthesize relevant local information. For example, the history, history, and cultural properties of the region will be read through web pages, such as web pages and encyclopedias, and blogs if you want to know what other users have experienced. In addition, if the document contains a place name or event venue, the location information should be further confirmed through the map service.

The local information service using the existing map service is Google Maps of Google. Google supports searches for businesses, addresses, and places around the world based on global map data. Photographs, videos, Wikipedia, etc. are displayed on the map with local information based on the location based on the keyword entered by the user, but the information displayed on the map is very simple and one-dimensional information. In other words, only one content and one location (coordinate) information is connected one-to-one, and it is difficult to determine the relevance between the contents related to the region. In addition, the company Google handles a lot of information in a very wide area, that is, around the world, so it's hard to get detailed information about any particular area.

Domestic portal companies' map service provides relatively detailed domestic local information rather than Google's map service. In the case of NAVER or Daum, the map service can show the address, phone number, public transportation, etc. of local attractions, major stores, and road traffic conditions in real time. NAVER recently provides thematic maps such as real estate maps, bicycle maps, mountain climbing maps, and web contents such as related blogs and cafes including place names keywords, and detailed information pages.

In addition, as a local information providing service, there is a blog, which is an Internet media media, and is currently being serviced by NAVER, Daum, Yahoo. A blog writes about a person's small daily life or interests, and includes a lot of information about events, travel, and restaurants. These blogs contain a variety of subjective content, including personal experiences, thoughts, and feelings, and include geographic data such as place names and business names. However, blog writers already know the location or address of the site they visited, and most don't provide a map that gives them an overview. Therefore, the other users sharing the blog is inconvenient to additionally confirm the geographical information through a map for the name or business name recorded in the blog. For example, suppose that there is a blog about "Hana Tour Expo" as shown in FIG. The blogger also recorded his experiences and also noted that the venue where the fair was held was 'KINTEX'. At this time, other users who have shared this blog will be able to know the location or address of the KINTEX through a map search to find the location information of the KINTEX while reading the post.

On the other hand, as a local information providing service using a map service, Gyeonggi 21 search system developed by Korea Aerospace University, the user searches for local web documents and compares the web documents with a map to obtain spatial information, and repeatedly This system was developed to solve the limitations of existing local information retrieval services that should be performed. The system utilizes geospatial words included in web geospatial information to simplify the process of searching for local information and is based on a three-tier spatial model called "reality, knowledge, and application" from the current web space. It became. By using this system, users can efficiently obtain web georegional information, various relational information about geographic characteristics, and spatial information simply by exploring the connection structure in which information between layers is related.

The Competition 21 Search System provides related keywords and expresses them on a map basis. The system does not consider providing various kinds of regional information at a glance. In other words, the system summarizes the information in the form of keywords and provides links to related web documents. In order to provide integrated local information, it is necessary to systematically and automatically organize various types of information.

The existing map service and blog service discussed above have drawn several problems as follows. First, it is necessary to read the web documents of the search results one by one, and the user is troublesome to analyze and synthesize various pieces of information that exist individually on the web. Second, there are inconveniences that the geographic information needs to be additionally checked through the map. Third, the search process is repeated to enter keywords and check results in order to obtain related information.

Therefore, there is a need for a system that enables a user to acquire detailed, comprehensive and easy information related to a specific region.

Accordingly, the present invention has been proposed in accordance with the need for building a local portal that can solve various problems occurring in the existing local information providing system as described above and satisfy various demands of users for local related information.

The problem to be solved by the present invention is the map interface and knowledge processing to connect the various local information included in the web document to display on a map to facilitate the user's information acquisition and to efficiently provide cultural tourism information It is to provide an integrated search method for local information.

Another problem to be solved by the present invention is to integrate information of various forms (blogs, photos, etc.) to easily grasp relevant information about a specific place, and utilize a map interface and knowledge processing, which eliminates an additional search process for obtaining geographic information. It is to provide an integrated search method for local information.

According to a preferred embodiment of the present invention for solving the above problems, "local area information integrated search method using a map interface and knowledge processing",

A first step of extracting a representative keyword representing a region and collecting and grouping various documents about the region from the web based on the extracted representative keyword;

Extracting a unique region keyword representing the region to generate related knowledge between the keywords;

And a third step of providing integrated regional information through a map-based interface so as to intuitively grasp relevant information of the extracted unique region keyword based on geographical information.

