Technical analysis method and device

Abstract

According to the embodiment of the invention, the technical analysis method and the technical analysis device can be provided, and the technical analysis method comprises the following steps: an information collection step: designing a meta-framework structure aiming at an analysis object keyword, exporting a technical structure based on the meta-framework structure by utilizing a generative AI model, and generating a retrieval keyword based on the exported technical structure, retrieving and acquiring a plurality of literatures based on the retrieval keywords; and a technical classification step of extracting the plurality of literatures according to the similarity between the literatures, executing preprocessing for utilizing the contents of the plurality of literatures, and then executing specific technical classification based on the result of executing the content-based clustering.

Description

Technical Field

[0001] The present disclosure relates to a technology analysis method and a device for processing the technology analysis method. Background Art

[0002] As society's awareness of the importance of intellectual property continues to grow, the need for non-IP experts to search patents is also increasing. Currently, a variety of patent search database services are provided for patent searches.

[0003] However, with the development of technology, technology is subdivided into various categories, and each subdivided technology accumulates a large amount of data. Therefore, manual processing of patent data or technical data consumes a lot of time and manpower.

[0004] In addition, when people extract large amounts of data, they may make subjective judgments or make mistakes, which may lead to reduced consistency of the data or errors. Summary of the Invention

[0005] (1) Technical issues to be resolved

[0006] According to various embodiments, methods and apparatus for performing technical analysis based on a generative AI model may be provided.

[0007] The technical problems to be solved by various embodiments are not limited to the technical problems mentioned. For those skilled in the art, other technical problems not mentioned can be clearly understood from the following description.

[0008] (2) Technical solution

[0009] According to one embodiment, a technology analysis method can be provided, which includes: an information collection step, designing a meta-framework structure for the analysis object keyword, using a generative AI model to derive a technology structure based on the meta-framework structure, generating retrieval keywords based on the derived technology structure, and retrieving and acquiring multiple documents based on the retrieval keywords; a technology classification step, refining the multiple documents according to the similarity between the documents, performing preprocessing for utilizing the content of the multiple documents, and then performing specific technology classification based on the results of performing content-based clustering; a trend analysis step, performing trend analysis based on the collection results and classification results performed in the information collection step and the technology classification step; and a candidate technology mining step, using the trend analysis results to mine promising candidate technologies.

[0010] In the information collection step, the meta-framework structure may be designed, comprising a plurality of meta-framework items derived from public reports of at least one enterprise or at least one institution registered in the technical field corresponding to the analysis object keyword.

[0011] The plurality of meta-framework structures may include at least one meta-framework item among lower-level technologies, technical elements, process elements, application industries, social impacts and related technologies.

[0012] In which, in the information collection step, the generative AI model can be used to generate multiple technical items related to the analysis object keyword in the technical field corresponding to the analysis object keyword, and the technical structure can be derived by associating at least part of the multiple technical items with at least part of the corresponding multiple meta-framework items included in the meta-framework structure.

[0013] In which, in the information collection step, the generative AI model can be used to generate at least one search keyword corresponding to each of the multiple technical items included in the technical structure, and at least a part of the at least one search keyword generated for each of the multiple technical items can be used to generate at least one search formula for each of the multiple technical items.

[0014] In the information collection step, based on the search keyword or the at least one search formula, documents including texts related to the analysis target keyword may be retrieved from a designated database or network.

[0015] In the technical classification step, the extraction, preprocessing and content-based clustering can be performed on at least one acquired document based on a deep learning AI model for each of the multiple technical items included in the technical structure.

[0016] In the technology classification step, at least one specific technology item can be classified for each of the multiple technology items based on the content-based clustering results and the generative AI model, and a technology name, technology summary and main keywords can be generated for each specific technology item in the at least one specific technology item using at least one clustering document classified according to the content-based clustering.

[0017] In the technology classification step, at least one main keyword can be generated from the clustered documents for each specific technology item in the at least one specific technology item based on at least a part of the generative AI and text mining algorithm.

[0018] In the technical classification step, at least a part of the technical name and the technical abstract can be generated for each specific technical item in the at least one specific technical item based on at least one main document and the main keywords identified for each specific technical item in the at least one specific technical item according to the content-based clustering.

[0019] In the technical classification step, the cluster center of each specific technical item in the at least one specific technical item can be determined based on the content-based clustering, and a specified number of clustered documents located at the cluster center can be determined as the main documents of the corresponding specific technical item.

