KR101051422B1 - Record media recording the automatic completion system, method and program for each query type with guaranteed search results - Google Patents

Abstract

Disclosed is a recording medium recording an automatic completion system, method, and program for each query type for which a search result is guaranteed. To this end, the search system that inputs the query automatically presents only the query result that guarantees the search result. In particular, the list and type of the query result in which the search result exists is presented and the document information is added or deleted. Since it can respond in real time, etc., it is effective to increase the reliability of the search results, to prevent the failure of the search, to speed up the search, and to provide the improved status by grouping the information by type.

Autocomplete, Search, Document Information, Index, Frequency, Query, Database, Server, Topic

Description

Recording system that records auto-complete system, method, and program for each type of query with guaranteed search results.

The present invention relates to a query search system. In detail, the present invention provides a list and types of query words in which a search result exists for an input query, and provides automatic search by type of query for which a search result can be added in real time. It relates to a recording medium recording the complete system, method and program.

The modern society is the age of the flood of information that can easily access a large amount of information through either the paid or the free method from the previously published information to the latest information by the development of data communication using computers, the Internet, and communication networks. .

Such information includes text, sound, pictures, movies, multimedia, and the like. In general, the content of information is recorded in text.

In addition, as the amount of information increases, there is a need for a device for recording and managing information and a technique for quickly and accurately finding necessary information when needed. Such a technique is generally called a database (DATABASE: DB), and the database dedicated computer is a server. It is called (SERVER), and anyone accesses the server from a long distance through a data communication network such as the Internet, and easily retrieves and utilizes necessary information.

The database DB includes all text, sound, image, video, and multimedia information, but mainly records text information in order to record and manage maximum information in a predetermined storage capacity.

Textual information includes a human-readable vocabulary and a computer-readable vocabulary that contains a program.

In order to quickly search for the desired contents from a large amount of information stored in a database, a standardized vocabulary term is required for both humans and computers to exchange information and know at the same time. This standardized vocabulary is called a collection of concepts or an ontology. In order to selectively search for desired information on the web of the Internet in which a large amount of information is shared using the ontology, a semantic web technology is used.

In general, there is a need for information retrieval in order to obtain professional knowledge or to make accurate judgments and decisions on management, and to quickly find desired or necessary technical information from a large amount of accumulated technical information. It becomes a technical field.

Vocabulary or terms input from the database server for information retrieval are also referred to as QUERY, ENTITY, or recommendation.

In the autocomplete display method, the searched query is displayed by determining that the query part is partially input as a completed state.

In the case of a query such as a name, an address, a name, etc. repeatedly inputted in a web browser or other data retrieval software, the autocomplete method displays a list of previously searched queries and displays one of the displayed query lists. To make a quick selection.

The autocomplete method is implemented using Asynchronous JavaScript and XML (AJAX) technology, which is one of the Web 2.0 technologies, and various fields including portal internet sites, digital libraries, ENTERPRISE WEB 2.0, and specialized application programs. It is widely applied in.

As the above-mentioned autocompletion method is proved as a preference in terms of enhancing the user experience, it may be used more widely in the future search interface.

However, the autocomplete methods provided up to now have been used to list the query terms using the stored log information or its own dictionary according to the user query input. Therefore, if the query to be searched is not displayed at the top of the displayed list, There is a problem that you must look through the autocomplete list sequentially.

In addition, unlike a portal site with a large amount of content, in the case of a private enterprise or a specific application field, the search function can be provided by presenting an autocomplete list based on the frequency of search terms even when the search success cannot be guaranteed due to relatively poor content. There is a problem of low reliability.

1 is a functional configuration diagram of a system for retrieving information from a general database system.

Referring to FIG. 1, a concept of retrieving necessary information from a database server by text information is described. A large amount of information composed of text is recorded and managed in a database server.

In order to retrieve necessary information from the text information recorded in the database server, a query is input to the computer terminal.

The computer terminal is provided with a search program (PROGRAM), analyzes the query word (object) input by the search program, and retrieves index information corresponding to the query word from the database server.

A plurality of index information including a query is searched and provided as a list. When any one of the lists is selected, the corresponding information is finally searched and output to the computer terminal.

The information retrieved is used as a reference for obtaining higher and more diverse knowledge or for making decisions and judgments by managers or managers.

The amount of information recorded and managed in the database server increases greatly as the knowledge and science develops, and there is a problem that it takes a lot of time to search and analyze the desired information by the input query.

FIG. 2 is a diagram illustrating a query word input and searched in an autocomplete manner by way of example.

Referring to FIG. 2 in detail, a query list searched in an autocomplete manner in a state where “large” is entered as a query is shown. The searched list is displayed classified as forward match and backward match. In FIG. 2, Korean is input as a query language, but a language such as English may be used.

The searched input words are automatically completed and displayed in a forward match method such as Korean Air, Supreme Court, Daesung, and Daegu Bank, and the rear matching methods include Yongdae, Girls' Generation, and Dankook University.

Queries, such as Korean Air, Daegu Bank, and Korea Express, which are searched as query words and marked as autocomplete, are classified as TYPE, Korea is classified as country type, and Daesung and Yongdae are classified as person types. Are classified into group types and the like.

If you enter a query to search in many portal sites of Korea, for example, Naver (Naver, www.naver.com ), the entered query is included and the search is completed and provided as a list. .

The query is completed while the input continues, and since a desired query is selected from a list of auto-completed queries even when a part is input, the query time for searching is shortened and it is convenient for use.

In the case of a company that does not have a large amount of content, such as a portal site, or a specific application, the search results (successful presentation of the search results) for the query input by the relative content poor cannot be guaranteed, but only by the frequency of input. Presenting a complete list reduces the reliability of the feature.

3 is a diagram illustrating a state in which a query is input and a search fails due to an example of the related art.

