KR102522954B1 - System and method for providing predictive information by artificial intelligence-based data analysis - Google Patents

Abstract

The present disclosure relates to a system and method for providing predictive information based on artificial intelligence-based data analysis, comprising: a user terminal that transmits at least one input data and receives predictive information in response thereto; and if at least one input data is received, at least one similar keyword for the at least one input data is extracted based on a database pre-established through the artificial intelligence-based first learning model to obtain first analysis information and second analysis information. After generating information, generating third analysis information by extracting at least one related keyword for the at least one input data based on web crawling through the first learning model and the second learning model, the first and a server configured to transmit prediction information including analysis information, the second analysis information, and the third analysis information to the user terminal, wherein the first analysis information includes at least one genetic keyword among the at least one similar keyword. and the second analysis information is information generated using at least one satisfactory keyword that satisfies a preset condition among keywords other than the at least one genetic keyword from the at least one similar keyword. can be

Description

System and method for providing predictive information by artificial intelligence-based data analysis {SYSTEM AND METHOD FOR PROVIDING PREDICTIVE INFORMATION BY ARTIFICIAL INTELLIGENCE-BASED DATA ANALYSIS}

The present disclosure relates to an artificial intelligence-based predictive information providing system and method. More specifically, the present disclosure relates to a system and method for providing predictive information based on artificial intelligence-based data analysis.

Generally, when a user inputs a keyword, all data including the keyword is searched on a database or web and provided.

The user has to check the provided data one by one to directly determine whether or not it corresponds to the data he or she is looking for, as well as to directly classify the data for each category.

Therefore, since the user has to perform cumbersome work to obtain desired data and classify it, and spend a lot of time on the work, time and economic waste inevitably occurs.

Therefore, based on artificial intelligence, by using the input data input from the user, the desired keyword is classified by preset category, provided along with the corresponding evidence data, and the generated information is built and managed on the database. technology needs to be developed.

Korean Registered Patent Publication No. 10-1007613 (registration date: January 05, 2011)

In the embodiments disclosed in the present disclosure, when data on research content or experiments is input, prediction information generated by extracting or predicting keywords from at least one evidence date using the input data, as well as the relationship between each keyword It is to provide a system and method for providing predictive information by artificial intelligence-based data analysis that can output how much significance there is by providing.

The problems to be solved by the present disclosure are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

A system for providing prediction information by artificial intelligence-based data analysis according to an aspect of the present disclosure for achieving the above technical problem includes a user terminal that transmits at least one input data and receives prediction information in response thereto; and if at least one input data is received, at least one similar keyword for the at least one input data is extracted based on a database pre-established through the artificial intelligence-based first learning model to obtain first analysis information and second analysis information. After generating information, generating third analysis information by extracting at least one related keyword for the at least one input data based on web crawling through the first learning model and the second learning model, the first and a server configured to transmit prediction information including analysis information, the second analysis information, and the third analysis information to the user terminal, wherein the first analysis information includes at least one genetic keyword among the at least one similar keyword. and the second analysis information is information generated using at least one satisfactory keyword that satisfies a preset condition among keywords other than the at least one genetic keyword from the at least one similar keyword. can be

In addition, a method for providing predictive information by artificial intelligence-based data analysis according to an aspect of the present disclosure, when at least one input data is received from a user terminal, based on a database previously built through an artificial intelligence-based first learning model extracting at least one similar keyword for the at least one input data; generating first analysis information and second analysis information based on the at least one similar keyword; generating third analysis information by extracting at least one related keyword for the at least one input data based on web crawling through the first learning model and the second learning model; and transmitting prediction information including the first analysis information, the second analysis information, and the third analysis information to the user terminal, wherein the first analysis information includes at least one of the at least one similar keyword. The second analysis information is information generated using a genetic keyword of, and the second analysis information is obtained by using at least one satisfying keyword that satisfies a predetermined condition among keywords excluding the at least one genetic keyword from the at least one similar keyword. It may be generated information.

In addition to this, a computer program stored in a computer readable recording medium for executing a method for implementing the present disclosure may be further provided.

In addition to this, a computer readable recording medium recording a computer program for executing a method for implementing the present disclosure may be further provided.

According to the above-described problem solving means of the present disclosure, when data on research content or experiments is input, prediction information generated by extracting or predicting keywords from at least one evidence data using the input data, as well as each keyword By providing a relationship between them, it is possible to output how much significance there is.

The effects of the present disclosure are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

1 is a diagram showing a network structure of a system for providing prediction information based on artificial intelligence-based data analysis according to an embodiment of the present disclosure.
2 is a flowchart illustrating a method of providing prediction information by artificial intelligence-based data analysis according to an embodiment of the present disclosure.
3 is a diagram showing the configuration of a server in a system for providing prediction information based on artificial intelligence-based data analysis according to an embodiment of the present disclosure.
4 is a diagram showing the configuration of first analysis information, second analysis information, and third analysis information generated based on at least one input data according to an embodiment of the present disclosure;
5 is a flowchart illustrating a method of providing prediction information by artificial intelligence-based data analysis according to an embodiment of the present disclosure.
6 is a flowchart illustrating an operation of generating first analysis information according to a first embodiment of the present disclosure.
7 is a flowchart illustrating an operation of generating second analysis information according to a first embodiment of the present disclosure;
8 is a flowchart illustrating an operation of generating third analysis information according to the first embodiment of the present disclosure.
9 is a flowchart illustrating an operation of generating first analysis information according to a second embodiment of the present disclosure.
10 is a flowchart illustrating an operation of generating second analysis information according to a second embodiment of the present disclosure.
11 is a flowchart illustrating an operation of generating third analysis information according to a second embodiment of the present disclosure.
12 is a flowchart illustrating an operation of generating third analysis information according to a third embodiment of the present disclosure.
13 is a diagram illustrating an example of a relationship diagram generated when a single keyword is received as input data according to the first embodiment of the present disclosure.
14 is a diagram showing an example of a relationship diagram generated when a single keyword is received as input data according to the second embodiment of the present disclosure.
15 is a diagram showing a network structure of an artificial intelligence-based database management system according to an embodiment of the present disclosure.
16 is a flowchart illustrating an artificial intelligence-based database management method according to an embodiment of the present disclosure.
17 is a diagram showing the configuration of an analysis set according to an embodiment of the present disclosure
18 is a diagram illustrating a specific operation of generating prediction information according to whether an open analysis set exists in an artificial intelligence-based database management method according to an embodiment of the present disclosure.
19 is a diagram illustrating a specific operation of generating prediction information according to whether at least one input keyword and at least one analysis keyword included in an open analysis set are identical in an artificial intelligence-based database management method according to an embodiment of the present disclosure.
20 is a diagram showing an example of a relationship diagram showing a relationship between analysis sets each including one single keyword according to an embodiment of the present disclosure;
21 is a diagram showing an example of a relationship diagram showing a relationship between analysis sets each including a plurality of keywords as input data according to an embodiment of the present disclosure;

Advantages and features of the present disclosure, and methods of achieving them, will become clear with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present disclosure is not limited to the embodiments disclosed below and may be implemented in various different forms, only the present embodiments are intended to make the present disclosure complete, and to those skilled in the art to which the present disclosure belongs. It is provided to fully inform the scope of the disclosure, which is only defined by the scope of the claims.