[0020] In which, in the technology classification step, the content-based clustering can be performed by distinguishing the categories of technical documents and network data including patents or papers for at least one acquired document of each of the multiple technical items. Based on the content-based clustering results and the generative AI model, at least one specific technical item can be classified according to the categories of the technical documents and network data. The main keywords can be generated for each specific technical item of the at least one specific technical item using at least one clustered document classified according to the content-based clustering.

[0021] According to another embodiment, a technology analysis device can be provided, which includes: a storage unit storing a generative AI model; an information collection unit designing a meta-framework structure for an analysis object keyword, utilizing the generative AI model to derive a technology structure based on the meta-framework structure, generating retrieval keywords based on the derived technology structure, and retrieving and acquiring multiple documents based on the retrieval keywords; a technology classification unit refining the multiple documents according to the similarity between the documents, performing preprocessing for utilizing the contents of the multiple documents, and then performing specific technology classification based on the results of performing content-based clustering; a trend analysis unit performing trend analysis based on the collection results and classification results performed in the information collection unit and the technology classification unit; and a candidate technology mining unit utilizing the trend analysis results to mine promising candidate technologies.

[0022] The information collection unit may design the meta-framework structure, which includes a plurality of meta-framework items derived from public reports of at least one enterprise or at least one institution registered in the technical field corresponding to the analysis object keyword.

[0023] The plurality of meta-framework structures may include at least one meta-framework item among lower-level technologies, technical elements, process elements, application industries, social impacts and related technologies.

[0024] In which, the information collection unit can use the generative AI model to generate multiple technical items related to the analysis object keyword in the technical field corresponding to the analysis object keyword, and derive the technical structure by associating at least part of the multiple technical items with at least part of the multiple meta-framework items included in the meta-framework structure.

[0025] In which, the information collection unit can use the generative AI model to generate at least one search keyword corresponding to each of the multiple technical items included in the technical structure, and can use at least a part of the at least one search keyword generated for each of the multiple technical items to generate at least one search formula for each of the multiple technical items.

[0026] The information collection unit may search and obtain documents including texts related to the analysis target keyword from a designated database or network based on the search keyword or the at least one search formula.

[0027] In which, the technology classification unit can perform the extraction, preprocessing and content-based clustering on at least one acquired document based on at least one deep learning AI model for each of the multiple technology items included in the technology structure.

[0028] In which, the technology classification unit can classify at least one specific technology item for each of the multiple technology items based on the content-based clustering results and the generative AI model, and can use at least one clustering document classified according to the content-based clustering to generate a technology name, technology summary and main keywords for each specific technology item in the at least one specific technology item.

[0029] The technology classification unit may generate at least one main keyword for each specific technology item in the at least one specific technology item from the clustered documents based on at least a portion of the generative AI and text mining algorithms.

[0030] In which, the technology classification unit can generate at least a part of the technology name and the technology abstract for each specific technology item in the at least one specific technology item based on at least one main document and the main keywords identified for each specific technology item in the at least one specific technology item according to the content-based clustering.

[0031] The technology classification unit may determine the cluster center of each specific technology item in the at least one specific technology item based on the content-based clustering, and determine a specified number of cluster documents located at the cluster center as the main documents of the corresponding specific technology item.

[0032] In which, the technology classification unit can perform the content-based clustering by distinguishing the categories of technical documents and network data including patents or papers for at least one acquired document of each of the multiple technical items, and can classify at least one specific technical item according to the categories of the technical documents and network data based on the content-based clustering results and the generative AI model, and can use at least one clustered document classified according to the content-based clustering to generate a technology name, technical abstract and main keywords for each specific technical item of the at least one specific technical item.

[0033] (3) Beneficial effects

[0034] According to various embodiments, through the technical analysis method and apparatus, the scope of various technical analyses can be expanded while reducing the time required for technical analysis.

[0035] According to various embodiments, the technical analysis method and apparatus can reduce the cost and time required for basic technical analysis based on raw data, thereby allowing data experts to focus on the analysis of advanced information, thereby improving work efficiency.

[0036] According to various embodiments, a technology analysis method and apparatus based on a generative AI model are provided, thereby enabling the mining of analysis indicators and methods reflecting the multifaceted nature of technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a diagram schematically showing the configuration of an apparatus according to one embodiment.

[0038] Figure 2 is a diagram schematically showing a detailed configuration of a processing unit and a storage unit of a device according to one embodiment.