Referring to FIG. 3, for example, the user accesses the 'best buyer ( www.bb.co.kr )' site, which compares the price of the product, and enters ' apina ' as a query for searching. For the input query word, 'apina, apina table' is searched for and displayed as autocomplete.

Searching for index information for a query is marked as failed. For example, the index information of the searched query word "apina" and the index information of the "apina table" are displayed as failed. The cause of such a failure may be, for example, deleted due to the lack of product sales, inventory exhaustion, expiration of the corresponding information, and the like.

In spite of the above causes, the user has a problem that the reliability of the system is lowered due to a search failure.

As an example of the prior art, according to A. Bangalore, A. Browne, and G. Divita, “UMLSKS SUGGEST: An Auto-complete Feature for the UMLSKS Interface Using AJAX”, In Proceedings of AMIA, 1106 As an autocompletion feature, a method for presenting only queries that have generated a successful search result is described.

In the prior art as described above, it is determined whether or not an autocomplete list is presented by setting a case where a query search is successful by a flag FLAG.

However, when content including a query that failed to be searched is added later, there is a problem that it is not presented as an autocomplete list until it is entered as a query and a search is attempted.

In addition, M. Takasi et al. The prior art by “Auto Complete Method for Web Application Form Based on Term Hierarchy”, In Proceedings of the Annual Conference on JSAI ( in Japanese ), 1106. relates to autocompletion and is compatible with query lists in different applications. It is related to a technology that supports format conversion so that it can be used, and its coverage is very different and unrelated because it does not provide a list of guaranteed autocompletions.

In addition, Kwang-Jo Lee, Jin-Woo Song, Han Jung-Seok, Yang Sung-Bong, “Location Based Query Recommendation System for Mobile Terminals”, Korean Information Science Society Fall Conference, 1107. It is a technology that overcomes the limitations of space and considers user location information, but also does not provide an autocomplete list that guarantees a search result, so the range of application is very different and unrelated.

Also, as prior art, W. Sung, H. Jung, P. Kim, I. Kang, S. Lee, M. Lee, D. Park, and S. Hahn, “A Semantic Portal for Researchers Using OntoFrame”, In Proceedings of the 6 ^th International Semantic Web Conference, 1107.

OntoFrame in the prior art is built to provide academic research information analysis services based on semantic web standard technologies XML, RDF (Resource Description Framework), OWL (Web Ontology Language), SPARQL Protocol and RDF Query Language (SPARQL), etc. Semantic web services framework.

In the prior art, an existing DB is collected by referring to a modeled ontology and converted into an RDF triple format, and OntoReasoner, an inference engine, uses this as knowledge.

In addition, as the prior art, there are Han Min-min, Kang In-soo, and Sung-Kyung Sung, "Assigning Topics and Sectors to Scientific and Technical Literature Using Thesaurus and Discipline Classification System", Korean Society for Linguistic Information, Summer Conference, 1106.

The prior art is responsible for collecting and converting information in a URI server, and performs a document classification function by extracting a query word from an original text and assigning it to the original text.

The OntoFrame service of the Semantic Web Services Framework provides a query-centered, integrated search function, similar to the vertical search of portal site Naver.

That is, to identify a specific query type and generate search results for that type. For example, if a user inputs the name of a person named “Christian Becker” as a query word, related researchers (Similar Researchers), researchers with citations ( Researchers in Citations, etc. If the topic “Semantic Web” is entered as a query, Topic Trends, See Also, Researchers by Topic, and Papers by Topic Topic), Researcher Network, etc.

In particular, when providing a search result for a query, the subject extraction result that was performed in the URI server is used. The result of extracting the key word is propagated to the person, organization, etc. by the inference engine, so that it is possible to compose the subject trend, thematic experts, and thematic papers. Therefore, there is a problem that a correct semantic web service cannot be constructed unless the document information is added. .

In the OntoFrame service according to the prior art, autocompletion was also provided, and among the queries in the query dictionary used for query extraction, those matching the user's query were presented as an autocompletion list.

However, even queries that are not included in the extracted queries are presented as an autocomplete list, which results in a case in which there is no search result or a poor search result.

In addition, a large number of query words matched with a string input as a query word causes a large autocomplete load.

Therefore, there is a need to develop a technique for solving the above problems using a method of controlling the autocomplete list with extracted query terms.

In addition, there is a need to develop a technology that enables autocompletion for each query type by performing a query type recognition at the stage of autocompletion presentation.

In addition, there is a need to develop a technology for automatically providing a result of a successful search for an input query.

The present invention has been made in order to solve the problems and necessity of the prior art, in particular, the automatic completion system for each type of query is guaranteed to ensure that the search results presented by the automatic completion only when the result is guaranteed by searching the input query and It is an object to provide a recording medium on which the method and program are recorded.

In addition, the present invention provides an autocompletion system for each type of query, which provides accurate, fast, and secure search results, and provides a recording medium recording the search results of an input query. Purpose.

In addition, the present invention records the automatic completion system, method and program for each type of query, which guarantees the search result that reflects the occurrence frequency value of the query by the addition and deletion of document information in real time and provides the improved information by grouping the information by type. The purpose is to provide a recording medium.

In order to achieve the above object, the present invention provides a document index server which extracts index word information and frequency information from registered document information and records the document information and records the index information in the index database and generates autocomplete list information from the extracted index word information. After searching the autocomplete database and the autocomplete database, which records the autocomplete list information generated by the document index server in association with the frequency information, extracts the autocomplete list information including the index information and converts it to the query language. It is configured to include an autocomplete server which provides the user interface by converting the selected input query word into an index word and searching document information including the index word.

Preferably, the document collecting unit for accessing the document index server to register the document information collected, and further comprising an index database for recording the index word information provided from the document index server and providing the index word information by the search of the autocomplete server Present the configuration.

The document collection unit may be configured to collect one or more content document information including web page document information, format document information, image document information, video document information, text document information, and multimedia document information.

