KR20020003915A - A method for providing search result including recommendation of search condition, and a server thereof - Google Patents

Abstract

PURPOSE: A method for providing a searching result for recommending searching conditions and a server for the same are provided to increase the convenience of a user by enabling the user to apply and search keywords or categories recommended, thereby enabling the user to reduce times for searching necessary information. CONSTITUTION: A user who wants to search Internet information connects to a searching server through a user interface(S510). The searching server transmits a connection window to the user interface(S515). The user receives the connection window and inputs searching conditions(S520). The conditions are transmitted to the searching server. In addition, the searching server judges whether the conditions are included in keyword input information(S525). In case that the conditions are included in the information, keywords to be recommended are extracted(S530). In case that the conditions aren't included in the information, it is judged whether the conditions are included in category selection information(S540). The searching server searches a database unit based on the conditions(S545). The searching server makes out a result document including searching results and so on(S550). The searching server transmits the document to the user interface(S555). The document is displayed to the user interface(S560). The user judges whether to newly perform a searching(S565).

Description

A METHOD FOR PROVIDING SEARCH RESULT INCLUDING RECOMMENDATION OF SEARCH CONDITION, AND A SERVER THEREOF}

The present invention relates to a search result providing method and server for recommending a search condition, and more particularly,

A network interface connected to a user interface via a network, the network interface transmitting and receiving data through the network; A keyword database containing keywords used for searching by the user; A user behavior information database including a user behavior code indicating a user behavior and attached data indicating information used for the user behavior; Keyword recommendation module; And a search server including a category recommendation module,

When the user searches for a keyword, the keyword importance is calculated for the keyword stored in the search server, and the keywords are recommended in the keyword importance order.

When the user selects a category provided by the search server, a category importance level is calculated for a category stored in the search server, and a search result providing method for recommending categories in the category importance order and a search server providing the method. It is about.

As will be appreciated, computers can exchange information between multiple computers via a network. However, as the Internet is widely known, sites providing information through networks such as the Internet have been created, and network addresses or URLs (Uniform Resource Locator) of each site can be easily found among countless sites that provide necessary information. And a search server that provides a database of the summary of the information provided by the site, categorized into its own classification scheme and provided.

However, in order to easily find the desired information through the search server, it is important to use an appropriate search term, and it is inconvenient to do a lot of trial and error in order to find a category containing a URL containing desired information in a classification system classified into categories. there was.

Recently, in order to reduce such inconvenience, a search server for recommending keywords or recommending categories has been introduced. However, in the method of recommending a keyword or category by the search server, a recommendation rank is determined only by the frequency of using the keyword or the category by other users, and the keyword or category is recommended by the ranking.

However, it is difficult to expect the effect of recommending an appropriate keyword or category because the recommendation based on the simple frequency of use is not properly reflected in the user's intention and behavior.

SUMMARY OF THE INVENTION The present invention has been made to solve such a problem, and an object of the present invention is to provide a search result providing method and a search method for recommending an efficient search condition by using a database that records the actions of other users with respect to a keyword input by a user or a category to be searched. Implement a server.

1 is a block diagram of a network search system to which a network search server according to an embodiment of the present invention is applied.

2 is a diagram illustrating an example of a user behavior information database used in a search method and a search server according to an embodiment of the present invention.

3 is a diagram showing an example of how the user action code and the attached data are created.

4 is a diagram illustrating an example of a keyword database used in a search method and a search server according to an embodiment of the present invention.

5 is a flowchart illustrating a search result providing method for recommending a keyword or a category according to an exemplary embodiment of the present invention.

6 is a flowchart illustrating detailed steps of a keyword recommendation module performing step in the method for providing a search result recommending a search condition according to an exemplary embodiment of the present invention.

7 is a flowchart illustrating a detailed step of performing a category recommendation module in the search result providing method recommending a search condition according to an embodiment of the present invention.

In order to achieve the above object, the search server according to the present invention is connected to a user interface via a network.

