Apparatus and method for providing a service for providing a search word network based on a search path

Abstract

To provide a service providing device and a service providing method for supporting a user so that the user can easily select a relevant search word that can improve efficiency of marketing.SOLUTION: A service providing device and method are to, on the basis of, a search path allowing a user to grasp, as distance, a relationship between a search word requested by the user and other search words related to the search word, generate and provide a search word network that visualizes a connection relationship between the search words and search words with high importance, thereby providing a search word network for a search path base to support the user to easily select a relevant search word with the highest relevance to reach the search word requested by the user and use the relevant search word for marketing.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a service providing apparatus and method for providing a search word network based on a search path, and more particularly, to a service providing apparatus and method for providing a search word network based on a search path, which generates and provides a search word network that visualizes connection relationships between search words and search words with high importance based on a search path that can grasp the relationship between a search word requested by a user and other search words related thereto as a distance, thereby enabling a user to easily select the most relevant related search word until the user's requested search word is reached and use it for marketing. [Background technology]

[0002] Recently, various strategies have been developed to promote products and services, and with the development of the Internet, keyword search has become an essential element of digital advertising / marketing strategies. In particular, in e-commerce, finding a single effective keyword for advertising can play such a big role that it can determine the results of your overall marketing.

[0003] Also, when a searcher searches for a product or service, the searcher repeats the process of exploration and evaluation before finally deciding on an action (subscription, purchase, etc.). For example, a search for a purchase often starts with a general keyword, then moves to a more specific brand or product keyword, and can be converted to a purchase through a journey of looking at reviews and rankings, or comparing competitors within a product line, etc. Here, the search words that ultimately contribute to the conversion are easy to find because they themselves have a relatively high search volume and conversion rate, but they may not be suitable in terms of advertising efficiency because the advertising unit price is high.

[0004] However, search terms that are in the process of exploration and evaluation around the conversion stage often have relatively low search volume or little advertising competition, and therefore may be more suitable in terms of advertising efficiency because the advertising cost is low.

[0005] As mentioned above, it is very important to secure search terms that are suitable for advertising and marketing in terms of advertising efficiency. However, the search terms in the process of exploration and evaluation around the conversion stage are often long-tail keywords that are relatively difficult to find, so there is a problem in using such keywords to create a marketing strategy or optimize advertising.

[0006] Although various existing search engines provide search terms with high advertising unit costs and related search terms, they only present simple related search terms, and because the related search terms are so diverse, it is difficult to separate out efficient related search terms until arriving at the search terms desired by marketers, sellers, etc., and it can be difficult to select search terms that can increase advertising efficiency.

[0007] In other words, marketing strategies need to be developed around conversion stages during a user's search journey (user needs for products and services, comparison items within a product range, and information needed before and after a purchase), and although there are a large number of keywords necessary for advertising optimization, the problem is that it is not easy for ordinary companies and users to discover and utilize them. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Korean Patent No. 10-1194296 Summary of the Invention [Problem to be solved by the invention]

[0009] The present invention generates a basic search word by adding different characters to a keyword, and then expands the search word by adding related search words provided by a search engine to the basic search word. Based on the search result searched through a search engine using the related search word as an input search word, a path is generated in which the relationship between the subsequent search word and the related search word is displayed as a distance based on the search result. Such paths are generated for various search words, and a search word network is generated based on a path list related to a search word requested by a user. Then, the path is visualized and provided so that the user can easily select related search words that are highly relevant until the requested search word is reached, thereby supporting the user to easily select related search words that can improve marketing efficiency. [Means for solving the problem]

[0010] The apparatus for providing a service for providing a search word network based on a search path according to an embodiment of the present invention automatically generates one or more basic search words by combining the keyword and characters by adding a series of characters to a keyword in different ways, obtains one or more related search words related to the basic search word through a preset search engine, applies the related search words to the search engine as input search words, extracts one or more subsequent search words based on returned search result list information, and calculates, as a distance, an exposure rank of the subsequent search word in a response result according to a response function corresponding to the subsequent search word among response functions used by the search engine to grasp the search intention of the input search word and a weighted value in a connection relationship between the input search word and the subsequent search word according to a function type of the response result. The apparatus generates relationship information on the distance between the input search word and the subsequent search word for each of one or more subsequent search words, and stores the relationship information in the search word relationship DB. The present invention may include a search path management unit that sets each of a plurality of search words as a node based on a plurality of relationship information, generates search path information for one or more search paths connecting a start node and an end node determined by a preset algorithm, and stores the search path information in a search path DB; and when receiving search word network request information including a requested search word and a search condition from a user terminal, extracts one or more search path information including the requested search word and satisfying the search condition from the search path DB, generates a search word network connecting different search words having a distance based on the extracted one or more search path information, clusters the search word network into a plurality of different search word groups using a preset clustering algorithm, and sets different attributes so that the plurality of different search word groups can be visually separated from each other, and provides the generated search word network information to the user terminal.

[0011] As an example related to the present invention, the search path management unit includes a search word extension unit that automatically generates one or more basic search words by combining the keyword and characters by adding a series of characters to a keyword in a different way, and obtains one or more related search words related to the basic search word through a preset search engine; and a search word extension unit that applies the related search words to the search engine as input search words and extracts one or more subsequent search words from one or more response results corresponding to the types of functions for one or more response functions used by the search engine based on returned search result list information by grasping the search intention of the input search word, and then exposes the subsequent search words in the response results corresponding to the subsequent search words. The distance calculation method may include a search information extraction unit that generates relationship information on the distance between the input search word and a subsequent search word for each of one or more subsequent search words and stores the relationship information in a search word relationship DB, and a search path extraction unit that generates a directed weighted graph in which a distance between the nodes is set by setting each of a plurality of search words as a node based on a plurality of relationship information stored in the search word relationship DB, and then generates search path information for one or more search paths connecting a start node and an end node determined by a preset algorithm in the directed weighted graph and stores the search path information in the search path DB.

[0012] As an example related to the present invention, the search information extraction unit may calculate the weighted value by multiplying a preset priority for a type of function corresponding to the subsequent search word and an exposure ranking of the subsequent search word in a response result corresponding to the subsequent search word, and set the calculated weighted value as the distance between the subsequent search word and the input search word.

[0013] As an example related to the present invention, the search information extraction unit may apply the subsequent search word to the search engine as another input search word, and based on the returned search result list information, extract one or more texts from each of one or more response results corresponding to the function types of one or more response functions used by the search engine to grasp the search intent of the input search word according to the distance calculation method as an additional subsequent search word corresponding to the subsequent search word, calculate a weighted value in a connection relationship between the input search word and the additional subsequent search word according to an exposure rank of the additional subsequent search word in the response result corresponding to the additional subsequent search word and a function type of the response result as a distance, and generate relationship information in the distance between the one or more additional subsequent search words and the input search word that is the subsequent search word, and store the information in the search word relationship DB.

[0014] As an example related to the present invention, the search conditions may include a connection direction and a number of hops with other nodes based on a node corresponding to the requested search word, and the search result providing unit may extract one or more search path information satisfying the requested search word and the search conditions, apply the extracted one or more search path information to a ForceAtlas2 algorithm to generate a search word network, and apply the search word network to a Louvain algorithm to generate search word network information in a search word network clustered into a plurality of different search word groups.

[0015] As an example related to the present invention, the search path management unit acquires and stores the search volume for each search word through the external server or the search engine, and the search result providing unit checks the search volume for each search word belonging to the search word network, and generates a search word network in which the size of a node corresponding to a search word is adjusted according to the search volume of the search word. After checking one or more peripheral nodes connected to a specific node corresponding to a search word for each search word included in the search word network, the search word network may be configured to calculate a network centrality index of a search word corresponding to the specific node considering the number of the peripheral nodes and the centrality of the specific node according to the position of each peripheral node and the distance between the specific node and the peripheral node, and to separate a node of a search word having a network centrality index equal to or greater than a preset reference value from another node having a network centrality index less than the reference value according to the network centrality index calculated for each search word.

[0016] As an example related to the present invention, the search result providing unit may be characterized by calculating a network centrality index of the search word using at least one of degree centrality, betweenness centrality, closeness centrality, and PageRank.

[0017] As an example related to the present invention, the search result providing unit applies search words belonging to the search word network to the search engine, and in one or more returned search result pages, based on a plurality of predetermined stage-based search intention judgment conditions for grasping search intention, checks whether one or more predetermined search result items exist from the top N search results according to the search intention judgment conditions, identifies one or more stage-based search intention judgment conditions satisfied by the one or more search result pages from the plurality of stage-based search intention judgment conditions, and sets a predetermined search intention corresponding to the identified one or more stage-based search intention judgment conditions from among the plurality of predetermined search intentions to the search words and includes them in the search word network, thereby generating a search word network in which the search intention for each search word is set.

[0018] As an example related to the present invention, the one or more search result items set as the search intent judgment criteria may include at least one of a knowledge encyclopedia page, a featured snippet, a knowledge panel, a video, an academic information search, a news article, a local map, a site link, an ad carousel, a keyword ad, and a pre-stored conversion page URL.