[0039] Figure 3 is a flowchart illustrating an operational flow of performing technical analysis based on an artificial intelligence (AI) model in a device according to one embodiment.

[0040] Figure 4 FIG. 1 is a diagram illustrating an operation flow of performing information collection in a device according to one embodiment.

[0041] Figure 5 FIG. 1 is a diagram illustrating an operation flow of performing technology classification in a device according to one embodiment.

[0042] Figure 6 FIG. 1 is a diagram exemplarily illustrating an operation of generating a search keyword in a device according to an embodiment.

[0043] Figure 7 FIG. 1 is a diagram exemplarily illustrating an operation of classifying a specific technology in a device according to an embodiment.

[0044] Figure 8 FIG. 4 is a diagram exemplarily illustrating an operation of classifying specific technologies based on technical documents in a device according to one embodiment.

[0045] Figure 9 FIG. 1 is a diagram exemplarily illustrating an operation of distinguishing technical literature data from network data to classify specific technologies in a device according to one embodiment.

[0046] Figure 10 FIG. 4 is a diagram exemplarily illustrating operations of performing trend analysis and mining promising candidate technologies in a device according to one embodiment.

[0047] Figure 11 FIG. 4 is a diagram exemplarily illustrating the results of mining promising candidate technologies and the generation operation of a customized report in a device according to one embodiment. DETAILED DESCRIPTION

[0048] The following embodiments will be described in detail with reference to the accompanying drawings. However, various modifications may be made to the embodiments, and the scope of the present patent application is not limited or restricted by these embodiments. It should be understood that all modifications, equivalents, or alternatives to the embodiments are included within the scope of the claims.

[0049] The terms used in the examples are for illustrative purposes only and should not be construed as limiting the present invention. Unless the context clearly indicates otherwise, singular expressions include plural expressions. In this specification, terms such as "including" or "having" are intended to specify the presence of the described features, numbers, steps, operations, components, parts, or combinations thereof, and should not be understood as excluding in advance the presence or additional possibility of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0050] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by those skilled in the art in the art to which the embodiments pertain. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and should not be interpreted in an idealized or overly formal manner unless explicitly defined in this application.

[0051] In addition, when describing with reference to the accompanying drawings, the same components are given the same reference numerals regardless of the figure numbers, and repeated descriptions thereof are omitted. When describing the embodiments, if it is determined that a detailed description of the relevant known technology would unnecessarily obscure the gist of the embodiments, the detailed description thereof will be omitted.

[0052] In addition, when describing the components of the embodiments, terms such as first, second, A, B, (a), and (b) may be used. These terms are only used to distinguish one component from another, and the nature, order, or sequence of these components are not limited by these terms.

[0053] When a component is described as being “connected,” “coupled,” or “accessed” to another component, it should be understood that the component can be directly connected, coupled, or accessed to the other component, but other components may also be “connected,” “coupled,” or “accessed” between the one component and the other component.

[0054] In addition, when a component is described as being “connected” or “accessed” to another component through communication, it should be understood that the component can be connected or accessed to another component through wireless communication or wired communication, but other components can also be “connected” or “accessed” between one component and another component.

[0055] In addition, when describing the components of the embodiments, "or" can refer to each component, two or more components, or all components. For example, "a, b, or c" should be understood to refer to any of "a", "b", "c", "a and b", "a and c", "b and c", or "a, b, and c".

[0056] In other embodiments, the same names are used to describe components included in one embodiment and components having the same functions as those in another embodiment. Unless otherwise specified, the description of one embodiment is applicable to other embodiments, and detailed description is omitted within the scope of repetition.

[0057] The "data" processed by the device and / or the device can be represented by the term "information." Here, "information" is a concept that includes data.

[0058] The present disclosure may provide a method and apparatus for performing technology analysis based on an artificial intelligence (AI) model. More specifically, a method and apparatus may be provided for retrieving and acquiring documents based on a meta-frame structure to perform technology analysis on analysis target keywords, classify the documents based on similarity and clustering, and perform trend analysis and prospective candidate technology mining based on the acquired documents and document classification results.

[0059] Various embodiments of the present disclosure will be described below with reference to the accompanying drawings. The drawings in the specification are intended to further describe the technical ideas of the present disclosure together with the specific implementation methods, and the present disclosure should not be interpreted as being limited to the contents shown in the drawings.