The various documents of the first step,

It includes blogs, news, dictionaries, web documents, local participation information, aerial photos, digital maps, and GIS data for the region.

Unique region keyword extraction representing the region of the second step,

It is characterized by extracting cultural relics and cultural facilities with unique local keywords representing the region.

Generating the associated knowledge of the second step,

Removing the tags of the collected web document, parsing the removed web document, and extracting text;

Analyzing the morpheme for the extracted text;

Extracting an associated keyword;

Calculating an association degree of the extracted association keyword.

The correlation calculation may be performed using a similarity matrix.

The related information of the extracted unique region keyword of the third step is

It is characterized by reconstructing related keywords centering on extracted unique region keywords and providing them as related information.

The map-based interface of the third stage,

It provides objective local information based on cultural and historical facts and subjective local information based on user's experience or opinion.

The subjective local information is characterized by including blog information.

The objective area information,

Encyclopedia, news, images, group information, including information.

The third step,

The relevance between the unique area keywords is displayed on a map, and the relevance between the keywords excluding the unique area keyword is displayed on an area except the map display area.

In addition, the map-based interface of the third step,

It is characterized by providing both semantic and semantic relations based on location or distance.

According to the present invention, since information generated by users and information providers is integrated and provided in a local information service specialized for a specific region, such as a living area information service and a regional portal provided by a local government, information separated in various places on a web space is provided. There is an advantage that can generate added value that cannot be obtained by the local people, improve the competitiveness of local governments, and improve the quality of life through the information activities of local people.

In addition, according to the present invention, since an additional search process for obtaining geographic information is unnecessary, there is an advantage of providing a user with convenience of searching.

In addition, according to the present invention, by connecting a variety of local information included in the web document to display on the map, there is an advantage that can be facilitated for the user information acquisition.

1 is a conceptual diagram showing a process of obtaining local information in a general web environment.
2 is an exemplary diagram of an existing blog service.
3 is a conceptual diagram of a system to which a method for integrated local retrieval information using a map interface and knowledge processing according to the present invention is applied;
4 is a flowchart illustrating a method for integrated regional retrieval information using a map interface and knowledge processing according to the present invention.
5 is an exemplary view to which the present invention is applied.
6 is an exemplary diagram of Goyang-shi numerical map.
7 is an explanatory diagram showing a subjective regional information processing process in the present invention.
8 is an exemplary view of a map interface in the present invention.
9- (1) is an exemplary view of the first screen of the user interface, and FIG. 9- (2) is an exemplary view of the screen when selecting the row destination acidity.
10 is a change diagram of data displayed on a map interface.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail.

FIG. 3 is a conceptual diagram of a system to which an integrated search method for local related information using a map interface and knowledge processing according to the present invention is applied. Is a model for helping to provide efficient service based on map interface when it wants to acquire relevant information of a specific area. 'Local information web document layer' grouping individual existing local information, 'unique regional keyword layer' where semantic correlation knowledge is generated between keywords, and 'map interface' that displays keywords connected by semantic relevance on map Layer '3-Layer.

The system to which the integrated search method for local information using the map interface and knowledge processing according to the present invention of this concept is applied generates a unique area information representing a region by using a digital map and a local government homepage, After collecting various types of web pages such as web documents, blogs, images, encyclopedias, and generating related keywords through stemming, extracting related keywords, and calculating relevance for each web page, it generates unique keywords and unique local keywords. Connect and generate association knowledge. After that, it provides a map-based user interface so that users can intuitively grasp the relevant information. Here, the user interface provides the local information by integrating the objective local information and the subjective local information so that the user can obtain detailed information, and provides the user with the convenience of information retrieval by providing related information at once.

A system for achieving the concept of the present invention uses a personal computer having a communication module capable of Internet and a control device capable of an information processing function. The control device includes an aerial photograph, a digital map, and geographic information (GIS data). Will be used.

FIG. 4 is a flowchart illustrating a method for integrated search of local information using a map interface and knowledge processing according to a preferred embodiment of the present invention. The representative keyword is extracted from the web based on the extracted representative keyword. A first step (S100) of collecting and grouping various documents about a region; A second step (S200) of generating a related knowledge between the keywords by extracting a unique region keyword representing the region; Based on the geographic information, the third step (S300) provides integrated area information through a map-based interface so as to intuitively grasp the related information of the extracted unique area keyword.