[0019] A service providing method for providing a search word network based on a search path of a service providing device according to an embodiment of the present invention includes the steps of: automatically generating one or more basic search words by combining a keyword with characters by adding a series of characters to a keyword in different ways, and acquiring one or more related search words related to the basic search word through a preset search engine; applying the related search words to the search engine as input search words, extracting one or more subsequent search words based on returned search result list information, and calculating, as a distance, a weighted value in a connection relationship between the input search word and the subsequent search word according to an exposure rank of the subsequent search word in a response result according to a response function corresponding to the subsequent search word among response functions used by the search engine to grasp a search intention of the input search word, and a function type of the response result; generating relationship information in a distance between the input search word and the subsequent search word for each of one or more subsequent search words, and storing the relationship information in a search word relationship DB; and a step of storing the search word network information in the search path DB; setting each of a plurality of search words as a node based on a plurality of relation information stored in the search word relation DB, generating search path information for one or more search paths connecting a start node and an end node determined by a preset algorithm, and storing the search path information in the search path DB; and when receiving search word network request information including a requested search word and a search condition from a user terminal, extracting one or more search path information including the requested search word and satisfying the search condition from the search path DB, generating a search word network based on the one or more extracted search path information, clustering the search word network into a plurality of different search word groups according to a preset clustering algorithm, and setting different attributes so that the plurality of different search word groups can be visually separated from each other, and providing the generated search word network information to the user terminal. Effect of the Invention

[0020] The present invention can obtain the relationship between search words as distance according to the search intent of an input search word and the exposure rank of subsequent search words corresponding to the input search word using a search engine, and based on that, generate and provide a search word network in which a plurality of search words with high interrelationships are grouped and visualized based on the distance between the search words according to search path information obtained for a plurality of different search words. The present invention can provide a search word network that allows a user to easily check search words that have a certain level of relevance to a requested search word and a certain level of search volume. In addition, the present invention can support a user to easily find search words with high marketing efficiency and low price competition from among search words located around the requested search word checked through the search word network, thereby greatly improving user satisfaction and usability.

[0021] In addition, the present invention can set and provide the most likely search intention among users, etc. when using search words belonging to a search word network in a search word network, thereby helping a seller who is trying to conduct marketing to easily find search words that have search intentions that match the business objectives of the seller and are efficient for marketing through the search word network.

[0022] Furthermore, the present invention not only provides a way to grasp the main search journey of users for products and services by expressing the connection relationships between search words beyond the level of simply providing related search words through a search word network generated based on a search path, but also has the effect of helping users to more intuitively grasp their needs, such as what they are requesting and what they are comparing. [Brief description of the drawings]

[0023] [Figure 1]2 is a configuration diagram of a service providing device for providing a search word network based on a search path according to an embodiment of the present invention. [Diagram 2] 4 is an exemplary diagram illustrating an operation of a service providing device according to an embodiment of the present invention; [Diagram 3] 11 is a diagram illustrating an example of a process of acquiring related search words of a service providing device according to an embodiment of the present invention; [Figure 4] FIG. 2 is an exemplary diagram of a SERP Feature (function type) type definition of a service providing device according to an embodiment of the present invention. [Diagram 5] 13 is a diagram illustrating an example of subsequent search word detection by a service providing device according to an embodiment of the present invention; [Figure 6] 13 is a diagram illustrating an example of generating a search word relation graph in a service providing device according to an embodiment of the present invention. [Figure 7] 11 is an exemplary diagram illustrating a search path generation of a service providing device according to an embodiment of the present invention. [Figure 8] 13 is an exemplary diagram illustrating a service providing device providing recommendation list information based on an API including a recommendation search path according to an embodiment of the present invention. [Figure 9] 13 is an exemplary diagram illustrating a recommended search password network UI of a service providing device according to an embodiment of the present invention. [Figure 10] 13 is an exemplary diagram illustrating a node highlight function in a recommended search password network UI of a service providing device according to an embodiment of the present invention. [Figure 11] 13 is an exemplary diagram illustrating a process of receiving search word network request information of a service providing device according to an embodiment of the present invention. [Figure 12] 1 is an exemplary diagram of a search word network generated by a service providing device according to an embodiment of the present invention; [Figure 13] 1 is an exemplary diagram of a search word network generated by a service providing device according to an embodiment of the present invention; [Figure 14] 13 is an exemplary diagram illustrating a process of generating a search word network including search intent of a service providing device according to an embodiment of the present invention. [Figure 15]13 is an exemplary diagram illustrating a process of generating a search word network including search intent of a service providing device according to an embodiment of the present invention. [Figure 16] 13 is an exemplary diagram illustrating a process of generating a search word network including search intent of a service providing device according to an embodiment of the present invention. [Figure 17] 13 is an exemplary diagram illustrating a process of generating a search word network including search intent of a service providing device according to an embodiment of the present invention. [Figure 18] FIG. 13 is an exemplary diagram of a search word network including search intent for each search word generated by a service providing device according to an embodiment of the present invention. [Figure 19] 1 is a flowchart of a service providing method for providing a search word network based on a search path according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0024] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, detailed embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a configuration diagram of a service providing device (hereinafter, referred to as a service providing device) for providing a search word network based on a search path according to an embodiment of the present invention.

[0025] As shown in the figure, the service providing device (100) according to an embodiment of the present invention may be configured to include a communication unit (20), a storage unit (30), a control unit (10), etc., but may be configured to include various components without being limited thereto.

[0026] First, the communication unit 20 can communicate with one or more user terminals and various external servers via a communication network.

[0027] Furthermore, the communication network described in the present invention may include wired / wireless communication networks. Examples of such wireless communication networks include Wireless LAN (WLAN), DLNA (Digital Living Network Alliance), Wibro (Wireless Broadband: Wibro), Wimax (World Interoperability for Microwave Access: Wimax), GSM (Global System for Mobile communication), CDMA (Code Division Multi Access), CDMA2000 (Code Division Multi Access 2000), EV-DO (Enhanced Voice-Data Optimized or Enhanced Voice-Data Only), WCDMA (Wideband CDMA), HSDPA (High Speed ​​Downlink Packet Access), HSUPA (High Speed ​​Uplink Packet Access), IEEE 802.16, Long Term Evolution (LTE (Long Term Evolution)), LTE-A (Long Term Evolution-Advanced), and Wireless Mobile Broadband (Wireless Mobile Broadband). These include Wireless Broadband Service (WMBS), 5G mobile communications services, Bluetooth (registered trademark), LoRa (Long Range), RFID (Radio Frequency Identification), Infrared Data Association (IrDA), Ultra Wideband (UWB), ZigBee (registered trademark), Near Field Communication (NFC), Ultra Sound Communication (USC), Visible Light Communication (VLC), Wi-Fi, and Wi-Fi Direct.Furthermore, examples of wired communication networks include wired LAN (Local Area Network), wired WAN (Wide Area Network), Power Line Communication (PLC), USB communication, Ethernet (registered trademark), serial communication, and optical / coaxial cables.

[0028] In addition, the storage unit (30) can store various information, and can be configured in various forms such as a hard disk drive (HDD) or a solid state drive (SSD), and can also be configured to include one or more DBs.

[0029] As an example, the storage unit (30) may be configured to include a plurality of DBs necessary for the operation of the service providing device (100), and may include various DBs such as a keyword DB (101) in which a plurality of different keywords are stored, a search word relationship DB (102) in which relationship information regarding the relationship between search words is stored, a function type DB (103) in which function types corresponding to search words and related information are stored, a search path DB (104) in which search paths and statistical information corresponding to search words are stored, and a search volume DB (105) in which search volume information including item-specific statistics preset for each search word is stored.

[0030] Here, the multiple DBs included in the storage unit (30) may each be configured as a separate database server, and the service providing device (100) may communicate with and be linked to the multiple different database servers via a communication network.

[0031] In addition, the control unit (10) performs the overall control function of the service providing device (100), and the control unit (10) may include a RAM, a ROM, a CPU, a GPU, and a bus, and the RAM, ROM, CPU, GPU, etc. may be connected to each other via a bus.

[0032] The communication unit (20) and the storage unit (30) may also be configured as being included in the control unit (10). Here, at least one of the control unit (10) and the various components constituting the control unit (10) can communicate with the user terminal and an external server via the communication unit (20), and in the following, the communication configuration through the communication unit (20) will be omitted.

[0033] As shown in the figure, the control unit (10) can be configured to include a search word expansion unit (111), a search information extraction unit (112), a search path extraction unit (113), and an optimal path recommendation unit (114). Here, the multiple components constituting the control unit (10) can be realized by a processor capable of data processing, and each component can be separated and realized by a different processor, or can be functionally separated within a single processor.

[0034] Based on the above-mentioned configuration, the detailed operation configuration of the service providing device (100) will be described with reference to the following drawings.

[0035] FIG. 2 is a detailed operational configuration diagram of the service providing device (100) according to the embodiment of the present invention, and the configuration of the control unit (10) which performs the substantial control function of the service providing device (100) will be mainly described.

[0036] As shown in the figure, the search word extension unit (111) can extract a keyword from a keyword DB (101) in which a plurality of different keywords for companies, brands, products, services, categories, etc. are stored, and can automatically generate one or more search words each of which is a combination of the keyword and one or more characters by adding a series of characters to the keyword in a different manner, as a basic search word.

[0037] Here, the search word expansion unit 111 can communicate with an external knowledge server that provides an online electronic dictionary, collect the keywords from the external knowledge server, and store the keywords in the keyword DB 101. Examples of such an external knowledge server include a server that provides DBPEDIA or Wikipedia (registered trademark).