Terminology used herein is for describing the embodiments and is not intended to limit the present disclosure. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase. As used herein, "comprises" and/or "comprising" does not exclude the presence or addition of one or more other elements other than the recited elements. Like reference numerals throughout the specification refer to like elements, and “and/or” includes each and every combination of one or more of the recited elements. Although "first", "second", etc. are used to describe various components, these components are not limited by these terms, of course. These terms are only used to distinguish one component from another. Accordingly, it goes without saying that the first element mentioned below may also be the second element within the technical spirit of the present disclosure.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used with meanings commonly understood by those skilled in the art to which this disclosure belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined.

Like reference numbers designate like elements throughout this disclosure. The present disclosure does not describe all elements of the embodiments, and general content or overlapping content between the embodiments in the technical field to which the present disclosure belongs is omitted. The term "unit" or "module" used in the specification means a hardware component such as software, FPGA or ASIC, and "unit" or "module" performs certain roles. However, "unit" or "module" is not meant to be limited to software or hardware. A “unit” or “module” may be configured to reside in an addressable storage medium and may be configured to reproduce one or more processors. Thus, as an example, a “unit” or “module” may refer to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. Functions provided within components and "units" or "modules" may be combined into smaller numbers of components and "units" or "modules" or may be combined into additional components and "units" or "modules". can be further separated.

Throughout the specification, when a part is said to be "connected" to another part, this includes not only the case of being directly connected but also the case of being indirectly connected, and indirect connection includes being connected through a wireless communication network. do.

In addition, when a certain component is said to "include", this means that it may further include other components without excluding other components unless otherwise stated.

Throughout the specification, when a member is said to be located “on” another member, this includes not only a case where a member is in contact with another member, but also a case where another member exists between the two members.

Terms such as first and second are used to distinguish one component from another, and the components are not limited by the aforementioned terms.

Expressions in the singular number include plural expressions unless the context clearly dictates otherwise.

In each step, the identification code is used for convenience of description, and the identification code does not explain the order of each step, and each step may be performed in a different order from the specified order unless a specific order is clearly described in context. there is.

Terms used in the following description are defined as follows.

In the description of the present disclosure, although limited to a server, it may further include a computer and/or a portable terminal, or may be in the form of any one of them.

Here, the computer may include, for example, a laptop computer, a desktop computer, a laptop computer, a tablet PC, a slate PC, and the like equipped with a web browser.

The server is a server that processes information by communicating with an external device, and may include an application server, a computing server, a database server, a file server, a game server, a mail server, a proxy server, and a web server.

The portable terminal is, for example, a wireless communication device that ensures portability and mobility, and includes a Personal Communication System (PCS), a Global System for Mobile communications (GSM), a Personal Digital Cellular (PDC), a Personal Handyphone System (PHS), and a PDA. (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), WiBro (Wireless Broadband Internet) terminal, smart phone ) and wearable devices such as watches, rings, bracelets, anklets, necklaces, glasses, contact lenses, or head-mounted-devices (HMDs). can include

Hereinafter, the working principle and embodiments of the present disclosure will be described with reference to the accompanying drawings.

1 is a diagram showing a network structure of a system for providing prediction information based on artificial intelligence-based data analysis according to an embodiment of the present disclosure.

Referring to FIG. 1 , a system for providing prediction information based on artificial intelligence-based data analysis (hereinafter referred to as 'system for providing prediction information') 10 may include a server 100 and a user terminal 200.

When the server 100 receives at least one input data from the user terminal 200, the server 100 generates at least one similar keyword for at least one input data based on a database previously constructed using the artificial intelligence-based first learning model. By extracting, first analysis information and second analysis information are generated. In addition, the server 100 extracts at least one related keyword for at least one input data based on web crawling using the first learning model and the second learning model to generate third analysis information, and then first Prediction information is generated using the analysis information to the third analysis information. After that, the server 100 transmits the prediction information to the user terminal 200 . Here, the first learning model may use a natural language processing (NLP) method based on long short-term memory (LSTM) and bidirectional encoder representations form transformers (BERT), and the second learning model is cosine similarity, jacquard similarity , at least one of Manhattan similarity, Euclidean distance, and Mahalanobis distance may be used. Specifically, in generating the first to third analysis information, keywords are extracted through the first learning model and similarities are calculated through the second learning model.

The user terminal 200 is a terminal of a user who wants to receive prediction information, and can request provision of prediction information for the at least one input data by transmitting at least one input data to the server 100 . According to the request, the user terminal 200 receives prediction information from the server 100 .

In this case, although only one user terminal 200 is shown in FIG. 1 , at least one may be used, and the number is not limited. However, in the following description of FIGS. 1 to 14 , for convenience of description, the description will be limited to the case where the server 100 provides prediction information to the user terminal 200 .

2 is a flowchart illustrating a method of providing prediction information by artificial intelligence-based data analysis according to an embodiment of the present disclosure.

Referring to FIG. 2 , the user inputs at least one input data through the user terminal 200 and transmits it to the server 100 (S101).

When at least one input data is received from the user terminal 200, the server 100 generates first to third analysis information by inputting the at least one input data to the artificial intelligence-based first learning model. .

Specifically, the server 100 generates first analysis information by extracting at least one similar keyword for at least one input data (S103), and then generates second analysis information (S105). Next, the server 100 generates third analysis information by extracting at least one related keyword for at least one input data based on web crawling (S107). However, this is just one embodiment, and steps S103 to S107 may be performed simultaneously and the order in which they are generated may be changed.

Next, the server 100 generates prediction information using the first analysis information, the second analysis information, and the third analysis information generated by steps S103 to S107.

3 is a diagram showing the configuration of a server in a system for providing prediction information based on artificial intelligence-based data analysis according to an embodiment of the present disclosure.

Referring to FIG. 3 , the server 100 according to an embodiment of the present disclosure may include a communication module 110, a storage module 130, and a control module 150.

The communication module 110 may include one or more components enabling communication with an external device, and may include, for example, at least one of a wired communication module, a wireless communication module, and a short-distance communication module. to transmit and receive signals.

Here, the wired communication module includes not only various wired communication modules such as a Local Area Network (LAN) module, a Wide Area Network (WAN) module, or a Value Added Network (VAN) module, but also a Universal Serial Bus (USB) module. Serial Bus), High Definition Multimedia Interface (HDMI), Digital Visual Interface (DVI), recommended standard (RS-232), power line communication, or plain old telephone service (POTS).