[0060] Figure 1 is a diagram schematically showing the configuration of an apparatus according to one embodiment. Figure 2is a diagram schematically showing a detailed configuration of a processing unit and a storage unit of a device according to one embodiment.

[0061] Reference Figure 1 , the apparatus 100 may include a processing unit 110, a storage unit 120 and a communication unit 130. More specifically, as Figure 2 As shown, the processing unit 110 may include at least one of an information collecting unit 201 , a technology classification unit 203 , a trend analyzing unit 205 , and a candidate technology mining unit 207 .

[0062] The processing unit 110 may include at least one processor (or controller) and may process at least one control instruction for the device 100 according to the embodiment or a configuration connected to the device 100 through at least one program (application, application, tool, plug-in, software, etc.).

[0063] At this time, at least one of the information collecting unit 201 , the technology classifying unit 203 , the trend analyzing unit 205 , and the candidate technology mining unit 207 may be configured by a single processor, or may be configured by independent or separate processors.

[0064] At this time, at least one program may be stored in the storage unit 120 of the device 100 and / or in a storage unit of a device connected to the device 100 .

[0065] According to various embodiments, at least one program may include at least a portion of a generative AI model and a deep learning AI model, and may be configured to perform technical analysis.

[0066] In the following, in various embodiments, the apparatus 100 performing an operation according to a control instruction may be understood as performing a specified operation through at least one program related to at least one control instruction processing of the apparatus 100 .

[0067] Here, although it is described that the control instruction processing is executed by the at least one program installed in the device 100, it is not limited thereto and may be executed by other programs pre-installed in the storage unit 120 or a temporarily installed program.

[0068] According to one embodiment, the control instruction processing may be performed by at least a portion of a database provided by an external device connected to the apparatus 100 for free or for a fee.

[0069] The operation of the device 100 can be performed based on data processing and device control by the processing unit 110 , and the processing unit 110 can also perform designated functions based on control instructions received through the input / output unit and / or communication unit of the device 100 .

[0070] In addition, the processing unit 110 may process data acquired through the communication unit 130 based on the identified user. For example, the processing unit 110 may identify the user based on user information received from at least one other device (e.g., a user device) connected through the communication unit 130 and / or user information acquired through an input unit (or input / output unit) connected to the device 100, and perform an operation according to a control instruction input by the identified user.

[0071] The storage unit 120 may store various data processed by at least one component of the device 100 (e.g., the processing unit 110 or the communication unit 130). The data may include, for example, a program for processing a control instruction, data processed by the program, and / or input data and output data related thereto.

[0072] Storage unit 120 may include artificial intelligence algorithms for processing control instructions, including artificial neural network algorithms, deep learning algorithms, and regression analysis algorithms, as well as at least a portion of the mechanisms, operators, language models, and big data associated therewith. Furthermore, storage unit 120 may include at least one deep learning model (or machine learning model) configured based on at least a portion of the aforementioned artificial intelligence algorithms.

[0073] For example, the generative AI model 221 or the deep learning AI model 223 may include at least a portion of a classification model, a detection model, and a generation model. Here, the classification model may include at least a portion of various algorithms based on transformer neural networks such as BERT (Bidirectional Encoder Representations from Transformers), ALBERT (A LiteBERT), GPT (Generative Pre-trained Transformer), and convolutional neural networks (CNN).

[0074] The detection model may include at least a portion of various algorithms based on Yolo (you only look once), R-CNN (Regions with Convolutional Neural Networks), and Transformer Neural Networks.

[0075] In addition, the generative model may include at least a portion of various algorithms based on GAN (Generative Adversarial Networks) and VAE (Variational Autoencoders) or transformer neural networks.

[0076] like Figure 2 As shown, the storage unit 120 may include a generative AI model 221 and may further include a deep learning AI model 223.

[0077] More specifically, the generative AI model 221 may include at least one of a generative AI model 221 for deriving technical structures related to analysis object keywords based on a meta-framework structure, a generative AI model 221 for generating retrieval keywords for document retrieval based on the derived technical structures, and a generative AI model 221 for classifying specific technical structures of technical structures based on content-based clustering results.

[0078] In addition, the deep learning AI model 223 may include at least one of the following deep learning AI models 223: a deep learning AI model 223 that extracts documents based on their similarity, a deep learning AI model 223 that performs preprocessing for utilizing or clustering document content, and a deep learning AI model 223 that performs clustering on documents.

[0079] The storage unit 120 may include data for confirming and processing control and operation specified by a signal received through an input / output unit (not shown) and / or various devices included in the communication unit 130 .