As described above, the "local related information integrated search method using a map interface and knowledge processing" according to the present invention comprises first extracting a representative keyword representing a region as a first step (S100) (S101) and extracting the extracted representative keyword. Various documents about the region are collected and grouped based on the web (S103).

In the present invention, the local information is divided into 'own area information' and 'general area information'. In order to map local information to a map, local information must include 'location' information. Cultural facilities (libraries, public sports facilities, etc.) and cultural relics (cultural assets, temples, acids, etc.) must have a location (location), and there is information on places held at local festivals and events. At the same time, if it includes information that characterizes a particular area, the area information can be mapped to a corresponding location on a map, which is referred to as 'unique area information' in the present invention. For example, information specific to a specific area, such as local festivals / events, cultural assets, and cultural facility information, is unique area information. 'General area information' is local information, not unique local information, and is general local information existing in any area. For example, there are Goyang City Hall and Goyang Police Station in Goyang City, and Gimpo City Hall and Gimpo Police Station in Gimpo City. The information on these facilities is not characteristic of a particular area. In this sense, general area information mainly includes educational institutions such as government offices, schools and academies, and company information such as hospitals and banks.

The local information web document layer of FIG. 3 is a layer for grouping local information related to web documents including local information by document type (encyclopedia, blog, image, video, etc.) based on the above definition. For example, an encyclopedia entry of 'Haengjusanseong', a news article called 'Hangjusanseong's 20,000 haunted house', and a blog titled 'Hangjusanseong Fortress' can be grouped with the same local information. By grouping related web documents and classifying them into document types, various types of local information such as encyclopedias, news articles, and blogs can be efficiently processed and provided to users.

Next, in the second step (S200) to extract the unique region keywords representing the region (S201), to generate the associated knowledge between the keywords (S203).

Each group of the local information web document layer grouping related local information by the same type of web document may be represented by a specific local related keyword. For example, web documents related to hangsan acidity may be represented by the keyword 'hangsan acidity'. As such, 'North Korea' and 'Ilsan Lake Park' may be keywords that represent and characterize Goyang. In this case, it is considered that such a keyword may be mapped to a specific position (coordinate) of the real life space. This is because the proposed model uses the map interface to efficiently display area-related information. That is, in the present invention, a keyword that can represent a group of related unique area information and can be mapped to a specific location (coordinate) of a real life space is defined as 'own area keyword'.

In the unique region keyword layer, semantic association knowledge between unique region keywords representing unique region information is generated. First, for the extraction of unique local keywords, web site, map platform, etc. are analyzed to extract the unique local keyword candidate set. After generating a set of unique local keywords, the association between the keywords is calculated and new knowledge is generated from them. For example, it is easy to see that 'Haengju Fortress' and 'North North Fortress' are related to acidity, but it is difficult to infer this systematically. This is because most of the information displayed on the map of existing map services is merely information based on location or distance. A method of generating a unique region keyword and generating relevant knowledge will be described later.

Next, in the third step (S300) to provide an integrated information through the map-based interface to intuitively grasp the relevant information of the extracted unique region keywords based on the geographic information.

'Map Interface Layer' is a layer that expresses the relevance between the keywords of the unique region based on the map. The keywords of the unique region selected by the user are mapped to the actual position (coordinates) of the map. Assume that the user selects a unique region keyword called 'hangsansanseong' among the keywords displayed and mapped on the map interface. At this time, the documents grouped by the hangju acidity are selected among the documents existing in the local information web document layer, and the keywords ranked at the top among the unique area keywords related to the hangju acidity are selected in the unique region keyword layer. The selected documents and associated keywords are finally represented on the map. And the keywords mapped to the map with the central keyword in this layer are keywords with some interesting associations.

Representation of the information based on the map in the map interface layer guarantees the intuitiveness of information acquisition. Information presented on the map can be grasped at a glance in the visual aspect, which helps the user's comprehensive judgment. In addition, if a document includes contents such as a place name or an event venue, an additional search for location information is not necessary.

Referring to the features of the present invention as described above in detail as follows.

For example, by applying the model proposed in the present invention, the integrated search service that provides information related to Goyang-si, Gyeonggi-do, can be implemented in a four-step process as shown in FIG.