In addition to the WiFi module and the WiBro module, wireless communication modules include global system for mobile communication (GSM), code division multiple access (CDMA), wideband code division multiple access (WCDMA), and universal mobile telecommunications system (UMTS). ), Time Division Multiple Access (TDMA), Wireless LAN (WLAN), Digital Living Network Alliance (DLNA), Wireless Broadband (WiBro), World Interoperability for Microwave Access (WiMAX), High Speed Downlink Packet Access (HSDPA), HSUPA ( It may include a wireless communication module supporting various wireless communication schemes such as High Speed Uplink Packet Access), Long Term Evolution (LTE), 4G, 5G, and 6G.

The short-range communication module is for short-range communication, and includes Bluetooth (Bluetooth??), RFID (Radio Frequency Identification), infrared data association (IrDA), UWB (Ultra Wideband), ZigBee, NFC (Near Field Communication), Wi-Fi (Wireless-Fidelity), Wi-Fi Direct, and wireless USB (Wireless Universal Serial Bus) technology may be used to support short-distance communication.

The storage module 130 stores data and/or various information supporting various functions of the server 100 . A plurality of application programs (application programs or applications) running in the server 100, data for operation of the server 100, and commands may be stored. At least some of these application programs may be downloaded from an external server through wireless communication. Meanwhile, the application program is stored in at least one memory provided in the storage module 130, installed on the server 100, and operated (or functions) by at least one processor provided in the control module 150. can be driven to do.

Meanwhile, the at least one memory is a flash memory type, a hard disk type, a multimedia card micro type, or a card type memory (for example, SD or XD memory). , RAM (Random Access Memory, RAM), SRAM (Static Random Access Memory), ROM (Read-Only Memory, ROM), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory) magnetic memory, It may include a storage medium of at least one type of a magnetic disk and an optical disk. In addition, the memory may store information temporarily, permanently or semi-permanently, and may be provided in a built-in or removable type.

The storage module 130 may build and have a database for providing prediction information through data analysis using artificial intelligence.

The control module 150 may include at least one processor, and controls all components in the server 100 to process input or output signals, data, information, etc., or to process commands, algorithms, and the like stored in at least one memory. It can execute various processes by executing application programs, provide prediction information through data analysis using artificial intelligence, or provide or process appropriate information or functions for database management.

Specifically, when at least one input data is received from the user terminal 200, the control module 150 inputs the at least one input data to the artificial intelligence-based first learning model, thereby providing first analysis information to third analysis information. Analysis information is generated to generate prediction information.

An operation of generating the first to third analysis information by the control module 150 will be described in detail with reference to FIGS. 5 to 12 below. Prior to that, the configuration of the first to third analysis information will be described based on FIG. 4 .

4 is a diagram illustrating configurations of first analysis information, second analysis information, and third analysis information generated based on at least one piece of input data according to an embodiment of the present disclosure.

The server 100 extracts at least one similar keyword 111 from a pre-built database using at least one input keyword 110, and also extracts at least one related keyword 112 from the web through web crawling. do.

Thereafter, the server 100 generates a first keyword list 1211 by extracting at least one genetic keyword from at least one similar keyword 111, and extracts at least one genetic keyword from the at least one similar keyword 111. A second keyword list 1221 is created by extracting at least one satisfactory keyword that satisfies a preset condition among the remaining keywords. In this case, the gene keyword may be a gene name, and the predetermined condition may be a specific technical field as a condition set so that a keyword that the user wants to use for data analysis can be extracted. For example, the preset conditions are keywords for the field of biology and may correspond to diseases, drugs, key words, microorganisms, and the like.

That is, the first keyword list 1211 is generated by generating at least one genetic keyword extracted from at least one similar keyword 111 in the form of a list, and the second keyword list 1221 includes at least one similar keyword 111 At least one satisfaction keyword extracted from is created in the form of a list.

Meanwhile, the server 100 extracts at least one genetic keyword from at least one related keyword 112 to generate first prediction data 1231, and extracts at least one genetic keyword from at least one related keyword 112. Second prediction data 1232 is generated by extracting at least one satisfactory keyword that satisfies a preset condition among the remaining keywords.

That is, the first prediction data 1231 is generated by generating at least one genetic keyword extracted from at least one related keyword 112 in the form of a list, and the first prediction data 1232 includes at least one related keyword 112 At least one satisfaction keyword extracted from is created in the form of a list.

Meanwhile, the server 100 may further generate third prediction data 1233 according to the shape of at least one input data 110 and include it in the third analysis information. In this case, the third prediction data 1233 is data generated using at least one similar sentence extracted from the web through web crawling using at least one input data 110 . Specifically, the third prediction data 1233 includes a sentence list in which at least one similar sentence is generated in the form of a list and a new sentence generated by inputting the at least one similar sentence to the third learning model.

Meanwhile, the keywords included in the first keyword list, the second keyword list, the first prediction data, the second prediction data, and the third prediction data, respectively, are sorted based on preset priorities, and the basis corresponding to each keyword Data can be written together. At this time, the evidence data may be various academic papers, books, projects, and the like.

5 is a flowchart illustrating a method of providing predictive information by artificial intelligence-based data analysis according to an embodiment of the present disclosure, which is an operation performed in the server 100 of the predictive information providing system 10.

Referring to FIG. 5 , when at least one input data is received from the user terminal 200 (S210), the server 100 uses the at least one input data based on the database previously built through the first learning model. At least one similar keyword is extracted with (S220).

Next, the server 100 generates first analysis information and second analysis information using at least one similar keyword extracted in step S220 (S230).

Next, the server 100 extracts at least one related keyword from the web through the first learning model using the at least one input data (S240), and uses the extracted at least one related keyword 3 Analysis information is generated (S250).

In this case, the first to third analysis information may be generated through different operations according to the type of at least one input data received in step S210. For example, it can be classified into a case in which at least one input data is a single keyword (first embodiment), a plurality of keywords (second embodiment), and a sentence/paragraph (third embodiment). However, in the case of the third embodiment, that is, when at least one input data is a sentence or paragraph, a single keyword or a plurality of keywords may be extracted and used from the sentence or paragraph.

6 is a flowchart illustrating an operation of generating first analysis information according to a first embodiment of the present disclosure.

Referring to FIG. 6, in the case of the first embodiment in which at least one input data is a single keyword, the server 100 includes a keyword and a single keyword constituting at least one evidence data stored in a database pre-built through the first learning model. The degree of relevance of each keyword is calculated (S2311), and at least one similar keyword having a degree of relevance equal to or higher than a preset first threshold is extracted based on the calculated degree of relevance (S2312).

Next, the server 100 extracts at least one genetic keyword from among the at least one similar keyword extracted in step S2312 through the first learning model (S2313), and uses the extracted at least one genetic keyword A first keyword list is created (S2314).

Next, the server 100 generates first analysis information using the first keyword list (S2315). At this time, the first analysis information may further include, in addition to the first keyword list, a relationship between at least one genetic keyword constituting the first keyword list and a single keyword generated through a third learning model.