[0080] The operations described by the storage unit 120 are processed by the processing unit 110 , and data used to process related operations, data being processed, processed data, preset data, etc. can be stored in the storage unit 120 as a database.

[0081] The data stored in the storage unit 120 may be changed, modified, deleted, and / or new data may be generated by the processing unit 110 based on the recognized user input.

[0082] The storage unit 120 may store device setting information of the device 100. The device setting information may be setting information of at least a portion of the device 100 and functions and services provided by the device 100.

[0083] The storage unit 120 may store user information (or user account) of at least one user. The user information may include user identification information (e.g., an identity identification number (ID) and password) and at least a portion of user customization information. The user customization information is setting information for at least a portion of the control permissions and / or functions of the device 100, and may be set and stored based on input from an administrator.

[0084] The storage unit 120 may include a volatile memory, a nonvolatile memory, and / or a computer-readable recording medium. In this case, the computer-readable recording medium may store a computer program for executing the operation of determining whether the battery circuit is abnormal based on the apparatus 100 of various embodiments.

[0085] The communication unit 130 may support establishment of a wired communication channel, establishment of a wireless communication channel, and communication through the established communication channel between the device 100 and at least one other device (eg, a user device or a server).

[0086] During communication, the communication unit 130 may perform modulation / demodulation, encryption / decryption, etc. The communication unit 130 may include one or more communication processors, which may operate independently of or in conjunction with the processing unit 110 and support wireless and / or wired communication.

[0087] According to one embodiment, when the communication unit 130 supports wireless communication, the communication unit 130 may include a wireless communication module, such as at least a portion of a cellular communication module, a short-range communication module, and a GNSS (global navigation satellite system) communication module.

[0088] When the communication unit 130 supports wired communication, the communication unit 130 may include a wired communication module, such as at least a portion of a LAN (local area network) communication module or a power line communication module, a CAN (controller area network) communication module, or an SPI (serial peripheral interface) communication module.

[0089] More specifically, the communication unit 130 can perform wired communication and / or wireless communication with external devices through a short-range communication network such as Bluetooth, BLE (Bluetooth Low Energy), WiFi, WiFi direct, IrDA (infrared data association), ZigBee, UWB, RF (Radio Frequency) and / or a long-range communication network such as a cellular network, the Internet or a computer network (for example, a LAN (local area network) or a WAN (wide area network)).

[0090] Various types of communication modules constituting the communication unit 130 may be integrated into one component (eg, a single chip), or may be implemented by a plurality of separate components (eg, a plurality of chips).

[0091] Reference Figure 1 The communication unit 130 is shown as communicating with an external device of the device 100, but is not limited thereto and may also communicate with, for example, the storage unit 120 and / or Figure 1 The present invention relates to a device 100 that is not shown in the figure and communicates with at least some of the various components of the device 100.

[0092] Although Figure 1 Although not shown in the figure, the apparatus 100 may include at least one input / output unit or may be connected to at least one input / output unit.

[0093] For example, the input / output unit may include at least one display unit (eg, display) for outputting information related to the operation of the technical analysis program. In this case, the display unit may be composed of a touch screen capable of touch input.

[0094] In addition, the input / output unit may include at least part of an input unit (not shown) for inputting data, such as a microphone, a keyboard, and a mouse, and an output unit (not shown) for outputting data, such as a speaker and a driving unit.

[0095] According to various embodiments, the apparatus 100 may include at least a portion of functions of a range of all information communication devices including mobile communication terminals, multimedia terminals, wired terminals, fixed terminals, and Internet Protocol (IP) terminals.

[0096] The apparatus 100 is an apparatus for processing control instructions. The apparatus 100 may be configured to include at least a portion of the functions of a workstation or a large database or be connected thereto via communication.

[0097] The user device connected to the device 100 can be a mobile phone, a personal computer (PC), a portable multimedia player (PMP), a mobile internet device (MID), a smartphone, a tablet PC, a phablet PC, a notebook, etc.

[0098] According to various embodiments, the user device may be described as a device connected to and communicating with the device 100. Although not shown in the drawings, the user device may be, for example, a user's smartphone connected to the device 100 via wireless communication.

[0099] The user device may be connected to the device 100 via at least one installed program, or may be connected to the device 100 via at least one online webpage provided by the device 100. To this end, the device 100 may include at least a portion of the functions of a server and / or a terminal for providing a charging control service.