A network interface for transmitting and receiving data through the network; A keyword database containing keywords used for searching by the user; A user behavior information database including a user behavior code indicating a user behavior and attached data indicating information used for the user behavior; Keyword recommendation module; And category recommendation module,

Search result providing method according to the present invention,

A keyword recommendation step of calculating keyword importance with respect to a keyword including a search keyword input from a user among the keywords included in the keyword interface, and arranging the keywords in order of the keyword importance; And

After extracting a user behavior information database item including a category ID item related to the search category in the attachment data, calculating the number of occurrences of the extracted user behavior information database item for each action code, and category importance based on the occurrence number. A category recommendation step of arranging the category IDs in order of category importance; Including,

The search result document including the keyword or category obtained in the keyword recommendation step or the category recommendation step is sent to the user interface.

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

1 is a block diagram of a network search system to which a network search server according to an embodiment of the present invention is applied.

As shown in FIG. 1, the network search system according to an embodiment of the present invention includes a network search server 100 and a user interface 190 connected to the network search server 100 through a network such as the Internet.

The network may include any network connected between a plurality of computers, such as the Internet, and may include an Internet providing server connected to the search server 100 and the Internet, and a wireless communication network capable of transmitting and receiving data with the Internet providing server. have.

The user interface 190 includes any terminal capable of transmitting / receiving data via the network or wireless communication network, but is preferably a user computer.

The search server 100 manages a data interface 105 for transmitting and receiving data through the network, a database unit 140 for providing a search function to a user, and data stored in the database unit 140. It includes a CGI unit 110 (Common Gateway Interface) for providing a search function to the user by using the data stored in the database unit.

The database unit 140 includes an index database 145 for storing index information of databases included in the database unit 140, a URL database 150 for storing Uniform Resource Locators (URLs) of the sites, and the URLs. A category database 155 storing a classification system for sorting by type, a document summary database 160 for summarizing and storing information of documents provided by the URL, and a record of a user's use of the search server 100. A user behavior information database 165, and a keyword database 170 that stores a keyword input and searched by a user using the search server 100 and history information about the keyword.

The index database 145, the URL database 150, and the document summary database 155 include information about a URL for providing content, and are provided by a user using the search server 100 according to the present invention, or The search server 100 includes a database that collects the URL information on a network such as the Internet or accumulates in any other way.

The category database 155 storing a classification scheme for classifying the URLs includes an arbitrary category scheme, whether or not created by the search server operator, and each category included in the category scheme includes a category ID. Is given and recorded.

2 is a diagram illustrating an example of a user behavior information database 165 used in a search method and a search server according to an embodiment of the present invention.

As shown in FIG. 2, the user behavior information database 165 includes a user identification code (hereinafter referred to as a user ID) for identifying a user, an action code for displaying the user's behavior, and information used for the user's behavior. Includes attached data referencing.

The user action code is a coded code of various actions that can be performed by the user in using the search server, and the attached data refers to data used in conjunction with the user action corresponding to the code.

3 is a diagram showing an example of how the user action code and the attached data are created.

As an example of the user action code and the attached data, as shown in FIG. 3, the action code is classified into A to E, and the action code A is the action that the user searches by keyword, and B is the user classified into the category. Click on a category in the taxonomy, C click on a category displayed in a keyword search result, D click on a URL contained in a document provided as a search result (hereinafter referred to as a search result document), and E And storing the result document in the search server, respectively, wherein the user action corresponding to the action code A is a keyword, B to C and E are corresponding category IDs, and D is a corresponding document ID. Include as data.

The user behavior information database 165 is updated whenever there is a user's use behavior such as searching for keywords while using the search server or clicking a hyperlink included in a document provided by the search server.

4 is a diagram illustrating an example of a keyword database 170 used in a search method and a search server according to an embodiment of the present invention.

As shown in FIG. 4, the keyword database 170 includes a keyword used for a user's search, a number of times that the keyword is selected and searched, and a hyper-link included in the search result document searched with the keyword. The number of clicks, and the number of times the search result document searched by the keyword is stored in the search server by the user.