In addition, the document index server includes a document register that inputs and registers new document information, a document index unit that extracts an index word from document information registered by the document register, and stores the index word in an index database, and the index word stored in the index database. And a database generation unit for retrieving the index word information provided, recording it in an autocomplete database, and updating and managing frequency information.

The document indexing unit may be configured to extract additional information including index word information from at least one selected from extracting an index word from document information registered in the document register and extracting index word information designated by text processing. It is done.

The document indexing unit may be configured to extract an index word from any document information registered in the document registration unit in one of a morpheme analysis method and an N-gram method and store the index word in the index database. .

In addition, the document indexing unit is configured to record additional information including the extracted index word in association with the document information and store it in the index database.

In addition, the document index server is characterized in that the configuration further comprises a document editing section for modifying and deleting the document information registered in the document register.

The document indexer may be configured to remove unnecessary index words included in the stopword dictionary from the index words extracted from the document information registered in the document register.

The database generating unit may be configured to accumulate and record the frequency information of the document information as an autocomplete list object of an autocomplete database unit.

The database generating unit may be configured to exclude from the target of completion of the autocomplete list that is cumulatively calculated when the value of the frequency information is zero.

In addition, the autocomplete server checks the frequency information of the query word input unit that receives the query to be searched into the user interface and converts it into an index, the database search unit that retrieves the index word from the query input unit from the autocomplete database, and the index information stored in the autocomplete database. Indexing unit that determines and provides the autocomplete list information, the presentation unit that converts the autocomplete list information provided by the index determination unit to the query interface to provide to the user interface, the input query and the query of the autocomplete list provided by the user interface user interface A selection unit for inputting the selected query word information together with an event signal and converting the index word into the index word; And a service linkage unit for searching and providing document information according to the index word information and the search event signal inputted by the selection unit.

In addition, the query input unit is characterized in that the query is composed of a unit letter input by any one selected from the phoneme, syllable, word, word.

In addition, the query input unit is characterized in that it comprises a configuration for retrieving the index by calling the autocomplete database in the AJAX method whenever the query is input.

The query input unit may include a configuration for inputting the query information into a user interface (UI).

In addition, the database search unit is characterized in that it comprises a configuration to create an autocomplete list by searching the index word in the forward matching and backward matching, respectively.

The index word determining unit may be configured to determine that the index information includes one or more frequency information in the autocomplete list.

The presenting unit may be configured to adjust the ranking in the autocomplete list by using at least one selected from the input statistical information of the query, the frequency information of the query, and the order of information of the query.

The service interworking unit may be configured to search for document information by calling the index word information by an API.

In order to achieve the above object, the present invention collects and registers document information, extracts index word information and frequency information from registered document information, stores it in an index database, and generates an autocomplete list information of index word information in an autocomplete database. Storing and modifying the registered document information, converting the query word input from the autocomplete database through the user interface into an index, including the index word searched with a frequency of 1 or more in the autocomplete list information, and the index word of the autocomplete list. The present invention provides a configuration including a process of converting a query language into a query interface, converting the selected query word from the outside into a user interface, and outputting document information retrieved from the converted index word.

Preferably, the modification process includes collecting document information by the document collecting unit, registering the collected document information by the document registering unit, extracting an index word from the registered document information by the document indexing unit, and storing the document information in the index database. Including extracting index words to be provided as an autocomplete list by the database generator from the index word information stored in the index database and storing them in the autocomplete database, and modifying or deleting registered document information by the document editing unit. Characterized in that made.

The index word extraction may be performed by extracting the index word by any one method selected from a morpheme analysis method and an N-gram index method.

In addition, the index word stored in the index database includes additional information and removes unnecessary index words using a stopword dictionary, and the updated frequency information of each document information is stored in the autocomplete database.

The input query word may be input in any one unit selected from a phoneme, a syllable, a word, and a word.

In addition, the input query is converted into an index word, characterized in that the search by calling from the autocomplete database unit in an Ajax manner.

In order to achieve the above object, the present invention collects and registers document information, extracts index word information and frequency information from registered document information, stores it in an index database, and generates an autocomplete list information of index word information in an autocomplete database. A process of storing and modifying registered document information, a process of converting a query word input from an autocomplete database through a user interface into an index word and including the index word searched with a frequency of 1 or more in the autocomplete list information, an index word of an autocomplete list A query with guaranteed search results, which includes a process of converting the query into the query and presenting it to the user interface, converting the selected query from the outside into the index, and outputting document information searched by the converted index. It presents the structure of a recording medium recording a program source of the type of auto-complete manner.

Preferably, the autocompletion list process of the query is characterized by a process of adjusting the ranking to any one or more selected from the input statistical information of the query, the frequency information of the autocomplete database, and the order of information of the query.

In addition, the input query process is characterized by a process of calling and searching the autocomplete database unit in an Ajax manner whenever it is input in any one unit selected from phoneme, syllable, word, and word.

In order to achieve the above object, the present invention establishes an autocomplete database by extracting index information and frequency information from registered document information after registering document information, converts a query word input from an external user interface into an index word, and autocompletes A server system and server for providing an automatic completion list of index words including index words from a database through a user interface, converting a query word selected through the user interface into an index word, and providing document information including the converted index word to the user interface. The public communication network which connects to the system and transmits and receives the query word and the data of the retrieved document information through the selected communication path among the wired communication path and the wireless communication path, and inputs the query word to access and search the public communication network through the user interface. It is a computer that displays the auto-completion list of query terms provided by the server system in the user interface, inputs one selected query word with event signal, provides it to the server system, and displays document information searched and provided by the server system. A configuration including a terminal device is provided.