[0100] The server is an entity existing on the network that performs the functions of a web server, a database server, and an application server. According to a preferred embodiment, the server can provide various contents to the device 100 and / or the user device based on the processing of the device 100.

[0101] Reference Figure 3 The operation of the above-mentioned apparatus 100 for performing technical analysis will be described in more detail with reference to the accompanying drawings.

[0102] to this end, Figure 3 is a flowchart illustrating an operational flow of performing technical analysis based on an AI model in a device according to one embodiment. Figure 4 FIG. 1 is a diagram illustrating an operation flow of performing information collection in a device according to one embodiment. Figure 5 FIG. 1 is a diagram illustrating an operation flow of performing technology classification in a device according to one embodiment. Figure 6 FIG. 1 is a diagram exemplarily illustrating an operation of generating a search keyword in a device according to an embodiment. Figure 7 FIG. 1 is a diagram exemplarily illustrating an operation of classifying a specific technology in a device according to an embodiment. Figure 8 FIG. 4 is a diagram exemplarily illustrating an operation of classifying specific technologies based on technical documents in a device according to one embodiment. Figure 9FIG. 1 is a diagram exemplarily illustrating an operation of distinguishing technical literature data from network data to classify specific technologies in a device according to one embodiment. Figure 10 FIG. 4 is a diagram exemplarily illustrating operations of performing trend analysis and mining promising candidate technologies in a device according to one embodiment. Figure 11 FIG. 4 is a diagram exemplarily illustrating the results of mining promising candidate technologies and the generation operation of a customized report in a device according to one embodiment.

[0103] First, refer to Figure 2 and Figure 3 The information collection unit 201 can design (or generate) a meta-framework structure for the analysis object keyword, use the generative AI model 221 to derive a technical structure based on the meta-framework structure, generate search keywords based on the derived technical structure, and retrieve and obtain multiple documents based on the search keywords (301).

[0104] Reference Figure 4 To describe in more detail, the information collecting unit 201 may acquire analysis target keywords for performing technical analysis through user input or the communication unit 103. Here, the analysis target keywords may include words in a specific technical field.

[0105] In addition, the information collection unit 201 can extract words in a specific technical field from phrases or sentences obtained through user input or communication unit 103, and determine the extracted words as analysis target keywords. To this end, the specific technical field for technical analysis can be set in the setting information of the device 100.

[0106] The information collection unit 201 may design a meta-framework structure based on the acquired analysis object keywords or the preset technical field (401). More specifically, the information collection unit 201 may design a meta-framework structure by outputting at least one preset meta-framework item for at least one of the analysis object keywords and the preset technical field.

[0107] In addition, the information collecting unit 201 may design the meta-framework structure, which includes a plurality of meta-framework items derived based on public reports of at least one enterprise or at least one institution registered in the technical field corresponding to the analysis target keyword.

[0108] More specifically, the information collecting unit 201 may derive at least one meta-framework item based on the content, subject, topic, title, or words included in the report published by the enterprise or the organization.

[0109] Here, the enterprise may include at least one of an enterprise for which technology analysis is to be performed, an enterprise conducting business in the technology field, and a competing enterprise.

[0110] In addition, the institution may include at least one of various institutions that conduct technology-related business, such as the Patent Office, the Patent Court, the Korea Intellectual Property Office, the Korea Patent Information Institute, and the Korea Patent Strategy and Development Institute.

[0111] According to one embodiment, Figure 6 As shown, the meta-framework structure may include at least one meta-framework item among lower-level technologies, technical elements, process elements, application industries, social impacts and related technologies.

[0112] The information collection unit 201 can derive a technical structure for technical analysis based on the meta-framework structure (403). More specifically, the information collection unit 201 can use the generative AI model 221 to generate multiple technical items related to the analysis target keyword in the technical field corresponding to the analysis target keyword.

[0113] The information collection unit 201 may derive the technical structure by associating at least a portion of the generated multiple technical items with at least a portion of the multiple meta-framework items included in the meta-framework structure. In this case, the information collection unit 201 may derive the technical structure based on the generative AI model 221 .

[0114] Refer to the following Figure 6 An example of selecting "battery" as an analysis target keyword will be described.

[0115] Reference Figure 6 The information collection unit 201 can process the keyword "battery" input as the analysis object keyword and at least part of the elements "lower-level technology", "technical elements", "process elements", "applied industry", "social impact", and "related technologies" stored or newly input as meta-framework items as input values ​​of the generative AI model 221 to derive at least one technical item of at least one meta-framework item.