That is, (1) extract keywords that can represent a particular region, and (2) extract related web pages after collecting relevant web pages. Next, (3) the correlation between keywords is calculated by applying the association knowledge generation algorithm. Finally, (4) construct a user interface using maps so that users can intuitively grasp. In relation to the proposed system model, steps (1) and (3) correspond to the unique region keyword layer, (2) the local information web document layer, and (4) the map interface layer.

First, the first step is to extract the names of cultural assets, cultural facilities, administrative institutions, mountains, etc. of Goyang-shi from the numerical map of Goyang-city in order to obtain unique-region keywords from the information on the unique region. The extracted words are compared with the words extracted from the Goyang City Hall homepage to obtain a final set of unique local keywords.

The concrete method is as follows. First, keywords that can be mapped to coordinates are extracted from the numerical map. A digital map is a computerized map that allows geospatial information to be classified by layer into a database and maps can be freely created according to the purpose of use. Since each object of the digital map is classified as a standard code determined collectively, related words can be extracted based on the code. FIG. 6 is a numerical map file including surroundings of Ilsan Lake Park, Janghang-dong, Ilsan-gu, of 1099 numerical maps in Goyang-si. If you zoom in on the part, you can see that the texts such as 'Ilsan-gu', 'Janghang-dong' and 'Janghang 2-dong' are mapped to the location. In this way, the numerical map was analyzed and 4614 words related to Goyang-si were extracted through text information such as facilities classified by codes.

The extracted words include the unique region keywords defined in the present invention, but also include keywords that are not unique region keywords. This is because the Goyang digital map contains all the facility information that can be mapped to the coordinates of the whole Goyang city, and contains many words other than the words that characterize Goyang. Therefore, in the present invention, the 'cultural tourism' guide page of the Goyang City Hall homepage was used to select only unique keyword. The reason for using this page is that the most representative of the sites that represent a particular area is the local government homepage, and most of these homepages contain a lot of information that represents the area. In addition, the homepages of public offices such as cities and counties have a strong publicity, and thus can be regarded as having high reliability because information of clearer and more objective information is delivered. In the 'Cultural Tourism' guide page of Goyang City Hall website, related words can be extracted from each page divided into categories such as 'Cultural Events / Festivals', 'Cultural Relics / Things', and 'Cultural Facilities'. By taking the intersection of the extracted words and the words extracted from the digital map, we finally obtained 133 keywords of unique regions in Goyang.

In the present invention, since only the 'cultural tourism' guide page in the Goyang City Hall homepage is used for extracting the unique region keyword, the unique region keyword can be extracted relatively accurately.

In the present invention, local information is divided into subjective and objective viewpoints. Blogs are categorized as subjective local information, and web documents such as encyclopedias, news, and images that contain other regional information are classified as objective local information. The reason why the blog is classified as subjective local information is to express it as a keyword frequently recorded among various contents such as the experiences, subjective thoughts, and feelings of the users recorded in the blog during the system implementation process so that it can be shared with other users. For sake. Relevant web documents were collected using NAVER open API. NAVER open API is an open interface to use various data, contents, and services of NAVER, and various Mash-up services can be developed through this.

To collect objective local information, we collect links in search results by querying each unique local keyword for encyclopedias, news, images, and other web documents. The collected links save the web document and extract only the text. From this, it is possible to know the frequency of occurrence of the unique area keywords included in the document, which is then used to select 10 related keywords displayed on the map in the process of calculating the degree of association. In addition, detailed information of Goyang-shi, which cannot be obtained from encyclopedias or news, supplements the objective area information using Goyang-gun data.

Similarly, blogs belonging to subjective local information are collected using the NAVER open API, and only the text of the document is extracted from each link. As shown in FIG. 7, the extracted text generates a related keyword set of a noun type through a process of tag removal, morphological analysis (noun extraction), and general noun removal. Next, it is necessary to extract only Goyang-related keywords from the generated noun set. Based on the reliability of existing search engine search results, if all the words in the generated noun set are sent to the search engine as a query, the query and the word 'Goyang' or 'Cat' are found on the search results page a certain number of times. It was determined as a keyword related to the Goyang region. These keywords are used to calculate the degree of association with the keyword centered in the map interface.

According to the present invention, ten unique local keywords most relevant to a keyword selected by a user are displayed on a map to indicate their relevance. In addition, 10 relevant keywords are extracted from the blog document and provided together. The criterion that any two keywords are related to is the aerodynamic of 'appearing in the same document'. Data mining techniques are used to determine the relevance between keywords. Here, we examine the results of data applying two methods of association rule and similarity matrix.