[0038] As an example, the search word expansion unit (111) can connect to the external knowledge server and extract and store the page name or the title of the knowledge panel as the keyword.

[0039] In addition, the search word expansion unit 111 can obtain one or more related search words related to the basic search word through a preset search engine. To explain this with reference to FIG. 3, the search word extension unit (111) extracts the keyword "Kookin Bank" stored in the keyword DB (101), adds "Ga" to the keyword to generate a basic search word such as "Kookin Bank Ga", and adds "Na" to the keyword to generate a basic search word such as "Kookin Bank Na".

[0040] In addition, the search word extension unit (111) can add characters (or letters) from "da" to "ha" in order to the keyword "Kookmin Bank" to generate basic search words corresponding to each added character. In addition to the above examples, it can also add a character with a final sound such as "gan" to the keyword, add a foreign word such as "A", or add multiple characters such as "ga-ga" and "AA".

[0041] In addition, the characters that the search word extension unit 111 can add to the keyword can include blank characters and special characters.

[0042] In addition, the search word expansion unit (111) can apply the search word "Kookmin Bank.ga" generated as described above to a search engine preset in the service providing device (100) or a search engine provided by an external server through communication with the external server, and obtain related search words related to the "Kookmin Bank.ga" from the search engine.

[0043] As an example of such a search engine, the search engines "NAVER" (registered trademark) and "GOOGLE" (registered trademark) can be used, and when the search engine is included in the service providing device (100), execution data related to the search engine can be stored in the storage unit (30) of the service providing device (100).

[0044] In addition, the related search words may be autocomplete search words (or autocomplete query languages) that the search engine autocompletes when a basic search word is input and meaningful search words (or query languages) are generated.

[0045] For example, when the search engine inputs the basic search word "Kookmin Bank Ga" from the search word expansion unit (111), the search engine automatically generates related search words such as "Kookmin Bank virtual account" and "Kookmin Bank household account book" which are auto-complete search words starting with "Kookmin Bank Ga" or including the basic search word, and the search word expansion unit (111) can obtain one or more related search words for one basic search word from the search engine.

[0046] As another example, when the basic search word “iPhone-ga” is input from the search word expansion unit (111), the search engine can automatically generate related search words such as “iPhone price” and “iPhone forced shutdown” that start with “iPhone-ga” or are autocomplete search words that include the basic search word.

[0047] Alternatively, the search engine may search a database that includes an external server that provides the search engine and stores various search words based on the basic search word, obtain related search words that start with the basic search word or that include the basic search word, and then provide the related search words to the search word extension unit (111).

[0048] In addition, the search engine may be included in an external server that provides the search engine based on a basic search word, and may search the DB in which various search words are stored to generate related search words in a sentence, or may extract the sentence including the basic search word from the DB and provide it to the search word extension unit (111) as related search words.

[0049] In addition, the search word extension unit (111) can store the related search words acquired through the search engine as keywords in the keyword DB (101).

[0050] Meanwhile, the search information extraction unit (112), in cooperation with the search word extension unit (111), may apply the one or more related search words corresponding to the basic search word to the search engine as input search words, and obtain search result list information including one or more search results corresponding to the input search word through the search engine.

[0051] In addition, the search results described in the present invention may be search result information generated by the search engine, and the search result information may include search language, search area, search engine name, search result position (rank), domain, feature type (SERP Feature), title, body text, etc., and may be composed of a text-based document.

[0052] In addition to providing document-based search results on websites or web pages for an input search word, the search engine may grasp the search intent of the input search word in the search engine, generate direct response results using a response function corresponding to the grasped search intent among one or more different unique response functions provided by the search engine, and then provide the result as the search result.

[0053] For example, as shown in FIG. 4, a search engine provides response results for one or more response features as search results through a SERP (Search Engine Result Page), and generates and provides response results according to a function type (SERP feature) corresponding to the search intent among a plurality of different response function types pre-set in the search engine, such as keyword advertisement (AD), advertisement carousel, application, featured snippets, image, job search, knowledge panel, people also search for, related search words, etc., based on the search intent.

[0054] For example, as shown in FIG. 5, when an input search word is input, a search engine can grasp the search intent of the input search word and generate multiple response results each corresponding to a response function such as “People also search for,” which is a response function in a collection of other search words searched together with the search word, and “related search words,” which is a response function in a collection of other search words related to the search word. Such response results may include an identifier of the type of function (response function) corresponding to the response result, a response processing result (or response content) obtained through the response function, etc.

[0055] That is, when an input search word for "women's perfume recommendations" is input, the search engine grasps the search intent of the input search word, and based on the grasped search intent, calculates a response result including one or more subsequent search words such as "women's perfume rankings 2021" or "women's perfume rankings 2021" using the "People also search for" response function, or calculates a response result including one or more subsequent search words such as "Olive Young women's perfume recommendations" or "women's perfume gifts" using the "related search word" response function, and returns the response result as a search result.

[0056] In addition, the search information extraction unit (112) may check the search result list information and delete or exclude related search words that do not search anything, and not store them in the keyword DB (101). That is, the search information extraction unit (112) may exclude or delete related search words whose search volume is 0, without storing them in the keyword DB (101).

[0057] In addition, when one or more response results are included in one or more search result list information corresponding to one or more related search words corresponding to the basic search word, the search information extraction unit (112) can identify function types from each of the one or more response results, generate function type information including the identified function types and the occurrence frequency of each identified function type, and match the generated function type information with the corresponding related search words, and store the generated function type information in a function type DB (103).

[0058] In addition, the search information extraction unit (112) may extract, as a subsequent search word, one or more texts from one or more response results each corresponding to a function type for one or more response functions used by the search engine to grasp the search intent of the input search word based on search result list information returned by the search engine for the input search word, and generate relationship information in the distance between the input search word and the subsequent search word for each of the extracted one or more subsequent search words using a distance calculation method that calculates, as a distance, a weighted value in the connection relationship between the input search word and the subsequent search word based on the exposure rank of a specific subsequent search word in the response result corresponding to the specific subsequent search word and the function type of the response result corresponding to the specific subsequent search word.

[0059] Here, the search information extraction unit (112) may calculate the weighted value by multiplying a preset priority for a type of function corresponding to the subsequent search word by an exposure rank (or a placement rank) of the subsequent search word in a response result corresponding to the subsequent search word, and set the calculated weighted value as the distance between the subsequent search word and the input search word.

[0060] As an example, the search information extraction unit (112) can identify multiple response results (search results) included in the search result list information of the input search word “women’s perfume recommendations” and the corresponding function types “others also searched for” and “related search words.”

[0061] Here, the search information extraction unit (112) may be preset with setting information in which one or more function types that are targets for extracting subsequent search words among a plurality of function types provided by the search engine are preset, and the search information extraction unit (112) may extract subsequent search words only from the function types that are targets for extracting the subsequent search words.

[0062] In addition, the setting information may preset different priorities among one or more types of functions from which the subsequent search words are extracted.

[0063] Accordingly, the search information extraction unit (112) assigns priority values ​​of 1 and 2 to each of the multiple function types, "others also search for" and "related search words", identified in the search list information corresponding to the input search word "women's perfume recommendations" according to the setting information, assigns an exposure rank (or an exposure rank weighting value) according to the ranking of the subsequent search word within the function type, extracts one or more subsequent search words from each of the multiple response results, such as ["Women's perfume ranking 2021", priority 1, exposure rank 1, distance 1], ["Women's perfume ranking 2021 for women in their 20s", priority 1, exposure rank 2, distance 2], ["Olive Young women's perfume recommendations", priority 2, exposure rank 1, distance 2], and ["Women's perfume gift", priority 2, exposure rank 2, distance 4], etc., calculates a weighting value, which is a value obtained by multiplying the priority and exposure rank for each of the extracted one or more subsequent search words, and then calculates the weighting value as a distance (distance weighting value) from the input search word.

[0064] Here, the search information extraction unit 112 may identify (detect) the placement order of the subsequent search words for each subsequent search word from the response result, and may determine the exposure order of the subsequent search words according to the placement order.

[0065] In addition, the priority may be defined as a priority weight, the exposure rank may be defined as an exposure rank weight, and the distance calculated as a product of the priority weight and the exposure rank weight may be defined as a distance weight.

[0066] In addition, the search information extraction unit (112) can generate relationship information including the calculated distance and the subsequent search word and the input search word for each of one or more subsequent search words corresponding to the input search word based on the input search word, and can store the relationship information in the search word relationship DB (102).

[0067] From the above configuration, the search information extraction unit (112) can apply the subsequent search word to the search engine as another input search word, and thereafter, applies the subsequent search word to the search engine as an input search word, and extracts one or more texts from one or more response results corresponding to the function types of one or more response functions used by the search engine to grasp the search intent of the input search word according to the distance calculation method based on the returned search result list information, as additional subsequent search words corresponding to the subsequent search word, and then calculates a weighted value in the connection relationship between the input search word and the additional subsequent search word according to the exposure rank of the additional subsequent search word in the response result corresponding to the additional subsequent search word and the function type of the response result corresponding to the additional subsequent search word as a distance, and generates relationship information in the distance between the one or more additional subsequent search words and the input search word that is the subsequent search word, and stores the information in the search word relationship DB (102).