Preferably, the server system registers document information, extracts index words and frequency information, and builds an autocomplete database. The document index server inputs query words from the outside into a user interface, converts them into index words, and extracts index word list information including index words. Autocomplete server that converts the selected input query into index and provides the retrieved document information, and automatically records the autocomplete list information generated by the document index server in association with the frequency information. Features include a complete database, a document collecting unit for registering collected document information by accessing a document index server, and an index database for recording index word information provided from a document index server and providing the search by an autocomplete server. do.

In addition, the public communication network includes a wireless communication network for connecting a server system and a terminal device to a wireless communication path and transmitting a data signal, and a wired communication network for connecting a server system and a terminal device to a wired communication path and transmitting a data signal. It features.

Therefore, according to the present invention, since the searched input query is automatically presented only when the result is guaranteed, there is a convenient effect of increasing the reliability of the search result.

In addition, the present invention has an industrial use effect to prevent the failure of the search and to speed up the search because it is presented as autocomplete only when the search results are guaranteed.

In addition, the present invention has a convenient effect of reflecting the occurrence frequency value of the query by the addition and deletion of document information in real time and grouping the information for each type to provide an improved state.

The terms or words used in this specification and claims are not to be construed as limiting in their usual or dictionary meanings, and the inventors may properly define the concept of terms in order to best explain their invention in the best way possible. It should be interpreted as meaning and concept corresponding to the technical idea of the present invention.

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

Example

In order to explain the present invention, FIG. 4 is a diagram illustrating a functional configuration of an automatic completion system for each type of query, in which a search result is guaranteed by an example of the present invention, and FIG. 5 is an example of the present invention. Figure 6 is a detailed configuration diagram of the server system of the complete auto-complete system for each query type, Figure 6 is a flow chart of the automatic completion method for each query type is guaranteed by the example of the present invention, Figure 7 is an example of the present invention document information An explanation of the frequency update status is given by adding and deleting.

Referring to FIG. 4, the automatic completion system for each query type for which the search result of the present invention is guaranteed includes a server system 100, a public communication network 110, and a terminal device 120.

The server system 100 receives document information, extracts an index word, constructs an autocomplete database, inputs a query word, converts it into an index word, and retrieves an index word including the index word from the autocomplete database and converts it into a query word.

In detail, the document information is registered and the index word and frequency information are extracted from the registered document information to build an autocomplete database. The query word input from the external user interface is converted into an index word and the index word including the index word from the autocomplete database is automatically generated. The complete list is provided through the user interface, and the selected query word is converted into an index word through the user interface, and the document information including the converted index word is provided to the user interface.

The public communication network 110 connects the server system 100 and the terminal device 300 to a communication path selected from a wired communication path and a wireless communication path, and transmits or receives a data signal.

The terminal device 300 inputs a query to be searched through a user interface (USER INTERFACE: UI) and provides the server system 100 to the server system 100 through the public communication network 110, which is provided by the server system 100. The information is displayed in a user interface (UI), and one of the displayed query words is selected and input to the user interface with a search event signal, and the computer terminal device checks the provided document information.

Referring to FIG. 5, the server system 100 includes a document collecting unit 110, a document index server 120, an autocomplete database 140, an autocomplete server 130, and an index database 150. Configuration.

The document collecting unit 110 registers the collected document information in the document index server 120, and the collected document information includes web page document information, form document information, image document information, video document information, text document information, Content (CONTENTS) document information including multimedia document information.

The document index server 120 registers document information, extracts index terms, and builds a database of autocomplete list information. The document register unit 121, the document index unit 123, the database generator 124, and the document editor The document registration unit 121 records and registers a document of new information provided by the document collecting unit 110 in a separate document information storage database not shown in the drawing.

The document indexing unit 123 extracts the additional information from the registered document information. There are a method of extracting an index word from the input document information and a method of extracting information designated by text processing. The additional information extraction operation is performed on at least one selected. In addition, the extracted additional information is recorded and stored in the index database in a state linked to document information.

Among the additional information extraction methods in which the document index unit 123 includes index word information from the document information registered by the document registration unit 121, the index word extraction method corresponding to the former is morphological analysis from the registered document information. In the selected index method of N-gram, the additional information including the index word is extracted and stored in the index database.

That is, the index word extraction method includes a morpheme analysis method and an N-gram index method, and the index word is extracted from the registered document information in a selected manner.

The morpheme refers to a word or part of a word that has meaning and no meaning in smaller units.

The morpheme analysis is also called the longest matching method, and is a method of adopting an analysis method including the longest number of characters when there is a plurality of possibilities as a method of classifying words or phrases in the morpheme analysis.

The N-gram indexing method refers to N adjacent syllables. For example, 'Korea Science and Technology' is a syllable of 'Korea', 'Kukwa', 'Science', 'Semester' and 'Technology', and each syllable is used as a query word. Among such syllables, a query word of meaningless gram (N-gram) can search for inappropriate document information, and thus weight each syllable to prevent such a state.

In addition, the document indexer 123 removes unnecessary index words included in the stopword dictionary from the index words extracted from the registered document information.

The stopword (STOPWORD, 不 用語) is a word that is not used as a search term when searching the Internet, for example, a word having no meaning as a search index word such as an article preposition investigation conjunction.

The database generation unit 124 retrieves the index words for each type provided by the autocomplete from the index words stored in the index database 150, records them in the autocomplete database 140, and calculates and manages document frequency information again. Each frequency information generated from each document information is recorded, and if the frequency value is 0, the frequency is excluded from the target of the autocomplete list.

It is the frequency of appearance (hereinafter, referred to as frequency) that indicates whether or not the specified index word is included in one document information. When a query word appears in the one document information, a value of 1 is given, and when it does not appear, 0 Is given by.

The autocomplete database 140 records the autocomplete list information of the index word generated by the document index server 120 together with the frequency information.