Association rules are a typical technique of data mining, which is a technique for finding associations between items purchased at department stores or supermarkets. Here, 'buy together' is called a transaction, and each element of the data we deal with is an item in a shopping cart. At this time, the data set included in one shopping cart is called a transaction. Association rule is a data mining technique that targets rules with a certain degree of support and counts them as one even if they appear several times on a page. Since it focuses on which products are purchased together rather than the number of products in the shopping cart of the product buyer, the occurrence number of keywords appearing on a web page several times is ignored. In addition, in general, the association rule may be applied in a case where it is universally valid, but the system to be implemented in the present invention is not a proper method because it is a special case of dealing with 'locality'.

The similarity matrix may indicate the degree of relevance between words such as 'the number of documents in which two words are held'. Creating a first document index term w _i m _ij is the matrix by measuring the frequency of occurrence in a few times that d _j, creating a transposed matrix m _{ij ij} m ^t of the product of two matrices s = m _ij × m makes _ij ^t. Each element of the matrix s is a measure of how many words appear in the same document. The matrix m _ij created using web documents and keywords is a sparse matrix with most elements zero. Therefore, by using only nonzero elements of a matrix, memory usage and computation can be greatly reduced. Using k documents and l index words to build a matrix and calculate the degree of association:

① Create a matrix m _ij by measuring the frequency of how many times the index word w _i appears in document d _j .

② Create the matrix m ^t = m _ji .

③ Calculation of Association s = m _ij × m _ji

The similarity matrix calculates the degree of association between all keywords by using keyword count information appearing on one page and generating a keyword-keyword matrix. The measure of how often two keywords appear in the same document at the same time is the measure. The more the two keywords appear in many documents at the same time, the higher the relevance is. Unlike the association rules, the number of times that appears in each document also contributes to the relevance, so that even if the number of times that appears in several documents is high, the number of times that appears in one document is high.

Calculation of the degree of association between the unique region keywords used in the present invention is as follows. First, create an A matrix of keyword-documents. The row of matrix A is 133 unique region keywords, and the columns represent each document collected by the relevant web document collection process, and the frequency of keywords appearing in each document is generated as a sparse matrix. Next, a matrix B of document-keywords, which is a transpose matrix of matrix A, is generated. Row B is each web document and column is a unique local keyword. The calculation method for obtaining the keyword-keyword matrix, that is, the A × B matrix is as follows. In the file recording the matrix A, information about the document in which the 'hangjusanseongseong' appears, such as document number 112, 1009, and 2080, is recorded as shown in Table 1 below. The frequency of appearance of the associated keywords appearing is recorded. In this case, the A × B matrix reads only the document in which 'rowability' appears, such as document numbers 112, 1009, and 2080, and calculates the degree of association by multiplying the frequency of appearance of other keywords appearing in the document.

Document information about the appearance of Haengsan Fortress Document number Appearance frequency Document number Appearance frequency 112 2 4568 5 1009 3 5871 12 2080 One 10551 3

Keywords in the region of Goyang, such as 'Haengju Fortress', 'Bukhansan' and 'Ilsan Lake Park', have a cultural, historical or geographical meaning. Also, keywords appearing with these unique local keywords on web pages are likely to have similar meanings. For example, in the case of Mt. North Korea, a list of unique local keywords appearing simultaneously on a web page is shown below.

Bukhansan Fortress, Kuknyeongsa Temple, Nojeoksa Temple, Taegosa Temple, Muryangsa Temple, Sangunsa Temple, Heungguksa Temple, Haengjusanseong Fortress, Seosamneung, Lake Park

These related keywords will appear more frequently when 'North Korea' appears relatively frequently in the document. This is because the document itself reflects cultural, historical and geographical specificities. At this time, Kuknyeongsa Temple, Nojaksa Temple, and Bukhansan Fortress are located close to Mt. Bukhansan through existing map services, so it can be intuitively assumed that they are geographically related. Notable is the relevance of Lake Park, which is about 20 km away from Bukhansan Mountain. Similarity matrices can be used to create not only distance-based associations, but also interesting relationships that are difficult to infer intuitively. As another example, meaningful related knowledge is created, such as Hongneung-Latin American Culture Center, Haengju Sanseong-lake Park, Goyang Aram Nuri-Mokam Art Museum.