7 is a flowchart illustrating an operation of generating second analysis information according to the first embodiment of the present disclosure.

Referring to FIG. 7 , in the case of the first embodiment in which at least one input data is a single keyword, the server 100 uses the first learning model to generate a sequence of at least one similar keyword extracted in step S2312 of FIG. 6 in step S2313. Only the remaining keywords except for the at least one genetic keyword extracted by are extracted (S2321), and the degree of association between the remaining keywords and each set keyword according to a preset condition is calculated (S2322).

Next, the server 100 extracts at least one satisfaction keyword having a degree of relevance greater than or equal to a preset second threshold value based on the degree of relevance calculated in step S2322 (S2323), and the extracted at least one satisfaction keyword A second keyword list is generated using the keywords (S2324).

Next, the server 100 generates second analysis information using the second keyword list (S2325). In this case, the second analysis information may further include, in addition to the second keyword list, a relationship between at least one satisfactory keyword constituting the second keyword list and a single keyword generated through a third learning model.

8 is a flowchart illustrating an operation of generating third analysis information according to the first embodiment of the present disclosure.

Referring to FIG. 8 , in the case of the first embodiment in which at least one input data is a single keyword, the server 100 calculates a similarity between a keyword existing on the web and a single keyword through a second learning model (S2331), At least one related keyword having a similarity within a preset first threshold range is extracted based on each calculated similarity (S2332).

Here, the second learning model may be configured in the form of any one of a single model, a composite model, and a Node2vec model, and the similarity is determined by using at least one of cosine similarity, Jacquard similarity, Manhattan similarity, Euclidean distance, and Mahalanobis distance. can be derived. In addition, any method that can calculate other similarity can be applied, and the type and number are not limited.

Next, the server 100 extracts at least one genetic keyword from among the at least one related keyword extracted in step S2332 through the first learning model (S2333), and uses the extracted at least one genetic keyword First prediction data is generated (S2334).

Next, the server 100 extracts only the keywords other than the at least one genetic keyword extracted in step S2333 from among the at least one related keyword extracted in step S2332 through the first learning model (S2335), and The similarity of each of the remaining keywords and the set keyword according to the preset condition is calculated through the 2 learning model (S2336).

Next, the server 100 extracts at least one predicted keyword having a similarity within a preset second threshold range based on the similarity calculated in step S2336 (S2337), and the extracted at least one predicted keyword Second prediction data is generated using (S2338).

Next, the server 100 generates third analysis information using the first prediction data and the second prediction data (S2339). In this case, each of the first prediction data and the second prediction data may further include each keyword included therein (first prediction data/second prediction data) and a relationship between single keywords.

9 is a flowchart illustrating an operation of generating first analysis information according to a second embodiment of the present disclosure.

Referring to FIG. 9, in the case of the second embodiment in which at least one input data is a plurality of keywords, the server 100 includes at least one keyword and The degree of relevance of each of a plurality of keywords is calculated (S2341), and at least one similar keyword having a degree of relevance equal to or higher than a predetermined first threshold value is extracted based on the degree of relevance of each of the calculated terms (S2342).

Next, the server 100 primarily extracts at least one genetic keyword from among the at least one similar keyword extracted in step S2342 through the first learning model (S2343), and the extracted at least one genetic keyword Among them, at least one first common keyword commonly associated with each of a plurality of keywords is secondarily extracted (S2344).

Next, the server 100 generates a first keyword list using at least one first common keyword extracted in step S2344 through the first learning model (S2345), and uses the first keyword list First analysis information is generated (S2346). At this time, the first analysis information may further include, in addition to the first keyword list, a relationship between at least one genetic keyword constituting the first keyword list and a plurality of keywords generated through a third learning model.

10 is a flowchart illustrating an operation of generating second analysis information according to a second embodiment of the present disclosure.

Referring to FIG. 10 , in the case of the second embodiment in which at least one input data is a plurality of keywords, the server 100 performs step S2343 of at least one similar keyword extracted by step S2342 of FIG. 9 through the first learning model. Only the remaining keywords except for the at least one genetic keyword extracted by are extracted (S2351), and the degree of association between the remaining keywords and each set keyword according to a preset condition is calculated (S2352).

Next, the server 100 primarily extracts at least one satisfaction keyword having a degree of relevance equal to or greater than a preset second threshold based on the degree of relevance calculated in step S2352 through the first learning model (S2353 ), and secondarily extracts at least one second common keyword commonly related to each of a plurality of keywords among the extracted at least one satisfying keyword (S2354).

Next, the server 100 generates a second keyword list using at least one second common keyword extracted in step S2354 through the first learning model (S2355), and uses the second keyword list Second analysis information is generated (S2356). In this case, the second analysis information may further include, in addition to the second keyword list, at least one satisfaction keyword constituting the second keyword list and a relationship between a plurality of keywords generated through the third learning model.

11 is a flowchart illustrating an operation of generating third analysis information according to a second embodiment of the present disclosure.

Referring to FIG. 11, in the case of the second embodiment in which at least one input data is a plurality of keywords, the server 100 calculates the similarity between the keyword existing on the web and each of the plurality of keywords through the second learning model (S2361 ), at least one related keyword having a similarity within a preset first threshold range is extracted based on the calculated similarity (S2362).

Next, the server 100 primarily extracts at least one genetic keyword from among the at least one related keyword extracted in step S2362 through the first learning model (S2363), and the extracted at least one genetic keyword Among them, at least one third common keyword commonly associated with each of a plurality of keywords is secondarily extracted (S2364).

Next, the server 100 generates first prediction data using at least one third common keyword extracted by step S2364 (S2365), and among at least one related keyword extracted by step S2362, S2363 Except for the at least one genetic keyword extracted by the step, only the remaining keywords are extracted (S2366).

Next, the server 100 calculates the similarity of each of the remaining keywords and the set keyword according to the preset condition through the second learning model (S2367), and based on the calculated similarity, within a preset second threshold range. At least one predicted keyword having a similarity of is firstly extracted (S2368).

Next, the server 100 secondarily extracts at least one fourth common keyword commonly associated with each of a plurality of keywords among the at least one predicted keyword extracted in step S2368 through the first learning model (S2369 ), second prediction data is generated using the at least one fourth common keyword (S2370).

Next, the server 100 generates third analysis information using the first prediction data and the second prediction data (S2371). In this case, each of the first prediction data and the second prediction data may further include respective keywords included therein (first prediction data/second prediction data) and a relationship between a plurality of keywords.

12 is a flowchart illustrating an operation of generating third analysis information according to a third embodiment of the present disclosure.

As described above, in the case of the third embodiment in which at least one input data is a sentence or paragraph, after extracting a single keyword or a plurality of keywords from the sentence or paragraph, analysis information can be generated. In this case, the server 100 may extract a single keyword or a plurality of keywords from the sentence or paragraph through the first learning model. That is, in order to extract specific keywords, a text corpus is produced, specific texts are extracted from sentences through natural language processing, and then classified into groups.