The autocomplete server 130 searches for the autocomplete database 140, extracts index word information including an index word, and converts the index word information into a query word, and provides a query word input unit 131; A database search unit 132; Index word determination unit 133; The presentation unit 134; Selector 135; The query interlocking unit 136 includes a service interworking unit 136, and the query input unit 131 receives a search query into a user interface and converts the search query into an index. The query unit is a unit selected from phoneme, syllable, word, and word. Whenever the text is entered through the user interface (UI), the autocomplete database is called by AJAX to search.

The Asynchronous JavaScript and XML (AJAX) application (APPLICATION) method is a method of processing data in a client after receiving and providing only necessary data to a web server.

Generally, a web page is created and provided by a searched or requested content on a web server, and a new web page is created and provided when a new content is requested.

In this case, the contents of the original web page and the contents of the new web page are often similar. In other words, the same HTML code is transmitted once again in a state where the HTML code is duplicated, which wastes a lot of bandwidth, and the waste of bandwidth causes time and financial loss, and makes it difficult to service a real-time conversation with a user.

This AJAX method processes some of the data processed by the web server in the client or the connected terminal, thereby reducing the amount of data exchanged between the web server and the client, reducing bandwidth, and reducing overall data throughput on the web server. This enhancement allows for interactive data exchange.

The AJAX method has disadvantages such as a browser that is not applicable, limited functionality of an HTTP client, security problems, and difficulty in debugging by writing a script. Delegation of data processing to client or terminal reduces server load, data processing time is short, asynchronous data communication is possible, bandwidth is reduced by small data volume, communication time is reduced, etc. By the advantage of the method is used a lot.

The database search unit 132 searches for the index words inputted and converted by the query input unit from the autocomplete database, and searches the index words in a forward match and a backward match, respectively, to create a list.

The index word determining unit 133 checks the frequency information of the index words stored in the autocomplete database 140, determines that the query frequency information is one or more values, and provides it as the autocomplete list information.

The presentation unit 134 converts the autocomplete list information provided by the index word determination unit 133 into query information and provides the user interface with one or more pieces of information selected from query input statistics information, frequency information, and order information. Use to adjust the order or order of display in the list.

The selector 135 provides an input query and an autocomplete list that the presentation unit 134 converts into a query language and provides the user interface, converts the selected query information into an index word, and converts the selected query information into an index word. ) Searches for index word information input and provided by the selection unit 135 by a search event signal, and provides document information searched from the index database 150. The search is searched by API call.

The application program interface (API) is a specific method predetermined by a computer operating system or another application program that enables a processing request to an operating system or another application program.

An API is different from a graphical user interface or an imperative interface that directly interacts with the user as an interface to the operating system or program.

The API refers to a language or message format used when an application program communicates with a system program such as an operating system or a database management system. An API is implemented by calling a function that provides a link to a specific subroutine for execution within a program.

That is, an API consists of several program modules or routines that must already exist or be linked in order to perform the work requested by the function call.

Generally, a server includes components that utilize a computer, communicate with a network, perform computer arithmetic operations, and perform various functions. Each of these components includes a processor and a memory of the server. (memory), the input and output means and the like.

The server system 100 according to the present invention having the above-described configuration registers and indexes document information and converts the query index inputted by the user into a index and converts the query index entered by the user into an index. Includes a document collection unit 110 to collect the document information and the automatic completion providing server 130 to provide a list of the index word is searched from the autocomplete DB 140 and converted into a query language, and includes information for each type of index word and index word And an index database 150 for recording the index word information.

The document index server 120 includes a document registration unit 121 for registering document information, a document indexing unit 123 for extracting and indexing index words in the document information, a DB generating unit 124, and a document editing unit 122. .

The document registration unit 121 receives document information through the document collection unit 110 including a document register, a knowledge management system, a document collector, and the like. The registered document information includes all contents such as web page document information, text document information, format document information, image document information, and video document information.

The document indexing unit 123 extracts index words (query words) detected by an index method selected from an index method such as morpheme analysis and N-gram, and stores the index words (query words) in the index database 150.

The document indexing unit 123 extracts an index word to perform additional information such as extracting a specific index word from registered document information or extracting specific information through text processing.

At this time, the additional information extracted from the document registration unit 121 is added to the index database 150, and unnecessary index words are removed in advance by using a stopword dictionary or the like.

The database generator 124 extracts the index information for each type provided by the autocomplete from the index word information stored in the index database 150, records the index word information with the index word information in the autocomplete database 140, and generates the document frequency. ) Recalculate the information.

The frequency information records frequency information generated in the document, and when the value is 0, the index information is excluded from the object to be presented as an autocomplete list.

The document editing unit 122 corrects or deletes already registered document information. As document information is modified or deleted, index words and frequency values in the document information need to be changed, thus affecting the database generator 124.

The present invention searches and extracts index words from document information input and recorded in the system, and separately displays and manages the searched index words on index words recorded in the index word dictionary managed by the system.

The query and index terms are slightly different, but are generally used in the same meaning. That is, the query is an interface with the user and is displayed to be input or selected by the user. The query is input to the system and converted into an index, and the index is converted into a query and presented or output.

In addition, when specific article information is input to the system, for example, the top five index words are detected from the article information and extracted as representative index words.

The searched, extracted and assigned top five index words place a frequency field in the index dictionary of the system and increase the frequency selected as the index word.

In the present invention as described above, when new document information is added, index word extraction is performed, and the frequency information of index words is automatically accumulated and changed in the index word dictionary, so that the autocomplete list is updated in real time. There is an advantage to this.

For example, when the specific document information is deleted, the frequency information of the corresponding index word in the frequency field is reduced by one each. In this way, a response according to the case where specific document information is added or deleted can be processed in real time, which will be described in detail later with reference to FIG. 7.

The index database in the search engine included in the system includes an index dictionary and a person dictionary. The person dictionary is composed of a person name received through a web service directly from a URI server. In other words, rather than having a list of index words obtained from corpus, etc., authors (persons) of service target document information such as article information are added in real time.