The CGI unit 110 manages data stored in the database unit 140 regardless of a language such as Java, Perl, or C, and uses the data stored in the database unit. It can be written as a program that provides a search function to the user.

The CGI unit 110 includes a search module 115 for searching according to a user's search condition, a keyword recommendation module 120 for extracting and recommending other keywords that can be searched for the keyword input by the user, and a user's selection. It includes a category recommendation module 125 for extracting and recommending a category that can be searched from the detailed categories included in the category.

The search module 115 executes a search by the action by inputting a keyword or clicking on an arbitrary category included in the categorized category system provided by the search server.

When the user action is a keyword input, a URL, a document summary, and a category including the keyword are searched from the URL database 150, the document summary database 160, and the category database.

If the user action is a category selection, the database unit 140 retrieves information about detailed categories and URLs included in the category.

When the search module 115 is performed, it is preferable to perform the search together with the keyword recommendation module 120 or the category recommendation module 125 according to the user action.

The keyword recommendation module 120 extracts a keyword to be recommended by a predetermined step based on the keyword input by the user and the data recorded in the keyword database 170.

The category recommendation module 125 extracts a category to be recommended from the detailed directory of the selected category of the user by a preset step based on the category selected by the user and the data recorded in the user behavior information database 165.

5 is a flowchart illustrating a search result providing method for recommending a keyword or a category according to an exemplary embodiment of the present invention.

As shown in FIG. 5, a user who first searches for Internet information accesses the search server 100 according to the present invention connected to the user interface 190 via a network through the user interface 190 (S510). .

The search server 100 receiving the connection of the user sends a connection window to the user interface 190 (S515). The access window includes input means for allowing the user to enter a keyword, and includes all or part of the classified category list.

In addition, when transmitting the access window, the search server preferably performs a category recommendation module, extracts a category to be recommended from the entire category range, and includes the same in the access window.

The user who receives the connection window inputs a search condition (S520). The search condition input includes inputting a keyword through the input means or selecting a category included in a category list included in the access window.

The search condition is transmitted to the search server 100, and the search server 100 determines whether the received search condition is keyword input information (S525).

If it is determined in the keyword input determination step (S525) that the search condition is keyword input information, a keyword to be recommended is extracted by performing a keyword recommendation module (S530), and after the keyword to be recommended is extracted, a search step (S545). Proceed to).

If it is determined in the keyword input determination step (S525) that the search condition is not keyword input information, it is determined whether the search condition is category selection information (S535).

If it is determined in the category selection determination step (S535) that the search condition is category selection information, a category to be recommended is extracted by performing a category recommendation module (S540), and after the category to be recommended is extracted, a search step (S545). Proceed to).

If it is determined in the category selection determination step S535 that the search condition is not category selection information, the flow advances to a search step S545.

In a search step S545, the search server searches the database unit 140 based on the search condition. When the search condition is keyword input information, the URL, document summary, and category including the keyword are searched from the URL database 150, the document summary database 160, and the category database.

If the search condition is category selection information, the database unit 140 searches for information on a detailed category and a URL included in the category.

The search server 100 searching the database unit 140 creates a result document including the search result and the execution result of the performed keyword recommendation module or category recommendation module (S550).

When creating the result document, the result document includes storage means for transmitting a storage signal to store the result document in the search server. The storage means may be any storage means capable of transmitting a signal requesting storage to the search server 100, a character string or an icon hyperlinked to the CGI unit 110 of the search server 100, etc. It can be a document element of.

In addition, when creating the result document, it is preferable to include an input window in which a user can input a new keyword.

The search server 100 creating the result document transmits the created result document to the user interface 190 (S555).

The result document is displayed on the user interface 190, and the user checks the result document through the user interface 190 (S560).

The user who checks the result document determines whether to perform a new search (S565).

If it is determined in step S565 that the new search is not to be performed, the search is terminated, and if it is determined that a new search is to be performed, the process proceeds to the search condition input step S520.