[0068] That is, the search information extraction unit (112) may apply the subsequent search word to a search engine again as an input search word each time it extracts a subsequent search word corresponding to the input search word, repeatedly extract other subsequent search words related to the subsequent search word according to the distance calculation method, generate relationship information in the distance relationship between the input search word related to the other subsequent search word and the other subsequent search word, and store the relationship information in the search word relationship DB (102).

[0069] Meanwhile, the search path extraction unit (113) may generate a directed weighted graph by setting each of a plurality of search words corresponding to a plurality of relationship information as a node based on a plurality of relationship information stored in the search word relationship DB (102) and setting a distance between the nodes, and may then calculate path information for one or more paths connecting a start node and an end node determined by a preset algorithm in the directed weighted graph, and may generate statistical information by averaging preset item-based statistics for a plurality of search words included in the path information in conjunction with an external server (including) that provides the search engine, and may then match the path information and store the statistical information in the search path DB (104).

[0070] Here, the search path extraction unit (113) can determine that a distance value exists between a plurality of different search words based on the relationship information, and can generate the directed weighted graph by connecting the plurality of different search words having distance values ​​with connecting lines and setting distance values ​​to the connecting lines.

[0071] Furthermore, the path may be configured to include one or more connecting lines for reaching the end node from the start node and nodes each corresponding to the one or more connecting lines, and hereinafter, the path is referred to as a search path, and path information corresponding to the path is referred to as search path information.

[0072] As an example, as shown in FIG. 6, the search path extraction unit (113) extracts connection relationships and distance information (weighted value information) between multiple search words included in multiple relationship information based on multiple relationship information stored in the search word relationship DB (102), sets each search word as a node, and generates a directed weighted graph with edges representing forward relationships in which the distance (distance value or distance weighted value) according to the relationship information is added to the distance between different nodes (or the search path connecting different nodes).

[0073] As an example, the directed weighted graph may be configured to include input search words and subsequent search words as nodes, relationships between the input search words and subsequent search words as forward edges, and distance weights between the input search words and subsequent search words may be set on the edges.

[0074] Here, the search path extraction unit (113) can store the directed weighted graph in the search path DB (104).

[0075] In addition, the search path extraction unit (113) may select an arbitrary search word as a start node according to a preset algorithm in the directed weighted graph, and then select N nodes with high relative importance as an end node from among the nodes connected in a forward direction within L hops from the start node according to the preset algorithm.

[0076] In addition, the relative importance of different nodes based on the start node can be determined using a Pagerank algorithm that determines the importance of a web document.

[0077] In addition, the search path extraction unit (113) may generate search path information for each of the one or more search paths connecting the start node and the end node at each of the one or more selected end nodes based on the selected start node according to the directed weighted graph.

[0078] In addition, the search path extraction unit (113) can change the start node and identify one or more end nodes corresponding to the start node through the algorithm, calculate K pre-defined search paths in order of shortest search path distance (or total search path distance) among the search paths connecting the start node and the end node, and then generate the search path information for each of the calculated search paths.

[0079] In addition, the search path extraction unit (113) can generate path list information including one or more pieces of search path information whose start node and end node are the same. That is, the search path extraction unit (113) can generate path list information by grouping one or more pieces of search path information whose start node is a first search word and whose end node is a second search word.

[0080] Also, as shown in FIG. 7, when generating the path list information, the search path extraction unit (113) can communicate with an external server that provides the search engine and request pre-set item-specific statistics for a plurality of search words included in the path list information from the external server, and can generate statistics list information based on the item-specific statistics collected (received) for the plurality of search words from the external server.

[0081] Here, the preset items may include search volume, advertising cost, competitive index, etc., and the search path extraction unit (113) may store the search volume information including statistics for each of the preset items for each of the plurality of search words from the external server in a search volume DB (105) included in the service providing device (100).

[0082] In addition, when the search engine is included in the service providing device (100), the control unit (10) of the service providing device (100) may further include a search volume generation unit that generates the search volume information in conjunction with the search engine, and the search volume generation unit may store the search volume information in the search volume DB (105) each time it generates search volume information for each of a plurality of search words in conjunction with the search engine.

[0083] The search volume generating unit may be configured to be included in the search path management unit (110) described below, and the search volume generating unit may communicate with the external server to request pre-defined item-by-item statistics for a plurality of search words included in the search path information, generate search volume information including the item-by-item statistics received from the external server, and store the search volume information in the search volume DB. The search volume information may include search words and the item-by-item statistics corresponding to the search words.

[0084] Also, the search path extraction unit (113) can match the statistics list information corresponding to the path list information with the path list information and store the matched information in the search path DB (104).

[0085] In addition, the search path extraction unit (113) can check the search volume of the search word corresponding to the end node from one or more search path information pieces having the same start node and end node, and can exclude the search path information pieces corresponding to the end node whose search volume is 0 from the path list information.

[0086] In addition, the search path extraction unit (113) may confirm the predetermined item-specific statistics of one or more search words included in the search path information in each of one or more search path information included in the path list information based on the statistics list information, and then average the confirmed search word-specific statistics for the predetermined items to generate statistical information including an average value of the item-specific statistics.

[0087] As an example, the statistical information may include an average search volume for multiple search words included in specific search path information, an average advertising cost for multiple search words included in specific search path information, an average competitive index for multiple search words included in specific path information, etc.

[0088] In addition, when generating the statistical information, the search path extraction unit (113) can match the statistical information with search path information corresponding to the statistical information and store the matched information in the search path DB (104).

[0089] Here, when generating search path information, the search path extraction unit (113) can set a unique identifier to the search path information, and can also set the unique identifier in statistical information matching the search path information, thereby enabling mutual matching of the search path information and the statistical information.

[0090] Alternatively, the search path extraction unit (113) can include the statistical information in search path information corresponding to the statistical information and store the information in the search path DB (104).

[0091] An example of this will be described with reference to FIG. 7. As shown in the figure, when the search path extraction unit (113) selects "women's perfume recommendations" as the start node (starting search word), it selects "faint women's perfume" as the end node (ending search word) according to the algorithm, and generates a plurality of search path information corresponding to a first search path ("women's perfume" → "women's perfume recommendation ranking 2021" → "faint women's perfume") and a second search path ("women's perfume" → "women's perfume ranking for women in their 20s 2021" → "faint women's perfume") connecting the start node and the end node based on the directed weighted graph, and then generates path list information including the plurality of search path information.

[0092] In addition, the search path extraction unit (113) can, in conjunction with the external server, identify a plurality of search words included in the path list information based on one or more search path information included in the path list information, collect item-specific statistics for each of the plurality of search words, and generate statistics list information.

[0093] In addition, the search path extraction unit (113) can check the statistics for each of a plurality of pre-set items for each of a plurality of search words including “women's perfume”, “women's perfume recommendation ranking 2021”, and “faint women's perfume” based on the first search path information corresponding to the first search path included in the path list information and the statistics list information corresponding to the path list information, calculate the average search volume (347), the average advertising cost (0.44), and the average competitive index (0.95), generate first statistical information including the calculated average search volume, average advertising cost, and average competitive index, match the first statistical information with the first search path information, and store it in the search path DB (104).

[0094] Similarly, the search path extraction unit (113) can calculate second statistical information including an average search volume (373), an average advertising price (0.34), and an average competitive index (0.91) based on second search path information corresponding to the second search path included in the path list information and statistical amount list information, and then match the second statistical information with the second search path information corresponding to the second statistical information and store it in the search path DB (104).

[0095] Meanwhile, when the optimal path recommendation unit (114) receives request information (search request information) including a requested search word and a ranking condition from a user terminal, it can generate recommendation list information (recommended path list information) by rearranging one or more search path information including the requested search word based on the ranking condition and statistical information, and provide the recommendation list information to the user terminal.

[0096] Here, the service providing device (100) may include a user input unit for receiving user input, and the optimal path recommender (114) included in the control unit (10) may receive the request information based on user input via the user input unit.

[0097] Here, the ranking conditions may be various conditions such as search volume, advertisement cost, competitive index, and the presence or absence of a type of function.

[0098] That is, when the optimal path recommendation unit (114) receives the request information from the user terminal, it extracts one or more search path information and one or more statistical information corresponding to the requested search word included in the request information from the search path DB (104), calculates scores for the extracted one or more statistical information according to a preset score calculation standard for one or more items set in the ranking conditions included in the request information, sorts the extracted one or more statistical information in order of scores, and then selects one or more search path information corresponding to statistical information that is ranked at or above a preset criterion rank from among the one or more search path information corresponding to each of the sorted one or more statistical information, and provides recommendation list information including the selected one or more search path information to the user terminal.

[0099] Here, the optimal path recommendation unit (114) can determine and rearrange the ranks of one or more search path information corresponding to the request information in a ranking determination method in which the calculated score for the statistical information corresponding to the search path information is compared with the score of another statistical information (etc.) to calculate the ranking of the statistical information, and then determine the calculated ranking as the ranking of the search path information corresponding to the statistical information, and select one or more search path information that is equal to or higher than a predetermined standard ranking from among the one or more search path information corresponding to the request information.

[0100] From the above-mentioned configuration, the optimal path recommendation unit (114) can extract only one or more search path information having the requested search word as an end node from the search path DB (104), and generate recommendation list information based on the extracted one or more search path information.

[0101] That is, the optimal path recommender 114 may select a search path consisting of one or more search words necessary to reach the requested search word as a recommended search path, and generate recommended list information.