However, like the index dictionary, the frequency information values are maintained in real time, and thus the list is provided by auto-completion and the method of adding and deleting document information is the same.

7 is an explanatory diagram of a frequency update state by addition and deletion of document information.

Referring to FIG. 7, in the case where document information 1 is registered, five index words are extracted and additionally recorded in the index database 150 and the autocomplete database 140. If the index words are type-specific information based on an autocomplete method, the information is extracted from one document information, and thus the frequency is 1.

When document number 2 in FIG. 7 is registered, five index words are extracted. Since “OWL” and “Semantic Annotation” are index words already extracted from document number 1, the frequency is 2, and the remaining 3 The dog index words are the first ones extracted, so the frequency information is one.

When the first document information in FIG. 7 is deleted, the frequency corresponding to five index words based on the first document information is decreased by one.

In the prior art, a method of automatically presenting only index terms that generate a search result has been introduced. Here, when a user query succeeds, a flag for checking whether the index is successful or failed is searched. A simple method of attachment was used.

In the prior art as described above, even if document information including an index word that caused a search failure is added later, there is a problem in that the automatic completion list is not presented until the user attempts to search by directly inputting the query word.

By overcoming the above problems, by automatically adjusting the frequency information of the query targets automatically completed at the time of document information registration or editing to provide an automatic completion list that guarantees the search results in real-time without time difference This is the technical idea of the present invention.

The prior art OntoFrame provides an object-oriented integrated search, which is a subset of the query.

Since the inputted query is matched with a specific query type such as a person or a subject, the search query page is configured, and if not, the general search result page is configured.

The type checking of the query is performed by calling a search engine, converting the input query into an index, and simultaneously referring to the index and person dictionary in the search engine for the index. At this time, while searching for the index word and the person name having a frequency value of 1 or more, type information such as whether the input query word is a local type or a person type is searched together.

The following is a search engine program according to an example of the present invention.

API: SearchResultList getAutoComplete (String SearchTerm)

Example call: getAutoComplete (“sem”)

Example result:

[Sem Borst, Person]

[Semantic Annotation, Topic]

[Semantic Web, Topic]

Semih Ergintav, Person

[Semyon M. Meerkov, Person]

In the auto-completion interface according to the present invention, an index word type based on a search result value may be recognized and displayed in an icon, color, tree classification, or the like manner.

The autocomplete server 130 is a query input unit 131 that receives a user query and converts it into an index, a database search unit 132 for searching the corresponding index in the autocomplete database 140 including the index, the autocomplete database An index word determination unit 133 for checking the frequency information stored in the 140 and determining whether to provide autocompletion, a presentation unit 134 for converting an autocompletion list into a query and providing the search interface through a user interface (UI), A selection unit 135 for providing a user interface (UI) to select a specific query among the suggested autocompletion lists including the suggested queries and converting the selected query into an index and an event by a search button or a keyboard operation Service interlocking unit 136 for providing the document information retrieved by the search service according to the signal It includes.

The query input unit 131 receives a query and converts it into an index through a search window provided in a user interface method. The input index word calls the database search unit 132 in an Asynchronous JavaScript and XML (AJAX) method whenever one letter is input by phoneme, syllable, word, and word unit.

The database search unit 132 searches the autocomplete database 140 for the input search word and searches whether the search words including the search word exist.

For example, the search may be performed in a forward-matching manner in which the front part coincides with the search as “Daepyeong Middle School” is searched for “Daepyeong” and in the back-matching method where the rear part is matched as if “simpyeongpyeong” is searched for “daepyeong”.

The index word determination unit 133 determines to present, as an autocomplete list, among the search words that are identified or matched as being included in the query through the database search unit 132 and searched for.

Meaning that the frequency value is 1 or more means that the document information including the corresponding index word (query word) exists in the search system.

The presenting unit 134 converts the index words obtained through the index word determining unit 133 into query terms and presents them as an autocomplete list. According to the general autocompletion method, the ranking or order of the placed and displayed in the autocompletion list is adjusted by using one or more selected from the user's input query information, frequency information, and order information. The ranking of the autocomplete list is adjusted by using the frequency information value or the alphabetical order information in the autocomplete database 140.

The manner of adjusting or determining the ranking may be applied when a new way is developed.

When the selector 135 selects and selects a specific query from an autocomplete list presented as a query to the terminal device 120 through a user interface (UI), the selector 135 converts the selected query into an index and inputs the index. In the selected query list, a specific query is designated using an up / down button or a mouse of a keyboard provided in the terminal device 120 and one query is selected.

The selected query (index) information is transmitted to the service interlocking unit 136 together with the corresponding event signal information.

The service interworking unit 136 inputs the selected query word and retrieves index information from the index database 150 through an API call according to an event signal generated by a keyboard operation such as a search button or an enter key and provides corresponding document information. Handles the service.

FIG. 6 is a flowchart illustrating an autocomplete method for each type of query in which a search result is guaranteed by an example of the present invention.

Referring to FIG. 6, a method of autocomplete for each type of query word for which a search result is guaranteed by an example of the present invention is modified. Decision making process; Output process; It includes a configuration.

The modifying process includes collecting and registering document information, extracting an index word, storing the index word in an index database, generating an index word for automatic completion, storing the index information in an autocomplete database, and modifying or deleting the registered document information. Collecting document information (S100), recording and storing the collected document information in a separate document information database (not shown in the drawing) by the document registration unit, and storing the document information by the document index unit from the registered document information. Extracting additional information including an index word and storing the index information in an index database (S110), and generating an index word to be automatically provided by a database generator from the information stored in the index database and storing the index word in an autocomplete database (S120). ), The document editing unit for the registered document information It includes a process of modifying or deleting by (S130).

The index word extraction method includes a morpheme analysis index method and an N-gram index method, and any one method selected from the morpheme analysis index method and the N-gram index method is used.

If the index word extraction method is not used, a text processing method may be used.