Similarly, for the set of related keywords extracted through the text processing of the blog, which is subjective local information, the calculation method of obtaining the A × B matrix is used. In this case, the row of matrix A is related keywords, the column is a document including the keywords, and the matrix B is a transpose matrix of matrix A. After relevance calculation process using similarity matrix, each relevant local keyword is ranked in order of high relevance related keywords, and then the top 10 keywords are displayed on the left side of the map interface to facilitate user synthesis and acquisition of local information. do.

On the other hand, in the currently implemented system, although it does not go all at once, the process corresponding to each chapter proceeds automatically. That is, the web document collection and association keyword extraction process is automatically executed at once, and the association calculation process is also automatically performed. Therefore, in terms of automation, the current system can guarantee the latest search results.

The user interface utilizes maps to provide an intuitive view of the relevant local information processed in this way. Implement in JavaScript, JSP, XML, MySQL (DB), etc. and write using AJAX technology. With AJAX technology, you can get the data you want without having to move pages, and you don't need to install any programs.

When a user selects a keyword on the map, the top 10 unique area keywords related to the keyword are displayed as markers on the map again. At this time, it is possible to obtain all related information by simply clicking on the marker, which increases the user's convenience of information search. Local information on the left and right side of the map is obtained by clicking the relevant information and obtaining the link of the relevant web page from the DB. 8 is the first screen shown when the user uses the service for the first time. When the present invention is loaded, the first central keyword is Goyang City Hall, and unique local keywords representing Goyang City are mapped to actual locations and displayed on a map. Among them, 10 keywords that are highly related to Goyang are connected with the central keyword by a line and marked with a red marker. In relation to the unique region keyword at the center of the map, the related keyword extracted from the blog as subjective regional information is arranged on the left side of the map, and the objective region information such as encyclopedia, news, image, and county is arranged on the right side.

When a user looks at a specific area using a map, most place names that have high spatial relevance and spatial proximity are displayed nearby. This relationship can be understood simply by looking at the map. However, if a semanticly relevant place name is provided regardless of spatial proximity, as shown in the present invention, it is easy to grasp the relevance even if it is far away from the map. Can find relevance for missing names. In addition, keywords are mapped with markers on the map. By simply clicking on the markers, the location of other names that are highly related to the name keywords can be easily identified.

The embodiment system implemented in the present invention was operated in a Windows Server having a specification of Intel Xeon 2.66Hz (8 CPUs), 3.25GB RAM.

Looking at the operation of the system in detail as follows. Referring to FIG. 9- (1), it can be seen that Haengju Fortress, Bukhansan, and Seoo-reung, which are connected to the central keyword 'Goyang City Hall', are the most representative keywords of Goyang City (①). Next, click on 'Encyclopedia' on the right side of the first screen centering on Goyang City Hall (②) and select 'Chungjangsa (Goyang City)' among the encyclopedia items related to 'Goyang City'. Can be. If you click (④) on 'Hangju Acidity', the keywords in the map area will be rearranged based on Haengju Acidity, one of the representative keywords of Goyang, as shown in Fig. 9- (2). It is connected by a line (⑤). If you click (⑥) on the right side of the map in this state, 10 images of 'Hangju Sansung' are displayed as thumbnails on the right side of the screen, and the image of the original size is provided when the thumbnail of each image is clicked. If you click (⑦) the keyword 'Feast Noodles' because you are curious about the relationship between Haengju Sanseong and 'Feast Noodles' among the blog related keywords located on the left side of the map, select one related blog (⑧) After reading it, we can see that there is a famous collection of aid soup in Haengju Fortress. This can be guessed through the title of 'Haengju Sansung Assisted Noodle House Collection' or 'Hangju Sanseong Delicious Restaurant / Noodles' in the list of blogs. FIG. 10 illustrates a change of data displayed on the map interface when the central keyword is changed from 'Goyang-si' to 'Hangju Sanseong' in FIG. 9.