Therefore, when a single keyword is extracted, the same operation as in FIG. 8 is performed to generate first prediction data and second prediction data, and when a plurality of keywords are extracted, the same operation as in FIG. 11 is performed to generate the first prediction data. Data and second prediction data are generated.

However, in the case of the third embodiment in which at least one input data is a sentence or a paragraph, an operation of generating third prediction data is additionally performed to generate third analysis information, and FIG. 12 shows the third prediction data It's about creating motion.

Referring to FIG. 12, the server 100 includes a sentence (or sentences included in a paragraph) received as at least one input data through the second learning model and at least one related keyword extracted by step S2362 above. The similarity of each sentence is calculated (S2372), and at least one similar sentence having a similarity within a preset third threshold range is extracted based on the calculated similarity (S2373).

Next, the server 100 generates third prediction data using the extracted at least one similar sentence (S2374), and uses the first prediction data, the second prediction data, and the third prediction data to generate the third prediction data. Analysis information is generated (S2375). At this time, the third prediction data includes a sentence list composed of at least one similar sentence extracted in step S2373 in the form of a list, and a new sentence generated by the third learning model using the extracted at least one similar sentence. do.

Previously, in the first to third embodiments, the first threshold and/or the second threshold used to generate the first analysis information and the second analysis information are the manager of the server 100 or the user terminal. It can be changed and set by the user of (200), and the value may be different or the same, but is not limited thereto. Similarly, the preset first threshold range and/or the second threshold range used to generate the third analysis information in the first to third embodiments are the manager of the server 100 or the user of the user terminal 200. It can be changed and set by , and the value may be different or the same, but is not limited thereto.

On the other hand, the degree of relevance referred to for extracting a keyword may be the number of times that at least one input data is explained on at least one ground data. That is, it can be determined that the higher the number of explanations, the higher the degree of association.

13 is a diagram illustrating an example of a relationship diagram generated when a single keyword is received as input data according to the first embodiment of the present disclosure.

Referring to FIG. 13, in the case of the first embodiment in which at least one input data 110 is a single keyword, the single keyword is displayed as target data, and based on this, first analysis information to third analysis information through the first learning model are displayed. At least one genetic keyword, at least one satisfaction keyword, and at least one prediction data, each extracted when generating analysis information, are displayed as extracted keywords.

13 shows a case where a total of 10 extraction keywords are extracted from the first extraction keyword to the 10th extraction keyword for the target keyword. In this case, the first extraction keyword described together in the target data and at least one source data , the second extraction keyword, the third extraction keyword, and the sixth extraction keyword have a direct relationship, and display a primary relationship with the target data by directly connecting to the target data. On the other hand, although not described together on the target data and at least one source data, the first extraction keyword, the second extraction keyword, the third extraction keyword, and the sixth extraction keyword having a primary relationship and the at least one source data are described together. The fourth extraction keyword, the fifth extraction keyword, the seventh extraction keyword, the eighth extraction keyword, the ninth extraction keyword, and the tenth extraction keyword are connected to the corresponding extraction keyword to display a secondary relationship with the target data.

However, although the second relationship is shown in FIG. 13, it may be extended and displayed from the first to the nth order according to the connection relationship between each keyword.

14 is a diagram illustrating an example of a relationship diagram generated when a single keyword is received as input data according to the second embodiment of the present disclosure.

Referring to FIG. 14, in the case of the second embodiment in which at least one input data 110 is a plurality of keywords, for example, assuming that two keywords are input, the plurality of keywords and the two keywords are respectively At least one genetic keyword, at least one satisfaction keyword and Display at least one predicted data as an extraction keyword, among which at least one first common keyword, second common keyword, third common keyword, and fourth common keyword commonly related to each of the first target keyword and the second target keyword are visually displayed differently.

FIG. 14 shows a case in which a total of 20 extracted keywords are extracted from the first to twentieth extraction keywords for the first target keyword and the second target keyword.

Specifically, the first extraction keyword, the second extraction keyword, the third extraction keyword, and the fourth extraction keyword described together in the first target data and at least one source data have a direct relationship and are directly connected to the first target data. to display the first order relationship with the first target data. In addition, although not described together on the first target data and at least one source data, on the first extraction keyword, the second extraction keyword, the third extraction keyword, and the fourth extraction keyword and at least one source data having a primary relationship The fifth extraction keyword, the sixth extraction keyword, the ninth extraction keyword, the thirteenth extraction keyword, the fifteenth extraction keyword, and the eighteenth extraction keyword described together are connected to the corresponding extraction keyword to obtain a secondary relationship with the first target data. display

On the other hand, the seventh extraction keyword, the eighth extraction keyword, the eleventh extraction keyword, and the twelfth extraction keyword described together in the second target data and at least one source data have a direct relationship, and are directly connected to the second target data The first order relationship with the second target data is displayed. In addition, although not described together on the second target data and at least one source data, the 7th extraction keyword, the 8th extraction keyword, the 11th extraction keyword, and the 12th extraction keyword and at least one source data having a primary relationship The 10th extraction keyword, the 6th extraction keyword, the 9th extraction keyword, the 13th extraction keyword, the 14th extraction keyword, the 16th extraction keyword, the 17th extraction keyword, and the 20th extraction keyword described together are assigned to the corresponding extraction keyword. It is connected to display the secondary relationship with the secondary target data.

At this time, since the thirteenth extraction keyword having a connection relationship in common with the first target data and the second target data exists, it can be visually displayed as shown in A.

Similarly, in FIG. 14, the 2nd order relationship is displayed, but according to the connection relationship between each keyword, the 1st order to the nth order may be extended and displayed.

Meanwhile, even in the case of the third embodiment, since a single keyword or a plurality of keywords are extracted and used from a sentence or paragraph, the relationship diagram may be displayed as shown in FIG. 13 or FIG. 14 .

1 to 14 relate to generating and providing prediction information by analyzing data based on artificial intelligence, and after providing the prediction information to the user terminal 200, a database is built or a previously built database is updated. Analysis sets are created and stored for management.

Hereinafter, an artificial intelligence-based database management system and method will be described with reference to FIGS. 15 to 21.

15 is a diagram showing a network structure of an artificial intelligence-based database management system according to an embodiment of the present disclosure.

Referring to FIG. 15, an artificial intelligence-based database management system (hereinafter, referred to as a 'database management system') 20 according to an embodiment of the present disclosure includes a server 100, a first user terminal 210, and a second A user terminal 220 may be included.

When at least one input data is received from the first user terminal 210, the server 100 checks at least one input keyword from the at least one input data using the artificial intelligence-based fourth learning model, An open analysis set including at least one of the at least one input keyword is identified on the database. That is, at least one analysis set may be stored in a pre-constructed database. Among them, an open analysis set including at least one input keyword as at least one analysis keyword is obtained by checking an analysis keyword of an open analysis set set open among them. Check.