The determining process is to search for an index word input from the autocomplete database 140 to the user interface and determine a frequency value of 1 or more as a search index word, wherein the input index word is any one selected from phoneme, syllable, word, word. Whenever it is input in one unit, the database search unit 132 is called in an AJAX manner (S140 to S160).

The outputting step is to convert the determined index word into a query word and present it as an autocomplete list, search for information on the selected index word, and output corresponding document information (S170 to S190). The autocomplete list of the query word is input by a user. The rank (order) is controlled by one or more selected from the query input statistics information, the frequency information of the autocomplete database, and the order information of the alphabet.

The present invention having the above configuration is all performed by the above description of the system.

The method according to the present invention can be embodied as computer readable codes on a computer readable recording medium. The computer-readable recording medium is a recording device that stores data that can be read by a computer system. For example, ROM, RAM, Cache, hard disk, optical disk, floppy disk, magnetic tape and the like. In addition, the carrier wave may be implemented in the form of a carrier wave, for example, transmission through the Internet. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

Although the present invention has been described in detail only with respect to the described embodiments, it will be apparent to those skilled in the art that various changes and modifications can be made within the technical scope of the present invention, and such modifications and modifications belong to the appended claims.

1 is a functional configuration diagram of a system for retrieving information from a general database system;

FIG. 2 is a state diagram showing, by way of example, an autocompletion method of a query entered and retrieved to retrieve data;

3 is a state in which a query is input and a search fails due to an example of the related art.

4 is a functional configuration diagram of an automatic completion system for each type of query word with a guaranteed search result according to an example of the present invention;

5 is a detailed configuration diagram of a server system of an autocomplete system for each query type with a guaranteed search result according to an example of the present invention;

FIG. 6 is a flowchart illustrating an autocomplete method for each type of query, in which a search result is guaranteed by an example of the present invention;

Fig. 7 is an explanatory diagram of a frequency update state by addition and deletion of document information by one example of the present invention.

          ** Explanation of symbols on the main parts of the drawing **

100: server system 200: public communication network

300: terminal device 110: document collection unit

120: document index server 121: document register

123: document indexing unit 124: database generation unit

122: document editing unit 140: autocomplete database

130: autocomplete server 131: query input unit

132: database search unit 133: index word determination unit

134: presentation unit 135: selection unit

136: service interlocking unit 150: index database

Claims (31)