6 is a flowchart illustrating the detailed steps of the keyword recommendation module performing step S530 in the method for providing a search result recommending the search condition according to an embodiment of the present invention.

As shown in FIG. 6, when execution of the keyword recommendation module 120 starts, the keyword recommendation module 120 reads an item of the keyword database 170 (S610).

The item includes a keyword, the number of times a user searches with the keyword, the number of clicks on a search result document searched with the keyword, and the number of times the search result document searched with the keyword is stored.

The keyword recommendation module 120 that reads the item of the keyword database 170 determines whether a keyword item among the items includes a keyword (hereinafter, referred to as a search keyword) input by the user for searching (S615).

The determination of whether the keyword item includes the search keyword may be determined based on whether the string of the keyword item includes the search keyword string. If the search keyword string matches the consecutive components of the string of the keyword item, it is determined that the keyword item includes the search keyword, and each character constituting the search keyword string is included in the keyword item string. In addition, it may be determined that the keyword item includes a search keyword.

If it is determined in the determination step S615 that the keyword item does not include a search keyword, in order to search for the next keyword in the keyword database 170, the currently read keyword item is the last of the keyword database 170. In step S660, it is determined whether the item is an item.

If it is determined in step S615 that the keyword item includes a search keyword, it is determined whether the number of selections of the keyword item is greater than zero (S620).

If the number of selections is greater than zero in the determination step (S620), a weighting value for the number of selections is calculated (S625). The selection frequency weighting value is calculated by multiplying the selection frequency by a preset selection frequency weighting ratio.

If it is determined that the selection frequency weighting value or the selection frequency is not greater than zero in the determination step (S620), it is determined whether the search result document click count is greater than zero (S630).

If the number of clicks on the search result document is greater than zero in the determination step (S630), a weighting value for the number of clicks on the result document is calculated (S635). The result document click count weighting value calculation is calculated by multiplying the click count by a preset click count weighting ratio.

The weight of the result document click count is calculated, or if it is determined in step S630 that the result document click count is not greater than zero, it is determined whether the search result document storage count is greater than zero (S640).

If the search result document storage count is greater than 0 in the determination step (S640), a weighting value for the result document storage count is calculated (S645). The resultant document storage frequency weighting value calculation is calculated by multiplying the storage frequency by a preset storage frequency weighting ratio.

When the storage frequency weighting value is calculated or when it is determined in the determination step S640 that the storage frequency is not greater than 0, the importance of the keyword item is calculated based on each of the calculated weighting values (S650).

The importance calculation may be calculated by adding weight values calculated in the weighting calculation steps S625, S635, and S645.

After calculating the importance, the importance is corrected based on the number of clicks on the result document and the number of result document storages (S655).

The correction is based on a correlation between the result document click count and the result document storage count, and modifying the selection count weighting ratio, click count weighting ratio, and storage count weighting ratio according to a preset rule. Each weight is calculated to recalculate importance.

As an example of the preset rule, when the result document click count is within a preset range of the result document storage count, the weighting ratio of the click count may be increased to a preset magnification.

As an example of the preset range and the preset ratio, when the result document click count exceeds 80% of the result document save count, the click count weighting ratio is 1.8 times, and the click count is equal to the stored count. Less than 80% and greater than 60%, the click weighting ratio is 1.6 times, and if the number of clicks is less than 60% of the storage frequency and more than 40%, the weighting ratio is 1.4 times, If the number of clicks is 40% or less of the stored number and exceeds 20%, the click count weighting ratio may be 1.2 times.

After correcting the importance (S655), or if it is determined in the determination step (S615) that the keyword item does not include the search keyword, the keyword item currently being processed is the last keyword item of the keyword database 170. It is determined whether or not (S660).

If it is determined in step S660 that it is not the last keyword item, the process proceeds to step S610 of reading an item of the keyword database.