As a result of the check, when the open analysis set is confirmed, the server 100 generates prediction information using the identified open analysis set, and when the open analysis set is not confirmed, at least one input keyword is entered through web crawling. New prediction information is generated by generating analysis data (including all of the first to third analysis information, respective relationship diagrams, etc.) for. A detailed operation of generating the prediction information will be described below with reference to FIGS. 18 and 19.

Meanwhile, the server 100 stores new prediction information and at least one input keyword used to generate the prediction information as an analysis keyword as an analysis set. In this case, when the server 100 stores the analysis set, the corresponding analysis set may be set as an open analysis set or a closed analysis set according to a user's setting.

Meanwhile, the server 100 checks whether merging is possible between open analysis sets stored in a pre-established database or whether merging is possible between an open analysis set and a closed analysis set is possible according to a preset cycle, and when it is confirmed that merging is possible, A merge possibility notification is transmitted to each user terminal. When a merge request is received from each user terminal as a response thereto, the previously constructed database may be updated by merging the two analysis sets. In this case, when it is confirmed that the open analysis set and the closed analysis set can be merged, merge possibility notifications transmitted to the user terminal corresponding to the open analysis set and the user terminal corresponding to the closed analysis set may be different. That is, the mergeable notification includes information on the two analysis sets, respectively. For example, information on the private analysis set may be selectively included in the mergeable notification to be transmitted to the user terminal corresponding to the open analysis set. . In addition, the merge possibility notification may be privately transmitted to the user terminal corresponding to the undisclosed analysis set.

Meanwhile, if the open analysis set identified to generate prediction information according to the request of the first user terminal 210 is an analysis set corresponding to the second user terminal 220, a merge possibility notification is sent to the first user terminal 210. and the second user terminal 220, respectively, and merging between the two analysis sets may be performed based on a response thereto.

On the other hand, the first user terminal 210 and the second user terminal 220 are terminals of users who want to receive prediction information, and each transmits at least one input data to the server 100 to receive prediction information therefor. Of course, when a merge possibility notification for the analysis set including the prediction information is received, it is checked and decision information on whether or not to merge is transmitted. That is, when merging is desired, a merging request may be transmitted to the server 100 so that merging is performed.

Previously, in FIG. 15, only the first user terminal 210 and the second user terminal 220 were shown, but there may be at least one or more, and the number is not limited.

However, although FIG. 15 shows two user terminals for convenience of description, the number is not limited.

16 is a flowchart illustrating an artificial intelligence-based database management method according to an embodiment of the present disclosure.

Referring to FIG. 16 , when at least one input data is received from the first user terminal 210 (S301), the server 100 checks at least one input keyword through the at least one input data (S303). . However, this is one embodiment, and at least one input keyword may be directly received from the first user terminal 210 .

Next, the server 100 includes, as at least one analysis keyword, at least one input keyword identified in step S303 among at least one open analysis set stored in the database previously constructed using the fourth learning model, and An open analysis set is checked (S305), and predictive information is generated using the open analysis set according to the check result, or predictive information is generated using at least one input keyword confirmed in step S303 (S307). . That is, as a result of the confirmation in step S305, if the open analysis set is confirmed, predictive information is generated using this, and if the open analysis set is not confirmed, at least one input data (at least One input keyword) is used to generate prediction information.

Next, the server 100 transmits the prediction information to the first user terminal 210, and generates an analysis set including the prediction information and at least one input keyword corresponding to the prediction information as an analysis keyword. stored in the database (S311). In this case, the analysis set is stored as an open analysis set or a closed analysis set according to the setting of the user using the first user terminal 210 . Accordingly, at least one open analysis set and/or at least one closed analysis set may be stored in the pre-built database.

Next, the server 100 checks whether it is possible to merge between open analysis sets stored in a pre-established database or between an open analysis set and a closed analysis set according to a preset cycle (S313), and as a result of the check, When analysis sets that can be merged are identified, the previously built database is updated by performing merging between the analysis sets that can be merged using the fourth learning model (S321). In this case, the fourth learning model is a model pre-learned for data merging, and can be generated by merging analysis sets that can be merged into one analysis set.

However, if a non-public analysis set exists among analysis sets that can be merged, a merge possibility notification is transmitted to the user terminal corresponding to the non-public analysis set prior to merging, and if merging is permitted, merging is performed at that time.

For example, as a result of the check in step S313, if it is confirmed that the merge between the open analysis set of the first user terminal 210 and the private analysis set of the second user terminal 220 is possible, the analysis set is disclosed when the analysis set is stored. A mergeable notification is not transmitted to the first user terminal 210 that has been allowed and stored as an open analysis set, but a mergeable notification is transmitted only to the second user terminal 220 that has been stored as an undisclosed analysis set (S315).

Next, after the second user terminal 220 confirms the merging possible notification (S317), and when the merging is allowed, it transmits a merging request as a response to the merging possible notification (S319). Thereafter, when a merge request is received from the second user terminal 220, the server 100 performs merging (S321).

That is, steps S315 to S319 of FIG. 16 are performed only when there is an undisclosed analysis set among mergeable analysis sets, and may be omitted if all mergeable analysis sets are open analysis sets.

17 is a diagram showing the configuration of an analysis set according to an embodiment of the present disclosure.

Referring to FIG. 17, an analysis set includes prediction information including first analysis information 121, second analysis information 122, and third analysis information 123, and at least one received to generate the prediction information. The input data of is included as the analysis keyword 110 (at least one input data is stored as an analysis keyword, and its identification number is identically indicated).

18 is a flowchart illustrating a specific operation of generating prediction information according to whether an open analysis set exists in an artificial intelligence-based database management method according to an embodiment of the present disclosure.

Referring to FIG. 18, the server 100 checks the open analysis set including at least one of the at least one input keyword (S410), and if the open analysis set is confirmed as a result of the check, the identified open analysis set is selected. Prediction information is generated by using (S420).

On the other hand, if the open analysis set is unconfirmed, that is, if it is not confirmed, the server 100 generates new prediction information by generating analysis data for at least one input data based on web crawling (S430), and the new prediction Updating the previously built database using the information (S440)

19 is a flowchart illustrating a specific operation of generating prediction information according to whether at least one input keyword and at least one analysis keyword included in an open analysis set are identical in an artificial intelligence-based database management method according to an embodiment of the present disclosure. am.

Referring to FIG. 19, the server 100 checks at least one analysis keyword included in the confirmed open analysis set (S431), and checks whether the at least one input keyword and at least one analysis keyword are identical (S432). ).

As a result of the confirmation, if the analysis keyword included in the open analysis set and at least one input keyword are completely the same, the server 100 generates prediction information by using the analysis data included in the confirmed open analysis set as it is (S433). ).

On the other hand, as a result of the confirmation, if the analysis keyword included in the open analysis set is identical to at least one input keyword only in part, extracting the remaining keywords excluding the keyword included in the analysis keyword as at least one search data (S434), Data is collected using at least one search keyword based on web crawling (S435). In this case, the collected data may be at least one of at least one keyword and sentence/paragraph having a degree of relevance with at least one search keyword.