The proposed system solves the limitations of existing services as follows. First, the system proposed by the present invention provides a service that automatically aggregates and organizes information desired by a user for various types of information existing on the web. For example, you can easily access related blogs by simply selecting the blog information about the specific place name you want to know as a related keyword on the left side of the map. In addition, the encyclopedia, news, images, Goyang group information is shown on the right side. As such, by automatically integrating and arranging various types of information on the web and providing relevant information to the user, it is possible to reduce the user's self-organization and synthesis of various pieces of information existing on the web. Second, by providing local information based on a map, an object to be searched by a user is displayed on a real location on a map, thereby reducing additional geographic information search. Third, the proposed system processes names / non-name keywords that are semantically related to the keywords selected by the user through an association calculation algorithm, and automatically selects and provides only relevant information. In addition, since the relevant local information can be easily obtained by simply clicking on a relevant local keyword or a blog related keyword on the left side of the map, unnecessary keyword input occurring to obtain related information can be reduced.

Further matters required by the present invention are as follows. First, information about 'news' or 'blog' among objective local information is static information at present. For example, you can't easily keep up-to-date on what's going on in a place. This will be solved by conducting research for providing information reflecting real-time information by scrapping and analyzing articles related to local newspaper events / festival periodically and providing them in a form that users can easily grasp. The second is a requirement for a customized interface that can change the menu configuration according to the user's preference. This is further implemented through a methodology study to provide only the information desired by the user such as blog, encyclopedia, and news. This can be solved.

According to the present invention as described above, in order to provide a system that enables the user to obtain detailed, comprehensive and easy to obtain the relevant information of a specific region, a three-layer of integrated search for the region-related information using a map interface and knowledge processing We propose a system model. Based on the proposed model, we proposed the four-step construction method of extracting unique keywords, collecting related web pages and generating related keyword sets, calculating the degree of association between keywords using the associated knowledge generation algorithm, and creating a user interface using maps. . In addition, by establishing and operating the system in Goyang-si, Gyeonggi-do, it has advantages in terms of convenience of searching relevant local information, intuitive information recognition, search by information type, and comprehensive information judgment. Confirmed. Through this, we verified the validity of the proposed model and the knowledge processing algorithm of local information using the similarity matrix and the UI for the user's convenience.

The present invention simply fuses local information existing as individual 'information', produces and organizes new 'knowledge' and provides it to users. Through this, the user can be provided with various detailed local information, and the related information can be easily obtained.

It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined in the appended claims and their equivalents. Of course, such modifications are within the scope of the claims.

Claims (10)

Extracting a representative keyword representing a region, and collecting and grouping various documents about the region from the web based on the extracted representative keyword;
Extracting a unique region keyword representing the region to generate related knowledge between the keywords;
And a third step of providing the integrated local information through a map-based interface so as to intuitively grasp the related information of the extracted unique local keyword based on the geographical information,
The map-based interface,
An integrated retrieval method for local information using a map interface and knowledge processing, which provides objective local information based on cultural and historical facts and subjective local information based on user's experience or opinion.
The method of claim 1, wherein the various documents of the first step,
Integrated search method for local information using a map interface and knowledge processing, including blogs, news, dictionaries, web documents, participation information of local people, aerial photographs, digital maps, and GIS data.
The method according to claim 1, wherein the unique region keyword extraction representing the region of the second step,
Integrated retrieval method for local information using a map interface and knowledge processing, characterized in that the extraction of cultural relics and cultural facilities as unique local keywords representing regions.
The method according to claim 1, wherein the association knowledge generation of the second step,
Removing the tags of the collected web document, parsing the removed web document, and extracting text;
Analyzing the morpheme for the extracted text;
Extracting an associated keyword;
Comprising a step of calculating the degree of association of the extracted relevant keywords, the integrated search method for local information using the map interface and knowledge processing.
The method according to claim 4, wherein the correlation calculation is performed using a similarity matrix.
The method according to claim 1, wherein the relevant information of the extracted unique region keyword of the third step,
Integrated retrieval method for local information using a map interface and knowledge processing, characterized in that the relevant keywords are reconstructed based on the extracted unique area keywords and provided as related information.
delete The method according to claim 1, wherein the objective area information,
Integrated search method for local information using a map interface and knowledge processing, including encyclopedia, news, images, and county information.
The method of claim 1, wherein the third step,
Integrated search method using local map interface and knowledge processing, characterized in that the relation between the keywords of the unique region is displayed on the map, and that the relations between the keywords except the unique region keyword are displayed in the region except the map display area. .
The method of claim 9, wherein the third step,
An integrated retrieval method for local information using a map interface and knowledge processing that provides location and distance-based and semantic relevance together.

Images