If it is determined in step S660 that the keyword item is the last keyword item, keyword items including the search keyword are arranged in the order of the modified importance level (S665).

7 is a flowchart illustrating a detailed step of performing a category recommendation module (S540) in the method for providing a search result recommending a search condition according to an exemplary embodiment of the present disclosure.

As shown in FIG. 7, when the execution of the category recommendation module 125 is started, the category recommendation module 125 rearranges the items of the user behavior information database 165 in category order (S710).

The rearrangement extracts only an item string in which the attached data item is a category ID from the user behavior information database 165, and rearranges the extracted item strings in the order of the category ID. Therefore, as the rearrangement result, the user behavior codes related to one category ID are arranged, and the user behavior codes related to another category ID are rearranged.

After rearranging the user behavior information database 165, the user behavior code and category ID are read for each category in the rearranged behavior information database (S715).

After reading the user action code and category ID, it is determined whether the category ID is related to the category searched by the user (hereinafter referred to as a search category) (S720).

The determination determines that the category ID is relevant when the category ID corresponds to the ID of the detailed category of the search category, and determines that the category ID is not relevant.

If it is determined in step S720 that it is not relevant, in order to search for an item related to the next category in the rearranged user behavior information database, determining whether the category ID currently being processed is the last category (S740). Proceed.

If it is determined in the determination step (S720) that the relevant, the user action code corresponding to the category ID currently being processed is read, and the number of occurrences for each action code is measured (S725).

After measuring the number of occurrences for each action code, the importance of the category represented by the category ID currently being processed is calculated (S730).

The category importance calculation may be calculated by multiplying the number of occurrences of each action code by a preset weighting ratio of each action code, and adding all the multiplied values for each action code.

After calculating the category importance (S730), or if it is determined in step S720 that the category ID currently being processed is not related to the search category, the category ID currently being processed is stored in the rearranged user behavior information database. It is determined whether it corresponds to the last category (S740).

If it is determined in step (S660) that it does not correspond to the last category, the process proceeds to step (S715) of reading the items of the user behavior information database rearranged by the categories.

If it is determined in the determination step (S660) that it corresponds to the last category, the category ID items are rearranged in the order of the calculated category importance (S745).

As described above, a preferred embodiment related to a search result providing method and a search server recommending a search condition of the present invention has been described. However, the present invention is not limited to the above embodiment, and the present invention is not limited to the above embodiment. Any change in the scope deemed to be equivalent and readily changed by a person with knowledge of the subject matter shall be included.

As described above, according to the embodiment of the present invention, the keyword or category can be efficiently recommended by utilizing a database that records the actions of other users with respect to the keyword input by the user or the category to be searched.

In addition, by allowing the user to search by using the recommended keyword or category, there is an advantage that the user can shorten the time spent searching for necessary information.

Claims (17)