Next, the server 100 generates an additional analysis set using the data collected in step S435, and merges the confirmed open analysis set and additional analysis set using the fourth learning model (S437). In this way, it is possible to generate prediction information for the two merged sets (S438).

At this time, the server 100 does not simply store the two analysis sets together, but integrates them into one analysis set to expand the analysis data, after integrating the analysis keywords 110, the specific operations of FIGS. 5 to 12 The first analysis information, the second analysis information, and the third analysis information are generated through , and respective relationship diagrams thereof are created and stored. In this case, the first learning model is used to extract keywords, and the second learning model is used to generate new sentences when generating third prediction data.

As described above, each piece of information includes a relationship diagram for the extracted keywords, and the relationship can be formed as positive, negative, neutral, or unknown through the third learning model. To this end, during learning, at least one learning data is directly labeled to input the relationship between keywords in each sentence, and the labeled at least one learning data is learned through a transformer-based third learning model. Predict the relationship between keywords for all sentences. All combinations of the keywords output through the relationship prediction are counted as a number from the learning data, and the similarity between the keywords is calculated based on the counted learning data. Here, the similarity means a correlation between keywords. The closer the similarity score is to 1, the closer the two keywords are to each other, and the closer to 0, the closer the two keywords are to each other. That is, when at least one keyword is input to the third learning model, the keywords are output in order of similarity, and the relationship between the keywords may output a complex relationship from the first to the nth order. Therefore, as the number of keywords input to the third learning model increases, intersectingly related information can be elaborately provided.

20 is a diagram illustrating an example of a relationship diagram showing a relationship between analysis sets each including one single keyword according to an embodiment of the present disclosure.

21 is a diagram illustrating an example of a relationship diagram showing a relationship between analysis sets each including a plurality of keywords as input data according to an embodiment of the present disclosure.

20 and 21 are similar in basic concept to FIGS. 13 and 14 described above, but FIG. 20 visually displays the connection relationship between single keywords respectively included in the two analysis sets when merging the two analysis sets. , FIG. 21 visually displays the connection relationship between a plurality of keywords included in each of the two analysis sets.

Specifically, referring to FIG. 20 , when each analysis keyword included in the first analysis set and the second analysis set is a single keyword, a single keyword in each of the first analysis set and the second analysis set is set to the 1-1 target. The keyword and the 2-1st target keyword are displayed, and the 1-1st target keyword and the 2-1st target keyword are connected with at least one extraction keyword having a 1st to nth order connection relationship.

In the case of FIG. 20, the 8th extraction keyword having a primary connection relationship with the 1-1 target keyword and the 2nd extraction keyword having a primary connection relationship with the 2-1 target keyword are connected through the 9th extraction keyword. Accordingly, the ninth extracted keyword may be a common keyword of the 1-1 target keyword and the 2-1 target keyword, and may be visually displayed differently. In addition, as the 10th extraction keyword having a primary connection relationship with the 1-1 target keyword and the 4th extraction keyword having a primary connection relationship with the 2-1 target keyword are connected through the 14th extraction keyword, 14 The extraction keyword may also be a common keyword of the 1-1 target keyword and the 2-1 target keyword, and this may also be visually displayed differently. In addition, the respective intersection regions can be indicated as A and B.

Meanwhile, referring to FIG. 21 , when each of the analysis keywords included in the first analysis set and the second analysis set is a plurality of keywords, the plurality of keywords in the first analysis set are the 1-1 target keyword and the 1-1 target keyword. 2 target keyword, displaying a plurality of keywords of the second analysis set as the 2-1 target keyword and the 2-2 target keyword, and then the 1-1 target keyword, the 1-2 target keyword, and the second target keyword The -1 target keyword and the 2-2 target keyword are connected with at least one extraction keyword having a 1st to nth order connection relationship.

21, the first extraction keyword having a primary connection relationship with the 1-1 target keyword, the fourth extraction keyword having a primary connection relationship with the 1-2, and the primary connection relationship with the 2-1 target keyword As the 5th extraction keyword having and the 9th extraction keyword having a primary connection relationship with the 2-2 target keyword are all connected through the 10th extraction keyword, the 10th extraction keyword is the 1-1 target keyword, the 1st extraction keyword. It may be a common keyword of the -2 target keyword, the 2-1 target keyword, and the 2-2 target keyword, and it may be visually displayed differently. In addition, each intersection region can be expressed as shown in A.

Meanwhile, the disclosed embodiments may be implemented in the form of a recording medium storing instructions executable by a computer. Instructions may be stored in the form of program codes, and when executed by a processor, create program modules to perform operations of the disclosed embodiments. The recording medium may be implemented as a computer-readable recording medium.

Computer-readable recording media include all types of recording media in which instructions that can be decoded by a computer are stored. For example, there may be read only memory (ROM), random access memory (RAM), magnetic tape, magnetic disk, flash memory, optical data storage device, and the like.

As above, the disclosed embodiments have been described with reference to the accompanying drawings. Those skilled in the art to which the present disclosure pertains will understand that the present disclosure may be implemented in a form different from the disclosed embodiments without changing the technical spirit or essential features of the present disclosure. The disclosed embodiments are illustrative and should not be construed as limiting.

10: prediction information providing system 100: server
200: user terminal 101: communication module
103: storage module 105: control module
110: at least one input data 111: at least one similar keyword
112: at least one related keyword 121: first analysis information
122: second analysis information 123: third analysis information
1211: first keyword list 1221: second keyword list
1231: first prediction data 1232: second prediction data
1233: third prediction data
20: prediction information providing system 100: server
210: first user terminal 220: second user terminal

Claims (10)