[0102] In addition, the optimal path recommendation unit (114) may include at least one of statistical information corresponding to search path information corresponding to each of one or more search path information included in the recommendation list information, search volume information including item-by-item statistics for each of a plurality of search words included in the search path information, and function type information of the function types of the response results corresponding to each of a plurality of search words included in the recommended search path information, in the recommendation list information, and then provide the recommendation list information to the user terminal in the form of a UI or an API.

[0103] Here, the optimal path recommendation unit (114) may generate analysis result information including at least one of the following information in the form of a UI or API: statistical information corresponding to the search path information for each of one or more search path information included in the recommendation list information; search volume information including item-by-item statistics for each of a plurality of search words included in the search path information; and function type information for the function types of the response results corresponding to each of a plurality of search words included in the search path information, and provide the analysis result information to the user terminal.

[0104] To this end, the optimal path recommendation unit (114) can provide the recommendation list information to the user terminal by including one or more statistical information corresponding to one or more search path information included in the recommendation list information, thereby allowing the user terminal to display specific statistical information corresponding to specific search path information selected by the user from the one or more search path information included in the recommendation list information together with the specific search path information.

[0105] In addition, the optimal path recommending unit (114) can request statistics for each item preset for a plurality of search words included in the recommendation list information from the external server, generate statistics list information based on the statistics for each item collected (received) for a plurality of search words corresponding to the recommendation list information from the external server, and provide the statistics list information to the user terminal by including it in the recommendation list information. Here, the preset items can include search volume, advertisement cost, competition index, etc.

[0106] In addition, the optimal path recommendation unit (114) can extract search volume information for each of a plurality of search words included in the recommendation list information from the search volume DB (105), and generate the statistics list information based on the extracted search volume information for each of a plurality of searches, and then include the statistics list information in the recommendation list information and provide it to the user terminal.

[0107] Thereby, when the user terminal receives the recommendation list information, it displays specific search path information selected by the user from one or more search path information included in the recommendation list information, and when a specific search word included in the specific search path information is selected by the user, it can display and provide statistics (or search volume information) for each item related to the specific search word.

[0108] As an example of the above, as shown in FIG. 8, when the optimal path recommender (114) receives request information, when the requested search word "Anbang Grill" according to the request information and the ranking condition "order by cost per click (CPC)" according to the request information are received, the optimal path recommender (114) acquires one or more search path information corresponding to the requested search word "Anbang Grill" as recommended search path information from the search path DB (104), and then acquires one or more statistical information corresponding to the one or more recommended search path information as interest statistical information from the search path DB (104).

[0109] In addition, the optimal path recommendation unit (114) can generate and return recommendation list information in the form of an API (Application Programming Interface) in which the one or more pieces of recommended search path information are sorted in order of advertising cost, which is a ranking condition, based on the one or more pieces of interest statistical information.

[0110] As an example, the CPC values ​​of the search terms “grill”, “grill”, “grill recommendation”, “Anbang Grill”, “Anbang Grill 501”, “Anbang Grill AB301MF”, and “fish grill” from the first recommendation search path with the highest ranking in the recommendation list information, “grill” → “grill” → “grill recommendation”, “Anbang Grill”, “Anbang Grill 501”, “Anbang Grill AB301MF”, and “fish grill”, are 2.503594, 0.119853, and 0.12749, respectively. 8, 0.248646, 0, 0.118544, and 0.28, the average CPC value of the search words etc. included in the first recommended search path, which is 0.4854, can be calculated (returned) as the score of the first recommended search path, and for each of the other recommended search paths included in the recommendation list information, the score can be calculated in the same manner as the score of the first recommended search path, and one or more recommended search paths etc. included in the recommendation list information can be rearranged in order of score.

[0111] Here, the optimal path recommendation unit (114) can check the average CPC value of the search words, etc. included in the first recommended search path from statistical information (interest statistical information) corresponding to the first recommended search path information, and calculate it as a score.

[0112] In addition, in the above-mentioned configuration, the optimal path recommendation unit (114) acquires one or more search path information corresponding to the requested search word from the search path DB (104) as recommended candidate path information, calculates a score for each of the one or more recommended candidate path information as described above, selects one or more recommended candidate path information whose score is calculated to be equal to or greater than a preset reference value as recommended search path information, includes only the selected recommended search path information in the recommendation list information, and provides the recommendation list information to the user terminal as a final result (final result information) corresponding to the requested information.

[0113] As a result, the optimal path recommendation unit (114) can provide only search paths that are highly relevant to the requested search words and ranking conditions in the final results, and can exclude search paths that are less relevant to the requested search words and ranking conditions from the final results.

[0114] Also, as shown in FIG. 9, when the optimal path recommendation unit (114) receives the requested search word “Anbang Grill” and the ranking condition “order by cost per click (CPC)”, it can diagram and output a word network UI based on the recommendation list information acquired corresponding to the requested search word.

[0115] For example, when the first recommended search path information included in the recommendation list information acquired in response to the request information is 'grill' → 'grill' → 'grill recommendation' → 'Anbang Grill' → 'Anbang Grill 501' → 'Anbang Grill AB301MF' → 'fish grill' and the second recommended search path information included in the recommendation list information is 'oil-free grill' → 'Anbang Grill' → 'Anbang Grill drawbacks' → 'electric grill drawbacks', the optimal path recommendation unit (114) may generate and provide recommendation list information configured in a UI (User Interface) in the form of a flow chart in which each search word is represented by a box, a previous search word in 'Anbang Grill', which is a specific node corresponding to the requested search word on the path, is arranged to the left of the specific node, and a subsequent search word in the specific node is arranged to the right of the specific node, and different nodes (search words) are interconnected by edges according to the first and second recommended search path information.

[0116] In addition, the optimal path recommendation unit (114) can represent the height of the box by the size (or number) of the incoming edge or outgoing edge of each search word, or represent the color of the box by information of the search words on the path (search volume, advertising cost, competitive index, etc.).

[0117] Also, as shown in FIG. 10, the optimal path recommendation unit (114) can extract function type information corresponding to a search word for each of a plurality of search words included in the recommendation list information from the function type DB (103), match the search word, add it to the recommendation list information, and provide it to the user terminal. Thereby, when a UI according to the recommendation list information is displayed on the user terminal, one or more boxes corresponding to one or more search words corresponding to a specific function type selected by the user may be highlighted and displayed, or one or more boxes corresponding to the remaining search words excluding the one or more search words corresponding to the specific function type may be filtered and displayed.

[0118] As described above, the present invention applies each basic search word obtained by adding different characters to a keyword to a search engine, obtains related search words related to the basic search words, expands the search words, applies the expanded search words to the search engine as input search words, and obtains the relevance between one or more subsequent search words obtained based on the search results searched through the search engine and the input search word as a distance according to the search intention of the input search word grasped by the search engine and the exposure rank of the subsequent search word. Based on the distance between such search words, the search words having a distance are interconnected to generate a search path, and then provides a recommended search path list related to the requested search word requested by the user and the advertising attribute that the user values. Through the recommended search path list, it is possible to easily select related search words that are most related to the user's requested search word and have high advertising efficiency from among the search words located on the path to reach the user's requested search word, and use them for marketing.

[0119] In addition, the present invention provides an optimal search path corresponding to a user's requested search word in the form of a UI or API, thereby enabling the user to easily obtain a list of search words suitable for advertising / marketing, search information for the search words (search volume, advertising cost, advertising competitiveness, etc.), and function type information.

[0120] Furthermore, the present invention goes beyond simply providing related search words through a word network UI generated based on a search path, and by expressing the connection relationships between search words, can provide a way to grasp the main search journey of users for products and services, as well as assist users in more intuitively grasping their needs, such as what they are requesting and what they are comparing.

[0121] Meanwhile, the service providing device according to the present invention can provide a visualized search word network that helps a user easily sort and select search words corresponding to the search conditions desired by the user from other search words based on the search path information generated as described above, which will be described in detail with reference to the following drawings.

[0122] First, as shown in FIG. 1, the control unit (10) can be configured to include a search path management unit (110) including a search word expansion unit (111), a search information extraction unit (112), a search path extraction unit (113) and an optimal path recommendation unit (114), and a search result providing unit (120) that generates and provides search word network information related to a search word network based on the search path generated via the search path management unit (110).

[0123] As described above, the search path management unit (110) can set each of a plurality of search words as a node based on a plurality of relationship information stored in the search word relationship DB, generate search path information for one or more search paths connecting a start node and an end node determined by a preset algorithm, and store the search path information in the search path DB.

[0124] Here, the search path information may include a plurality of nodes connected via connecting lines, a plurality of search words set to correspond to each of the plurality of nodes, and connecting lines (edges) connecting adjacent nodes.

[0125] In addition, when the search result providing unit (120) receives search word network request information including a requested search word and search conditions from a user terminal, it can extract one or more search path information including the requested search word from the search path DB, and generate search network information in a search word network that connects different search words that have a distance based on the extracted one or more search path information.

[0126] In addition, the search result providing unit (120) may cluster the search word network information into a plurality of different search word groups using a preset clustering algorithm, and then set different attributes so that the plurality of different search word groups can be visually separated from each other, and transmit the generated search word network information to the user terminal.

[0127] As an example, the search result providing unit (120) may receive search word network request information from the user terminal, the search word network request information including search conditions in which a connection direction and a number of hops with respect to other nodes are set based on a requested search word and a node corresponding to the requested search word.