[0116] In this regard, Figure 6 Taking the analysis object keyword "battery" as an example, the information collection unit 201 can apply the "lower-level technology" meta-framework item to the generative AI model 221 to derive a technical structure including technical items such as "chemical substances and electrolytes", "battery design and architecture", "charging and discharging management", "reliability and safety technology" and "new materials and production".

[0117] In addition, the information collection unit 201 can derive a technical structure of at least part of the technical items including "electric vehicles and transportation", "energy storage and grid connection" and "mobile power supply and portable devices" for the "application industry" meta-framework item.

[0118] In addition, refer to Figure 6Although the figure shows that the technical structure is derived for the meta-framework items of "lower-level technology" and "application industry", the information collection unit 201 can derive a technical structure including at least one technical item for other meta-framework items, such as at least one meta-framework item of "technical elements", "process elements", "social impact" and "related technologies".

[0119] Refer again Figure 4 , the information collection unit 201 may generate search keywords based on the derived technical structure ( 405 ). At this time, the information collection unit 201 may generate search keywords based on the generative AI model 221 .

[0120] The information collecting unit 201 may generate at least one search keyword corresponding to each of a plurality of technical items included in the technical structure using the generative AI model 221 .

[0121] For example, refer to Figure 6 The information collection unit 201 processes the technical items of "chemical substances and electrolytes", "battery design and architecture", "charging and discharging management", "reliability and safety technology", "new materials and production", "electric vehicles and transportation", "energy storage and grid connection" and "mobile power supplies and portable devices" derived based on the generative AI model 221 as input values ​​of the generative AI model 221 for deriving keywords to generate at least one search keyword (or search term) for each technical item.

[0122] Reference Figure 6 , it can be confirmed that the information collection unit 201 generates search keywords such as “battery chemical substance development”, “electrolyte type” and “battery chemical reaction research” for the “chemical substance and electrolyte” technology item.

[0123] In addition, it can be confirmed that the information collection unit 201 generates search keywords such as “electric vehicle battery technology” and “electric vehicle charging management” for the technical item “electric vehicle and transportation”.

[0124] In addition, if Figure 6 As shown, it can be confirmed that the information collection unit 201 generates search keywords related to the other technical items, such as at least one of the technical items "battery design and architecture", "charging and discharging management", "reliability and safety technology", "new materials and production", "energy storage and grid connection" and "mobile power and portable devices".

[0125] In addition, the information collecting unit 201 may generate at least one search formula for each of the multiple technical items included in the technical structure using at least a portion of the at least one search keyword generated for each of the multiple technical items included in the technical structure.

[0126] At this time, the information collecting unit 201 may generate at least one search formula for searching for documents from at least a portion of a preset database, a web page (or network), and a search engine.

[0127] Here, the search engine may include a patent search engine, such as Korea Intellectual Property Rights Information Service (KIPRIS), World Intellectual Property System Online (WIPS On), Google Patents, USPTO Patent Full-Text and Image Database (USPTO), Espacene, and the like.

[0128] Refer again Figure 4 The information collection unit 201 may search and obtain (or collect and load) relevant documents (407) based on the generated search keywords and at least a part of the search formula.

[0129] In this case, information collection unit 201 may search and obtain documents containing text related to the analysis target keyword from a designated database, search engine, or network based on the search keyword and at least a portion of the search formula. Furthermore, information collection unit 201 may search and obtain at least one document containing text related to the analysis target keyword for each of the multiple technical items.

[0130] In addition, the information collection unit 201 can search and obtain documents related to the search keyword and search formula, the analysis target keyword, and at least a portion of the technical items included in the technical structure. In this case, the information collection unit 201 can search and obtain documents based on the generative AI model 221 and the deep learning AI model 223.

[0131] Afterwards, refer to Figure 2 and Figure 3 The technology classification unit 203 can refine the multiple documents according to the similarity between the documents, perform preprocessing for utilizing the contents of the multiple documents, and then perform specific technology classification (303) based on the result of performing content-based clustering.

[0132] Reference Figure 5In more detail, the technology classification unit 203 can refine data based on the similarity of the acquired documents (501). At this time, the technology classification unit 203 can perform data refinement based on the deep learning AI model 223.

[0133] The technology classification unit 203 may perform at least a portion of vectorization and similarity measurement operations on the acquired documents to perform data refinement.