In the predictive information providing system by artificial intelligence-based data analysis,
A user terminal that transmits at least one input data and receives prediction information in response thereto; and
When at least one input data is received, at least one similar keyword for the at least one input data is extracted based on a database pre-established through the artificial intelligence-based first learning model, and the first analysis information and the second analysis information are extracted. generating, extracting at least one related keyword and at least one sentence for the at least one input data on the web based on web crawling through the first learning model and the second learning model; After generating third analysis information based on a keyword and the at least one sentence, a server for transmitting prediction information including the first analysis information, the second analysis information, and the third analysis information to the user terminal. contains,
The first analysis information is information generated using at least one genetic keyword among the at least one similar keyword,
The second analysis information is information generated using at least one satisfactory keyword satisfying a predetermined condition among keywords other than the at least one genetic keyword from the at least one similar keyword,
The third analysis information includes first prediction data and second prediction data generated based on at least one related keyword extracted from the web, and generated based on the at least one sentence extracted from the web. Information including third prediction data,
The at least one sentence is extracted through the first learning model of the server, and is a sentence having a similarity with the at least one input data within a preset third threshold range,
The first prediction data is data generated by extracting at least one genetic keyword from among the at least one related keyword through the first learning model,
The second prediction data is data generated by extracting at least one or more satisfactory keywords satisfying the preset condition from among the at least one related keyword through the first learning model;
The third prediction data is data that is further generated only when the at least one input data is a sentence or a paragraph, and the at least one sentence is generated by a third learning model based on a sentence list composed of the at least one sentence and a transformer. Characterized in that it includes a new sentence generated based on,
Predictive information providing system based on artificial intelligence-based data analysis.
According to claim 1,
The at least one input data,
It is in the form of at least one of a single keyword, a plurality of keywords, sentences and paragraphs,
The server,
Characterized in that, when the at least one input data is a sentence or a paragraph, the single keyword or the plurality of keywords are extracted from the sentence or the paragraph through the first learning model and then used.
Predictive information providing system based on artificial intelligence-based data analysis.
According to claim 2,
The server,
When the at least one input data is a single keyword,
Based on the pre-constructed database through the first learning model, the at least one similar keyword having a degree of association with the single keyword and a predetermined first threshold value or higher is extracted, and the at least one similar keyword is extracted from the extracted at least one similar keyword. The first analysis information is generated by extracting at least one genetic keyword and creating a first keyword list, and the at least one having a degree of association of a preset second threshold value or higher with a set keyword according to the preset condition in the remaining keywords. Generating the second analysis information by extracting one satisfactory keyword and generating a second keyword list;
The first prediction data and the second prediction data are obtained by extracting the at least one related keyword having a similarity with the single keyword on the web based on the web crawling through the second learning model and within a predetermined threshold range. Characterized in that the third analysis information is generated by generating each,
Predictive information providing system based on artificial intelligence-based data analysis.
According to claim 2,
The server,
When the at least one input data is a plurality of keywords,
Based on the pre-built database through the first learning model, the at least one similar keyword having a degree of relevance equal to or higher than a predetermined first threshold value is extracted for each of the plurality of keywords, and the at least one similar keyword is extracted from the at least one similar keyword. A first keyword by first extracting at least one genetic keyword, and secondarily extracting at least one first common keyword commonly associated with at least two keywords included in the plurality of keywords among the at least one genetic keyword. Generating the first analysis information by generating a list, and primarily extracting the at least one satisfaction keyword having a degree of association greater than or equal to a preset second threshold value with a set keyword according to the preset condition from the remaining keywords; generating the second analysis information by generating a second keyword list by secondarily extracting at least one second common keyword commonly associated with at least two keywords included in the plurality of keywords among the at least one satisfying keyword; and ,
Based on the web crawling through the second learning model, the at least one related keyword having a similarity within a preset first threshold range is extracted for each of the plurality of keywords on the web, and through the first learning model After the at least one genetic keyword is primarily extracted from the at least one related keyword, the at least one genetic keyword is excluded from the at least one related keyword, and the at least two keywords included in the plurality of keywords are common. At least one third common keyword associated with is secondarily extracted to generate first prediction data, and the at least one genetic keyword is excluded from the at least one related keyword through the second learning model. The at least one prediction keyword having a similarity with a set keyword according to a condition within a predetermined second threshold range is primarily extracted, and included in the plurality of keywords among the at least one prediction keyword through the first learning model. Characterized in that the third analysis information is generated by generating second prediction data by secondarily extracting at least one fourth common keyword commonly associated with at least two keywords,
Predictive information providing system based on artificial intelligence-based data analysis.
delete According to claim 4,
The pre-built database includes at least one source data,
The degree of association is the number of times the at least one input data is explained on the at least one ground data,
The server,
When there are a plurality of keywords respectively included in the first keyword list, the second keyword list, the first prediction data, the second prediction data, and the third prediction data, a plurality of keywords are selected based on preset priorities. Sort, and write together the evidence data corresponding to each of the plurality of keywords,
Characterized in that the evidence data includes at least one of academics, theses, books, and projects,
Predictive information providing system based on artificial intelligence-based data analysis.
According to claim 4,
The first analysis information further includes a relationship between the at least one input data and the at least one genetic keyword extracted from the at least one similar keyword,
The second analysis information further includes a relationship between the at least one input data and the at least one satisfaction keyword extracted from the at least one similar keyword;
The third analysis information includes a relationship between the at least one input data and the at least one genetic keyword extracted from the at least one related keyword,
Characterized in that the relationship diagram is connected and displayed in the first to nth order according to whether the keyword is directly related or indirectly related to each of the at least one input data,
Predictive information providing system based on artificial intelligence-based data analysis.
According to claim 7,
The first learning model is a model learned to extract a keyword or sentence having a degree of association with the at least one input data from the pre-built database,
The second learning model is a model learned to extract a keyword or sentence having a similarity with the at least one input data on the web,
The third learning model generates the relationship diagram when generating the first analysis information, the analysis information, and the third analysis information, and generates the new sentence when generating the third prediction data. Characterized in that it is a model learned for
Predictive information providing system based on artificial intelligence-based data analysis.
According to claim 1,
The server,
Updating the pre-built database through web crawling according to a preset first cycle, generating the prediction information every time the update is performed or every preset second cycle, and transmitting it to the user terminal ,
Predictive information providing system based on artificial intelligence-based data analysis.
In the method of providing predictive information by artificial intelligence-based data analysis, performed by a server of a system for providing predictive information by artificial intelligence-based data analysis,
If at least one input data is received from a user terminal, extracting at least one similar keyword for the at least one input data based on a database previously constructed through a first artificial intelligence-based learning model;
generating first analysis information and second analysis information based on the at least one similar keyword;
At least one related keyword and at least one sentence for the at least one input data are extracted from the web based on web crawling through the first learning model and the second learning model, and the at least one related keyword and the at least one sentence are extracted. generating third analysis information based on one sentence;
Transmitting prediction information including the first analysis information, the second analysis information, and the third analysis information to the user terminal,
The first analysis information is information generated using at least one genetic keyword among the at least one similar keyword,
The second analysis information is information generated using at least one satisfactory keyword satisfying a predetermined condition among keywords other than the at least one genetic keyword from the at least one similar keyword,
The third analysis information includes first prediction data and second prediction data generated based on at least one related keyword extracted from the web, and generated based on the at least one sentence extracted from the web. Information including third prediction data,
The at least one sentence is extracted through the first learning model of the server, and is a sentence having a similarity with the at least one input data within a preset third threshold range,
The first prediction data is data generated by extracting at least one genetic keyword from among the at least one related keyword through the first learning model,
The second prediction data is data generated by extracting at least one or more satisfactory keywords satisfying the preset condition from among the at least one related keyword through the first learning model;
The third prediction data is data that is further generated only when the at least one input data is a sentence or a paragraph, and the at least one sentence is generated by a third learning model based on a sentence list composed of the at least one sentence and a transformer. Characterized in that it includes a new sentence generated based on,
A method for providing predictive information through artificial intelligence-based data analysis.

Images