[0128] Here, the connection direction may refer to the connection direction of a connection line (edge) that constitutes a search path and connects a specific node corresponding to the requested search word to other nodes, and at least one of a forward direction and a reverse direction may be selected as the connection direction.

[0129] Also, a hop may refer to the number of nodes that are successively connected to a specific node corresponding to the requested search word to form a path.

[0130] Accordingly, when the search result providing unit (120) receives a search condition in which both the forward and reverse directions are selected as the connection direction and the number of hops is set to 2 hops, it can extract search path information from the search path DB, which includes the requested search word and is composed of a plurality of nodes connected within 2 hops forward or backward based on a node corresponding to the requested search word, from one or more path information stored in the search path DB.

[0131] Here, the search result providing unit (120) can identify search path information from the search path DB, among the one or more search path information, that includes the requested search word and includes nodes that are more than two hops away based on the requested search word, and extract the identified search path information from the search path DB after removing nodes that are more than two hops away based on the requested search word from the identified search path information.

[0132] Meanwhile, the search result providing unit (120) extracts one or more search path information corresponding to the search word network request information from the search path DB, generates a search word network based on the extracted one or more search path information, clusters the search word network into a plurality of different search word groups according to a preset clustering algorithm, and sets different attributes so that the plurality of different search word groups can be visually separated from each other, and provides the generated search word network to the user terminal.

[0133] In addition, the search result providing unit (120) can generate a search interface including the search word network and transmit search interface information in the search interface to the user terminal.

[0134] The search word network described in the present invention may be search word network information, and a search interface including the search word network may be configured in the form of a UI. Also, the search interface including the search word network may be configured as search word network information. Also, the search interface may refer to a search word network (search word network information) configured in the form of a UI.

[0135] An example of the above will now be described with reference to the following drawings. First, as shown in FIG. 11, the search result providing unit (120) can provide the user terminal with search interface related data capable of generating search word network request information for a search word network request, and can receive the search word network request information generated based on the search interface from the user terminal.

[0136] In addition, when the search result providing unit (120) receives search word network request information including “men’s slacks” as a requested search word, including search conditions in which both forward and reverse directions are set as the connection directions and the number of hops is set to 2, it can extract one or more search path information that satisfies the requested search word and search conditions according to the search word network request information from the search path DB, as shown in the figure.

[0137] Thereafter, as shown in FIG. 12, the search result providing unit (120) extracts the requested search word and one or more search path information that meets the search conditions, and then applies the extracted one or more search path information to the ForceAtlas2 algorithm to generate a search word network that connects (with connecting lines) different search words that have a distance between them using the relationship between the search words according to the extracted one or more search path information, and applies the search word network to the Louvain algorithm to generate a search word network that is clustered into multiple different search word groups.

[0138] In addition, the search result providing unit (120) can generate a search word network based on a plurality of search words included in search path information and distances between the plurality of search words, and can generate a search word network by merging the same search words between a plurality of different search path information into one node.

[0139] In addition, the search result providing unit (120) may cluster the search words by grouping search word groups with high connection density together using a community detection method based on the Louvain algorithm in the search word network.

[0140] In addition, the search result providing unit (120) may assign a unique number and a unique color to each of the different search word groups to easily distinguish between the different clustered search word groups, and generate a search word network in which different attributes are set between the search word groups.

[0141] That is, the search result providing unit 120 may set numbers or colors differently between different search word groups to generate a search word network in which different attributes are set between the groups.

[0142] In addition, the search result providing unit (120) can, in conjunction with the search path management unit (110), obtain the search volume for each of the plurality of search words included in the search word network from search volume information stored in a search volume DB, or from a search engine or an external server, and check the search volume corresponding to each of the plurality of search words included in the search word network, and can determine the size of a node corresponding to a search word according to the search volume of the search word, and can vary (adjust) the size of the node included in the search word network.

[0143] In addition, the search result providing unit (120) may check one or more peripheral nodes connected to a specific node corresponding to a search word for each search word included in the search word network, and then calculate a network centrality index of the search word corresponding to the specific node considering the number of the peripheral nodes, the centrality of the specific node according to the position of each of the peripheral nodes, and the distance between the specific node and the peripheral nodes. In addition, the search result providing unit (120) may generate a search word network configured to separate search word nodes having a network centrality index equal to or greater than a preset reference value from other nodes having a network centrality index less than the reference value according to the network centrality index calculated for each search word.

[0144] In addition, network centrality is a measure of how close each node in a network is to the center of the network, and can be calculated as an index indicating the importance of each node in the network function.

[0145] In addition, the network centrality index may be calculated using at least one of a number of calculation methods commonly used to calculate the network centrality index, such as degree centrality, betweenness centrality, closeness centrality, PageRank, etc.

[0146] Thereby, the search result providing unit (120) can utilize the network centrality index for the purpose of easily finding important search words in a visualized search word network by distinctively expressing nodes corresponding to search words with high centrality on the search word network.

[0147] As an example, the search result providing unit (120) may set important nodes corresponding to search words whose network centrality index is equal to or greater than a preset reference value in the search word network in the shape of a doughnut, and may set nodes having a network centrality index less than the reference value in a shape different from the shape of the important nodes, such as a general circle or square.

[0148] Accordingly, the search result providing unit 120 can provide a search word network so that the user can easily sort out influential search words on the search word network by checking the shape corresponding to important nodes and the size of each node through node setting according to search volume and network centrality index.

[0149] In particular, the search word network shown in FIG. 13 is a diagram illustrating a search word network generated for a requested search word different from that shown in FIG. 12. As shown in the figure, the search result providing unit 120 can classify related search words of "Tiffany" of "Girls' Generation" and related search words of "Tiffany" which is a jewelry brand into completely different groups according to distance, and can accurately select other search words which are highly related to the requested search word while belonging to the same category as the requested search word of the category desired by the user, and provide them by grouping them with the requested search word. In addition, by setting nodes according to search volume and network centrality index, a search word network can be provided so that the user can easily classify influential search words from the main search word and peripheral search words which are highly related to the main search word based on the shape corresponding to the important node and the size of each node.

[0150] In addition, the search result providing unit (120) may transmit the search word network generated in response to the search word network request information received from the user terminal as described above to the user terminal, and may generate a search interface including the search word network in the form of a web page and then transmit it to the user terminal, for example.

[0151] According to the above-mentioned configuration, as shown in FIG. 12, the service providing device (100) according to the present invention provides a search word network to a user, which allows the user to easily sort and identify search word groups consisting of one or more other search words having a certain level or higher of relevance to the requested search word requested by the user by grouping multiple search words having high interrelationships based on the distance between the search words according to the search path information, thereby enabling the user to easily identify search word groups consisting of one or more other search words having a certain level or higher of relevance to the requested search word through the search word network. Thus, the search words having a certain level or higher of search volume among the search words having high relevance to the requested search word can be easily identified through the search word network, and the search words having high marketing efficiency but low price competition can be easily found from among the search words located around the requested search word.

[0152] Meanwhile, in the above-mentioned configuration, the search result providing unit (120) can set various additional information to the search word network according to additional search conditions (or search setting conditions) set by the user through a search interface including the search word network, and transmit the set information to the user terminal, which will be described in detail below.

[0153] First, when gender is selected as the additional search criterion through the search interface, the search result providing unit 120 may receive additional search request information (or additional search criterion information) for the additional search criterion including the gender from the user terminal.

[0154] Accordingly, when the search result providing unit (120) receives the additional search request information, it checks the gender set as the additional search condition, communicates with an external server providing the search engine, checks the interest level of men and women, such as the search volume and conversion rate, for each of the multiple search words constituting the search word network, sets the gender with the higher interest level among men and women to a node corresponding to the search word, creates a search word network in which the gender with the highest interest level is set for each search word (or node), and transmits it to the user terminal.

[0155] In addition, when the search result providing unit (120) receives additional search request information for the additional search criteria in which an age group is set, it checks the age group set as the additional search criteria, communicates with an external server providing the search engine, checks the interest level, such as search volume, conversion rate, etc., of each of the multiple different age groups pre-set for the multiple search words constituting the search word network, sets the age group with the highest interest level to a node corresponding to the search word, and generates a search word network in which the age group with the highest interest level is set for each search word (or node), and transmits the search word network to the user terminal.

[0156] Also, the search result providing unit (120) can generate a web page including a search word network set for the age group or gender, and transmit the web page to the user terminal.

[0157] In addition, the additional search criteria may be included in the search word network request information, and when the search result providing unit (120) transmits a search word network corresponding to the search word network request information, it may set the gender and age group for each node included in the search word network and transmit it to the user terminal.

[0158] Meanwhile, when the search result providing unit (120) receives a request related to the additional search conditions in the search intent for each search word from the user terminal, it can check the search intent for each search word, and generate and provide the search word network in which the search intent for each search word is set, which will be described in detail below.

[0159] First, when a user selects a search intent as an additional search criterion through a search interface, the search result providing unit 120 may receive additional search request information in which the search intent is set as an additional search criterion from a user terminal.