Networked to the user interface, In a search server including a network interface for transmitting and receiving data over the network, and a keyword database including keywords used for searching by the user, (a) receiving a search keyword from a user; (b) reading a keyword from the keyword database; (c) determining whether the read keyword includes a search keyword; (d) calculating keyword importance when it is determined to include in step (c); (e) arranging the keywords in order of keyword importance; (f) transmitting a search result document including the arranged keywords to the user interface; Search results providing method for recommending a search condition comprising a. In claim 1, (B) to (d), the search results providing method for recommending a search condition characterized in that it is performed repeatedly for all the keywords in the keyword database. The method of claim 1 or 2, The keyword database further includes a number of selections for which a keyword is used for searching, a number of clicks for clicking on a result document searched with the keyword, and a number of storing for storing the result document, In step (d), (g) calculating a keyword selection frequency weighting value; (h) calculating a weight of the result document click count; (i) calculating a weight of the result document storage frequency; (j) assigning each weighting value to a keyword importance level; Search results providing method for recommending a search condition comprising a. In claim 3, In the step (g), the keyword selection count weighting value is calculated by multiplying the keyword selection count by a preset selection count weighting ratio. In the step (h), the result document click count weighting value calculation is performed by multiplying the result document click count by a preset click count weight ratio. And calculating the result document storage frequency weighting value in the step (i) by multiplying the result document storage frequency by a preset storage weight weighting ratio. In claim 3, After step (j), (k) modifying the keyword importance according to a preset rule; Search results providing method for recommending a search condition further comprising. In claim 5, In the step (k), The predetermined rule is a search result providing method for recommending a search condition, characterized in that if the number of clicks on the result document is within a predetermined range of the number of storing the result document, the weighting ratio of the number of clicks is increased to a preset magnification. . Networked to the user interface, In a search server including a network interface for transmitting and receiving data through the network, and a user behavior information database including a user behavior code indicating a user's behavior and attached data indicating information used for the user's behavior, (l) receiving a selection of a search category from a user; (m) extracting a user behavior information database item including a category ID item related to the search category in the attachment data; (n) calculating the number of occurrences of the extracted user behavior information database items for each action code; (o) calculating a category importance based on the number of occurrences; (p) arranging the category IDs in order of category importance; (q) sending a search result document including the arranged categories to the user interface; Search results providing method for recommending a search condition comprising a. In claim 7, (M) step, (r) arranging the user behavior information database in the order of category IDs indicated by the attached data items; (s) reading the user behavior information database item for each category ID; (t) determining whether the category ID is related to the search category, and if it is determined to be related, selecting a user behavior information database item including the category ID as attachment data; Search results providing method for recommending a search condition comprising a. The method of claim 7, wherein (M) to (o), the search result providing method for recommending a search condition, characterized in that it is repeatedly performed for all category ID types in the user behavior information database. In claim 9, The determination of whether the search category and the category ID item are related is determined as being related when the category ID corresponds to the ID of the detailed category of the search category, and otherwise determined as not related. How to provide search results that suggest conditions. In claim 9, In the step (o), The category importance calculation method of claim 1, wherein the number of occurrences of each action code is multiplied by a predetermined weighting ratio of each action code, and the multiplication value is added to each action code. Networked to the user interface, A network interface for transmitting and receiving data through the network; A keyword database containing keywords used for searching by the user; And keyword recommendation module, The keyword recommendation module calculates keyword importance with respect to a keyword including a search keyword input from a user among the keywords included in the keyword interface, and arranges the keywords in order of the keyword importance. Recommended search server. In claim 12, The keyword database further includes a number of selections for which a keyword is used for searching, a number of clicks for clicking on a result document searched with the keyword, and a number of storing for storing the result document, The keyword importance calculation may include: a selection count weighting value multiplied by the predetermined number of selection weighting ratios, a click count weighting value multiplied by a predetermined click count weighting ratio, and a preset number of times of storing the result document; A search server recommending a search condition characterized in that it is calculated by adding a storage frequency weighted multiplied by the storage frequency weighting ratio. In claim 13, And the click count weighting ratio is increased at a preset magnification when the number of clicks on the result document is within a preset range of the number of storing the result document. Networked to the user interface, A network interface for transmitting and receiving data through the network; A user behavior information database including a user behavior code indicating a user behavior and attached data indicating information used for the user behavior; And category recommendation module, The category recommendation module extracts a user behavior information database item including a category ID item related to the search category in the attachment data, calculates the number of occurrences of the extracted user behavior information database item for each action code, and generates A search server for recommending a search condition, wherein the category importance is calculated based on the number of times, and the category IDs are arranged in order of the category importance. The method of claim 15, The user behavior information database item extraction, After arranging the user behavior information database in the order of category IDs indicated by the attached data item, the user behavior information database item is read for each category ID, and then the user behavior information database item whose category ID is related to the search category selected by the user. Search server for recommending the search conditions, characterized in that the extraction by screening. The method of claim 16, The determination of whether the category ID is related to the search category may include determining that the category ID is related when the category ID corresponds to an ID of a detailed category of the search category, and otherwise determining that the category ID is not related. Search server to recommend search criteria.