A document index server for registering document information, extracting and recording index word information and frequency information from the registered document information, and generating autocomplete list information from the extracted index word information; An autocomplete database for recording the autocomplete list information generated by the document index server in association with the frequency information; And Search the autocomplete database to extract the autocomplete list information including the index information, convert it into a query language, provide the user interface, convert the selected input query word into an index word, and search for document information including the index word. Autocomplete server to provide a user interface; Guaranteed search results, including the automatic completion system for each type of query. The method of claim 1, A document collecting unit for registering the collected document information with the document index server; And An index database for recording the index word information provided from the document index server and providing the index word information to the autocomplete server; Autocomplete system for each type of query is guaranteed search results, characterized in that further comprises. The method of claim 2, wherein the document collecting unit, Automatic search results guaranteed by query type characterized in that the configuration consists of collecting one or more content document information, including web page document information, form document information, image document information, video document information, text document information, multimedia document information Finishing system. The server of claim 2, wherein the document color is a server. A document registration unit for registering the collected document information of the document collection unit; A document indexing unit which extracts the index word from the document information of the document registering unit and stores the index word in the index database; And A database generation unit for retrieving index word information provided as an autocomplete list from the index word stored in the index database, recording the index word information in the autocomplete database, and updating and managing frequency information; Autocomplete system for each type of query is guaranteed search results, characterized in that the configuration consisting of. The method of claim 4, wherein the document index unit, And extracting the additional information including the index word information from any one or more selected from extracting the index word from the document information registered in the document registration section and extracting the index word information designated by text processing. Autocomplete system for each query type with guaranteed search results. The method of claim 4, wherein the document index portion, The search result is configured to extract the index word from any of the document information registered in the document registration unit and a method selected from the N-gram method and store the index word in the index database. Autocomplete system by guaranteed query type. The method of claim 6, wherein the document index portion, And the additional information including the extracted index word is recorded and stored in the index database in association with the document information. The server of claim 4, wherein the document color is a server. A document editing unit for modifying and deleting the document information registered in the document registration unit; Autocomplete system for each type of query is guaranteed search results, characterized in that further comprises a configuration. The method of claim 4, wherein the document index portion, The automatic completion system for each type of query word guaranteed by the search result, characterized by removing the unnecessary index word included in the stopword dictionary from the index words extracted from the document information registered in the document registration unit. The method of claim 4, wherein the database generation unit, And the frequency information of the document information is cumulatively calculated and recorded as an object of the autocompletion list in the unit of the autocompletion database. The method of claim 10, wherein the database generation unit, When the value of the frequency information is 0, the automatic completion system for each type of query result guaranteed by the search result, characterized in that the configuration is excluded from the cumulative autocomplete list target. The method of claim 1, wherein the autocomplete server, A query input unit which receives the query word to be searched by a user interface and converts the query word into the index word; A database search unit for searching the index word of the query word input unit from the autocomplete database; An index word determination unit for checking frequency information of index words stored in the autocomplete database and determining and providing the information as autocomplete list information; A presentation unit which converts the autocomplete list information provided by the index word determination unit into the query language and provides the converted information to the user interface; A selection unit which provides the input query word and the query word of the auto-completion list provided by the presentation unit to the user interface, and inputs the query word information selected together with an event signal to convert the index word into the index word; And A service linker for searching and providing the document information according to the index word information and a search event signal inputted by the selector; Autocomplete system for each type of query is guaranteed search results, characterized in that consisting of a configuration. The method of claim 12, wherein the query input unit, The query completion type automatic completion system for each query type, characterized in that the query is composed of input unit by any one selected from phoneme, syllable, word, word. The method of claim 13, wherein the query input unit, The automatic completion system for each type of query result guaranteed by the search result, characterized in that it comprises a configuration for calling the auto-complete database in the AJAX method every time the query is input. The method of claim 12, wherein the query input unit, And a query result guaranteed automatic search system for each type of query, comprising: inputting the query information into the user interface (UI). The method of claim 12, wherein the database search unit, And a search result guaranteed query type according to a query type, characterized in that the index word is searched in the forward matching and backward matching methods, respectively. The method of claim 12, wherein the index word determination unit, The automatic completion system for each type of query word guaranteed by the search result, characterized in that the configuration to determine to provide the frequency information of the index word to the one or more in the autocomplete list. The method of claim 12, wherein the presentation unit, And the ranking is adjusted in the autocomplete list by using one or more selected from the input statistics information of the query, the frequency information of the query, and the order of information of the query. Autocomplete system for each type of query with guaranteed results. The method of claim 12, wherein the service interlocking unit, The search result guaranteed automatic completion system for each type of query, characterized in that the index information is configured to retrieve the document information by an API call (API). (a) collecting and registering document information, extracting index word information and frequency information from the registered document information, storing the index word information and frequency information in the index database, generating the autocomplete list information of the index word information, storing it in the automatic completion database, and registering the registered document information. Modifying; (b) converting a query word input from the autocomplete database through a user interface into the index word and including the index word searched for the frequency 1 or more in the autocomplete list information; And (c) converting the index word of the autocomplete list into the query language and presenting the query word to the user interface, converting the query word input from the outside into the user interface into the index word, and outputting the document information retrieved from the converted index word. Doing; Autocompletion method for each type of query with guaranteed search results, including. The method of claim 20, wherein step (a) is Collecting the document information by a document collecting unit; Registering the collected document information by a document registration unit; Extracting the index word from the registered document information by a document indexing unit and storing the index word in the index database; Extracting the index word to be provided to the autocomplete list by a database generator from the index word information stored in the index database and storing the index word in the autocomplete database; And Modifying or deleting the registered document information by a document editing unit; Autocomplete method for each type of query is guaranteed search results, characterized in that comprises a. The method of claim 21, wherein the index word extraction, An automatic completion method for each type of query word guaranteed by a search result, wherein the index word is extracted using one of a morpheme analysis method and an N-gram index method. The method of claim 21, Step (a) is The index word stored in the index database includes additional information, and removes the unnecessary index word by using a stopword dictionary and stores the updated frequency information of each document information in the autocomplete database. Guaranteed autocomplete by query type. The method of claim 20, wherein the query word input in the step (b), Autocomplete method for each type of query guaranteed by the search results, characterized in that input by phoneme, syllable, word, word in any one unit selected. The method of claim 24, wherein the input query is, An automatic completion method for each type of query word with a guaranteed search result, which is converted into an index word and retrieved by calling from the autocomplete database unit in an Ajax manner. A document that collects and registers document information, extracts index word information and frequency information from the registered document information, stores the index word information and frequency information in the index database, generates the autocomplete list information of the index word information, stores the information in the autocomplete database, and modifies the registered document information. Information collection and registration process; An index word extraction and storage process for converting a query word input from the autocomplete database through a user interface into the index word and including the index word searched for the frequency 1 or more in the autocomplete list information; And A search process for converting the index word of the autocomplete list into the query word and presenting it to the user interface, converting the query word selected from the outside into the user interface into the index word, and outputting the document information searched for in the converted index word; ; To Recording medium that records the program source of automatic completion method for each query type with guaranteed search results. 27. The method of claim 26, wherein the index extracting and storing process is a process of adjusting the rank to at least one selected from input statistical information of the query, query frequency information in the autocomplete database, and the Canadian order information of the query. A recording medium recording a program source of an autocompletion method for each query type with guaranteed search results. The search result according to claim 26, wherein the search process is a process of calling and searching the autocomplete database in an ajax manner whenever it is input in any one unit selected from phoneme, syllable, word, and word. Recording the program source of the autocomplete method for each guaranteed query type. Register the document information, extract the index word and the frequency information from the registered document information to build an autocomplete database, convert the query word input to the user interface from the outside into the index word and the index word including the index word from the autocomplete database. A server system for providing an autocomplete list through the user interface, converting a query word selected and input through the user interface into the index word, and providing the document information including the converted index word to the user interface; A public communication network connected to the server system and transmitting and receiving data of the query word and the retrieved document information through a communication path selected from a wired communication path and a wireless communication path; And Input the query to search and access the public communication network to the user interface and transmit the query to the server system to display an autocomplete list of the query provided by the server system on the user interface and display the selected query as an event signal. A terminal device configured to be inputted together with the server system to display the document information provided by the server system; Autocomplete system for each query type is guaranteed a search result including a. The server system of claim 29, wherein the server system comprises: A document index server that registers the document information, extracts the index word and the frequency information, and builds the autocomplete database; Input the query word from the outside into the user interface, convert the index word, extract index word list information including the index word, convert the query word into the query word, provide the query word to the user interface, and convert the selected input query word into the index word An autocomplete server providing the document information; The autocomplete database for recording the autocomplete list information generated by the document index server in association with the frequency information; A document collecting unit which registers the collected document information by accessing the document index server; And An index database for recording the index word information provided from the document index server and providing the index word information by searching the autocomplete server; Autocomplete system for each type of query is guaranteed search results, characterized in that the configuration consisting of. The method of claim 29, wherein the public communication network, A wireless communication network connecting the server system and the terminal device to the wireless communication path and transmitting the data signal; And A wired communication network connecting the server system and the terminal device to the wired communication path and transmitting the data signal; Guaranteed search results, characterized in that the configuration consisting of including the automatic completion system for each type of query.

Images