[0160] In addition, as shown in FIG. 14 and FIG. 15, when the search result providing unit (120) receives additional search request information in which the search intention is set as an additional search condition, the search result providing unit (120) applies search words belonging to the search word network as input search words to the search engine, and checks whether one or more predetermined search result items (each of the search result items) exist from the top N search results based on a plurality of predetermined search intention judgment conditions for grasping the search intention in one or more returned search result pages, identifies one or more stage-based search intention judgment conditions satisfied by the one or more search result pages from the plurality of stage-based search intention judgment conditions, and sets a predetermined search intention to the search word corresponding to the identified one or more stage-based search intention judgment conditions (one or more stage-based search intention judgment conditions satisfied by the one or more search result pages) from among the plurality of predetermined search intentions, and includes the search intention in the search word network. In this manner, the search intention can be set for each search word belonging to the search word network.

[0161] In addition, the multiple search intents are defined as four types: information seeking, navigational, commercial, and transactional.

[0162] The search result page may be a web page including search results obtained by the search engine based on the input search word, and the search result page may be the search result list information described above.

[0163] Also, as shown in FIG. 16 and FIG. 17, the search result providing unit (120) can check search results for different regions included in the search result page through region dividing lines provided by the search engine included in the search result page, and can identify the top N search results in order of arrangement for each of the regions corresponding to the plurality of search results.

[0164] In addition, the search result providing unit (120) may set search intent only for search words whose search volume exceeds 0 among search words belonging to a search word network.

[0165] In addition, since most clicks occur within two pages of search result pages, the search result providing unit (120) can check the search intent only for the search result pages within two pages returned from the search engine corresponding to the search word as the target for determining the search intent, and set it as the search word.

[0166] In addition, the one or more search result items set as the search intent judgment criteria may include at least one of a knowledge encyclopedia page, a featured snippet, a knowledge panel, a video, an academic information search, a news article, a local map, a site link, an ad carousel, a keyword ad, and a pre-stored conversion page URL (or a conversion page).

[0167] The search result item may also refer to a neighborhood type of the neighborhood corresponding to the search result.

[0168] As an example of identifying the search result items of the search results, the search result providing unit (120) may identify the search results included in the search result page, whose URL includes a pre-set knowledge encyclopedia (Wikipedia (registered trademark), Namwiki) domain, as a search result item corresponding to a "knowledge encyclopedia page."

[0169] Additionally, the search result providing unit (120) may identify a search result item that includes location information (map, contact information, etc.) of a particular location as a "local map." Additionally, the search result provider (120) may identify search result items that contain information on a particular entity in the form of information boxes as "knowledge panels."

[0170] In addition, the search result providing unit 120 may identify search result items that include results that can be moved as video information as "videos." In addition, the search result providing unit (120) can identify a search result item including results that can be moved as research papers and academic resource information as "academic information search."

[0171] Additionally, the search result providing unit (120) can identify search result items that include major news as "news articles." In addition, the search result providing unit (120) can identify a search result item of a search result that includes "highlighted snippet information" as a "highlighted snippet."

[0172] In addition, the search result providing unit 120 can identify a search result item that includes a link result that can move to a specific site as a "site link." In addition, the search result providing unit 120 may identify search result items including phrases related to 'promotion', 'advertising', etc., titles, descriptions, link URLs, etc. as 'keyword advertisements'.

[0173] In addition, the search result providing unit (120) may identify a search result item that includes phrases related to “promotion” or “advertisement”, product photos, link URLs, and other advertising content in the form of a carousel as an “advertising carousel”. In addition to the above-mentioned search result items, the search result providing unit 120 can identify search result items in the form of a title, description, and link URL related to a web / blog, etc., as a 'basic type'.

[0174] In addition, the search result providing unit (120) communicates (links) with an external server that provides the search engine, and extracts only the protocol, domain, and path from the URL of the "keyword advertisement" area of ​​all search result pages held during a unit period (or at a predetermined cycle), and stores the extracted URL as a "conversion URL" in a separate conversion page DB included in the service providing device (100).

[0175] In addition, the search result providing unit (120) extracts the conversion URL only from the search result page within a past period that is preset based on the current time point, and stores it in the conversion page DB, and the conversion URL extracted from the search result page that does not belong to the past period can be deleted from the conversion page DB.

[0176] In addition, the search result providing unit (120) can extract the protocol, domain, and path of the URL in the search result corresponding to the basic type from among the search results (etc.) included in the search result page, and compare them with the "conversion URL," thereby identifying the search result item of the search result corresponding to the basic type that includes the same URL as the conversion URL as the "conversion page."

[0177] As an example of the above, the search result providing unit (120) may perform a process of determining whether a preset first-stage search intent determination condition is satisfied based on whether a knowledge encyclopedia page such as Wikipedia (registered trademark) or Namwiki, a featured snippet, a knowledge panel, a video, an academic information search, or a news article is present among the top three results on one or more search result pages corresponding to a specific search word with a search volume exceeding 0, and then, if the first-stage search intent determination condition is satisfied, a process of determining whether a preset second-stage search intent determination condition is satisfied based on whether a local map or a site link is present among the top three search results.

[0178] In addition, the search result providing unit (120) may determine (identify) the search intent corresponding to the specific search word as informational and navigational if the second stage search intent determination condition is satisfied.

[0179] In addition, when the second-stage search intent judgment condition is not satisfied, the search result providing unit (120) may determine whether a preset third-stage search intent judgment condition is satisfied based on whether one of the following is satisfied: whether a conversion page is present in the top 20 search results, whether a keyword advertisement is present in the top 10 search results, or whether an advertisement carousel is present in the top 10 search results, for one or more search result pages.

[0180] In addition, the search result providing unit (120) may determine (identify) the search intent corresponding to the specific search word as informational if the third stage search intent judgment condition is not satisfied, and may grasp the search intent of the specific search word by determining whether or not one or more subsequent stages of search intent judgment conditions are satisfied if the third stage search intent judgment condition is satisfied.

[0181] In addition, the search result providing unit (120) may set a search intention for each search word belonging to a search word network in the above-mentioned manner, and may transmit the search word network in which the search intention for each search word is set to the user terminal as a response to additional search request information in which the search intention is set as an additional search condition.

[0182] In addition, the search result providing unit (120) may provide the user terminal with a search interface including a search word network in which a search intent for each search word is set, thereby allowing one or more search words corresponding to a specific search intent selected by a user through the search interface in the user terminal to be displayed separately from the specific search word and search words in which other search intents are set.

[0183] For example, as shown in FIG. 18, the search result providing unit (120) may generate a search word network in which a plurality of different colors corresponding to a plurality of search intentions are pre-defined, and the colors of the nodes between a plurality of search words having different search intentions are displayed in different colors, and transmit the search word network to the user terminal.

[0184] Thereby, the search result providing unit 120 can generate a search interface including a search word network that allows the user to check the search intent corresponding to the search word through the color set to the node of the search word, and transmit the generated search interface to the user terminal.

[0185] Meanwhile, when the search result providing unit (120) selects a node corresponding to a specific search word, it can provide the user terminal with a search word network that provides a path starting or ending at the node. As an example, when a user selects a search word node corresponding to “Concept One Slim Fit” via the search interface and then selects the selected node as an end point, the search result providing unit (120) can provide the user terminal with a search word network that provides a starting search word list with “Concept One Slim Fit” as an end point.

[0186] In addition, when the search result providing unit (120) selects an item (slacks brand recommendation → concept one slim fit) with “concept one slim fit” as the end point and “slacks brand recommendation” as the start point from the starting point search word list, it can generate a search word network to visualize and provide a search path starting from “slacks brand recommendation” and ending with “concept one slim fit” using a shortest distance algorithm, and provide the generated search result to the user terminal.

[0187] As described above, the present invention can obtain the relationship between search words as distance according to the search intent of the input search word and the exposure rank of the subsequent search word corresponding to the input search word using a search engine, and based on that, generate and provide a search word network in which a plurality of search words having high interrelationships are grouped and visualized based on the distance between the search words according to the search path information obtained for a plurality of different search words. It can provide search words, etc. that have a certain level or more of relevance to the requested search word requested by the user and a certain level or more of search volume through the search word network, and can easily find search words, etc. that have high marketing efficiency but low price competition from among search words, etc. located around the requested search word confirmed through the search word network, thereby greatly improving user satisfaction and usability.

[0188] In addition, the present invention can set and provide the most likely search intention among users, etc. when using a search word belonging to a search word network in a search word network, thereby supporting a seller who is trying to conduct marketing to easily find a search word that has a search intention that matches the business purpose of the seller and is efficient for marketing through the search word network.

[0189] Furthermore, the present invention not only provides a way to grasp the main search journey of users, etc. for products and services by expressing the connection relationships between search words beyond the level of simply providing related search words through a search word network generated based on a search path, but also helps users to more intuitively grasp their needs, such as what they are requesting and what they are comparing.

[0190] FIG. 19 is a flowchart of a service providing method for providing a search word network based on a search path in a service providing device 100 according to an embodiment of the present invention.

[0191] The service providing device (100) automatically generates one or more basic search words by combining the keyword and the characters by adding a series of characters to the keyword in different ways, and can obtain one or more related search words related to the basic search word through a preset search engine (S1).

[0192] In addition, the service providing device (100) applies the related search word to the search engine as an input search word, and based on the returned search result list information, the search engine grasps the search intent of the input search word and extracts one or more subsequent search words from each of one or more response results corresponding to the function type of one or more response functions used by the search engine, and then generates relationship information on the distance between the input search word and the subsequent search word for each of one or more subsequent search words in a distance calculation method that calculates a weighted value in the connection relationship between the input search word and the subsequent search word according to the exposure rank of the subsequent search word in the response result corresponding to the subsequent search word and the function type of the response result as a distance, and stores the relationship information in a search word relationship DB (102) (S2).

[0193] In addition, the service providing device (100) may generate a directed weighted graph by setting each of a plurality of search words as a node based on a plurality of relationship information stored in the search word relationship DB (102) and setting the distance between the nodes, generate search path information for one or more search paths connecting a start node and an end node determined by a preset algorithm in the directed weighted graph, generate statistical information by averaging statistics for preset items for a plurality of search words included in the search path information, and then match the statistical information with the search path information and store it in the search path DB (104) (S3).

[0194] In addition, when the service providing device (100) receives search word network request information including the requested search words and search conditions from the user terminal (S4), it extracts one or more search path information including the requested search words and satisfying the search conditions from the search path DB, generates a search word network based on the extracted one or more search path information, clusters the search word network into a plurality of different search word groups using a preset clustering algorithm, and sets different attributes so that the plurality of different search word groups can be visually separated from each other, and provides the generated search word network to the user terminal (S5).

[0195] The components described in the embodiments of the present invention may be embodied using one or more general-purpose or special-purpose computers, such as hardware including a storage unit such as a memory, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a Field Programmable Gate Array (FPGA), a programmable logic unit (PLU), a microprocessor, software including a command set, or a combination thereof, or any other device capable of executing and responding to instructions.

[0196] The above contents should be modifiable and changed by a person having ordinary skill in the art to which the present invention belongs, without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are for illustration purposes, not for limiting the technical phenomena of the present invention, and the scope of the technical phenomena of the present invention is not limited by such embodiments. The scope of protection of the present invention should be interpreted according to the following claims, and all technical phenomena within the scope equivalent thereto should be interpreted as being included in the scope of the present invention. [Explanation of symbols]

[0197] 100: Service providing device 10: Control unit 20: Communications Department 30: Preservation Department 101:Keyword DB 102: Search word related database 103: Function type DB 104: Search path DB 105: Search volume DB 110: Search path management section 111: Search word extension 112: Search information extraction unit 113: Search path extraction part 114: Optimal path recommendation section 120: Search results provider

Claims

1. a search path management unit which automatically generates one or more basic search words by combining a keyword and characters by adding a series of characters to a keyword in a different way, obtains one or more related search words related to the basic search word through a preset search engine, applies the related search words to the search engine as input search words, extracts one or more subsequent search words based on returned search result list information, and calculates, as a distance, a weighted value in a connection relationship between the input search word and the subsequent search word according to an exposure rank of the subsequent search word in a response result according to a response function corresponding to the subsequent search word among response functions used by the search engine to grasp a search intention of the input search word and a function type of the response result, generates relationship information in a distance with respect to the input search word for each of one or more subsequent search words and stores it in a search word relationship DB, and generates search path information for each of one or more search paths connecting a start node and an end node determined by a preset algorithm by setting each of a plurality of search words as a node based on the plurality of relationship information stored in the search word relationship DB, and stores the search path information in a search path DB; and When receiving search word network request information including a requested search word and a search condition from a user terminal, the search word network request information includes the requested search word and a search condition, and extracts one or more search path information including the requested search word and satisfying the search condition from the search path DB, generates a search word network connecting different search words having a distance based on the extracted one or more search path information, clusters the search word network into a plurality of different search word groups according to a preset clustering algorithm, and sets different attributes so that the plurality of different search word groups can be visually separated from each other, and provides the generated search word network information to the user terminal. A service providing device for providing a search word network based on a search path.

2. The search path management unit a search word expansion unit that automatically generates one or more basic search words by combining the keyword and the characters by adding a series of characters to the keyword in different ways, and acquires one or more related search words related to the basic search word through a preset search engine; a search information extraction unit that applies the related search words to the search engine as input search words, extracts one or more subsequent search words from one or more response results corresponding to the function types of one or more response functions used by the search engine based on returned search result list information, and calculates a weighted value in a connection relationship between the input search word and the subsequent search word according to an exposure rank of the subsequent search word in the response result corresponding to the subsequent search word and a function type of the response result as a distance, and generates relationship information in a distance between the input search word and the subsequent search word for each of the one or more subsequent search words and stores the relationship information in a search word relationship DB; and and a search path extracting unit for generating a directed weighted graph in which a distance between nodes is set by setting each of a plurality of search words as a node based on a plurality of relationship information stored in the search word relationship DB, and then generating search path information for one or more search paths connecting a start node and an end node determined by a preset algorithm in the directed weighted graph, and storing the search path information in the search path DB.

2. A service providing apparatus for providing a search word network based on a search path according to claim 1.

3. The search information extraction unit a weighted value is calculated by multiplying a preset priority level of a function type corresponding to the subsequent search word by an exposure rank of the subsequent search word in a response result corresponding to the subsequent search word, and the calculated weighted value is set as the distance between the subsequent search word and the input search word.

3. A service providing apparatus for providing a search word network based on a search path according to claim 2.

4. The search information extraction unit The subsequent search word is applied to the search engine as another input search word, and based on the returned search result list information, the search engine grasps the search intent of the input search word according to the distance calculation method, and extracts one or more texts from one or more response results corresponding to the function types of one or more response functions used by the search engine as additional subsequent search words corresponding to the subsequent search word, respectively. Then, a weight value in a connection relationship between the input search word and the additional subsequent search word according to an exposure rank of the additional subsequent search word in the response result corresponding to the additional subsequent search word and a function type of the response result is calculated as a distance, and relationship information in a distance between the one or more additional subsequent search words and the input search word that is the subsequent search word is generated for each of the one or more additional subsequent search words, and is stored in the search word relationship DB.

3. A service providing apparatus for providing a search word network based on a search path according to claim 2.

5. The search conditions include a connection direction and a number of hops with respect to other nodes based on the node corresponding to the requested search word, The search result providing unit extracts one or more search path information satisfying the requested search word and a search condition, applies the extracted one or more search path information to a ForceAtlas2 algorithm to generate a search word network, and applies the search word network to a Louvain algorithm to generate search word network information in a search word network clustered into a plurality of different search word groups.

2. A service providing apparatus for providing a search word network based on a search path according to claim 1.

6. The search path management unit acquires and stores search volume for each search word through an external server or the search engine, The search result providing unit checks the search volume for each search word belonging to the search word network, generates a search word network in which a size of a node corresponding to the search word is adjusted according to the search volume of the search word, checks one or more peripheral nodes connected to a specific node corresponding to the search word for each search word included in the search word network, calculates a network centrality index of the search word corresponding to the specific node considering the number of the peripheral nodes and the centrality of the specific node according to the position of each peripheral node and the distance between the specific node and the peripheral node, and generates a search word network configured to classify a node of a search word having a network centrality index equal to or greater than a preset reference value from another node having a network centrality index less than the reference value according to the network centrality index calculated for each search word.

2. A service providing apparatus for providing a search word network based on a search path according to claim 1.

7. The search result providing unit is The network centrality index of the search word is calculated using at least one of degree centrality, betweenness centrality, closeness centrality, and PageRank.

7. A service providing apparatus for providing a search word network based on a search path according to claim 6.

8. The search result providing unit is The search word network is applied to the search engine, and based on a plurality of predetermined search intent judgment conditions for grasping search intent in one or more returned search result pages, the search word network is generated by identifying one or more predetermined search intent judgment conditions satisfied by the one or more search result pages from among the plurality of predetermined search intent judgment conditions while checking whether the one or more search result items exist according to the search intent judgment conditions from the top N search results. The search intent corresponding to the identified one or more predetermined search intent judgment conditions from among the plurality of predetermined search intents is set to the search word and included in the search word network.

2. A service providing apparatus for providing a search word network based on a search path according to claim 1.

9. The one or more search result items set as the search intent determination criteria include at least one of a knowledge encyclopedia page, a featured snippet, a knowledge panel, a video, an academic information search, a news article, a local map, a site link, an advertisement carousel, a keyword advertisement, and a URL of a pre-stored conversion page.

9. The service providing apparatus for providing a search path-based search word network according to claim 8.

10. A service providing method for providing a search word network based on a search path by a service providing device, comprising: automatically generating one or more basic search words by combining the keyword and the characters by adding a series of characters to the keyword in different ways, and acquiring one or more related search words related to the basic search words through a preset search engine; applying the related search word to the search engine as an input search word, extracting one or more subsequent search words based on returned search result list information, and then generating relationship information on the distance between the input search word and the subsequent search word for each of the one or more subsequent search words in a distance calculation method in which the search engine grasps the search intent of the input search word and calculates, as a distance, a weighted value in a connection relationship between the input search word and the subsequent search word according to an exposure rank of the subsequent search word in a response result according to a response function corresponding to the subsequent search word among response functions used by the search engine and a type of function of the response result; setting each of the plurality of search words as a node based on the plurality of relation information stored in the search word relation DB, generating search path information for one or more search paths connecting a start node and an end node determined by a preset algorithm, and storing the search path information in the search path DB; and When receiving search word network request information including a requested search word and a search condition from a user terminal, extracting one or more search path information including the requested search word and satisfying the search condition from the search path DB, generating a search word network based on the extracted one or more search path information, clustering the search word network into a plurality of different search word groups according to a preset clustering algorithm, setting different attributes so that the plurality of different search word groups can be visually separated, and providing the generated search word network information to the user terminal. A service providing method for providing a search word network based on a search path.

Images