[0134] More specifically, the technology classification unit 203 may extract text from the acquired document and perform vectorization on the extracted text. For example, the technology classification unit 203 may convert the text extracted from the document into a vector form.

[0135] More specifically, the technology classification unit 203 can represent words as dense vectors based on word embedding technology (e.g., Word2Vec, GloVe), or can represent the entire document as a vector based on document embedding technology (e.g., TF-IDF, document embedding), thereby vectorizing the document.

[0136] The technology classification unit 203 can use the vectorized documents to measure the similarity between documents. For example, the technology classification unit 203 can measure the distance between vectors based on at least some of various similarity measurement techniques, such as cosine similarity, Jaccard similarity, Euclidean distance, Manhattan distance, edit distance, and term frequency-inverse document frequency (TF-IDF).

[0137] The technical classification unit can refine (or improve) the quality of the documents, for example, by deleting at least one information determined to be unnecessary based on the similarity of the documents measured according to the distance between the document vectors, or grouping repeated words for consistency or accuracy of the data, or correcting typos.

[0138] The technology classification unit 203 may perform pre-processing for utilizing document contents or performing clustering on the refined documents (503).

[0139] For example, the technology classification unit 203 may perform at least some of various pre-processing operations on the text extracted from the document, such as removing special letters, converting English capital letters to lowercase letters, tokenization, title extraction, morpheme analysis, or removing stop words.

[0140] At this time, the technology classification unit 203 can distinguish document types. For example, the technology classification unit 203 can distinguish documents into technical document data and network data.

[0141] Here, the technical documents may include patent documents, standard documents or papers. In addition, the technical documents may further include at least one document among examination regulations, reports, design documents, manuals or educational materials.

[0142] The technology classification unit 203 may classify the documents classified as technical documents into at least one of various document categories, such as patent documents, standard documents, papers, examination regulations, reports, design documents, manuals, and educational materials.

[0143] At this time, the technology classification unit 203 may classify the document based on various features such as metadata of the document, whether the document includes specific words preset for classifying the document, and the document structure.

[0144] Thereafter, the technology classification unit 203 may measure the similarity of the documents based on the distance between the document vectors. When the similarity of documents included in the same technology item is greater than a specified value, they may be determined to be the same documents, and at least one document may be deleted as a duplicate document.

[0145] In addition, the technology classification unit 203 can judge the similarity between the documents included in the technology item and the technology item and the search keywords generated for the technology item. When the similarity is less than the specified value, it is determined that the relevance with the technology item or the search keyword is low, and at least one document is deleted as an irrelevant document.

[0146] In addition, the technology classification unit 203 may perform at least a part of various pre-processing operations such as deleting URLs included in the document or replacing specific words or patterns.

[0147] Thereafter, the technology classification unit 203 may perform clustering on the documents based on the vectorized and pre-processed documents (505).At this time, the technology classification unit 203 may perform content-based clustering on the documents.

[0148] More specifically, the technology classification unit 203 may apply a clustering algorithm to group the documents. For example, the technology classification unit 203 may apply at least some of various clustering algorithms such as k-means, hierarchical clustering, and DBSCAN to group the documents.

[0149] The technology classification unit 203 can measure the distance between vectors based on at least a portion of various similarity measurement techniques such as cosine similarity, Jaccard similarity, Euclidean distance, Manhattan distance, edit distance, and term frequency-inverse document frequency (TF-IDF).

[0150] The technology classification unit 203 may measure the similarity of documents based on the distance between vectorized documents or the distance between vectorized texts.

[0151] Here, the technology classification unit 203 can perform document clustering (505) for each of the multiple technology items included in the technology structure. At this time, the technology classification unit 203 can perform data extraction (501), preprocessing (503), and clustering (505) operations on the documents based on at least one deep learning AI model 223.

[0152] Here, the technology classification unit 203 may perform content-based clustering by distinguishing at least a portion of categories of technical documents including patents, papers, etc. and web data for the acquired at least one document of each of the plurality of technical items.

[0153] The technology classification unit 203 may generate at least one specific technology item for each of the multiple technology items based on the content-based clustering result and the generative AI model 221 ( 507 ).

[0154] For example, refer to Figure 7 , it can be confirmed that the technology classification unit 203 generates specific technology items such as “battery technology”, “electric vehicle transmission system”, “charging and infrastructure”, “vehicle connectivity and autonomous driving”, and “renewable energy integration” for the “electric vehicle and transportation” technology item.

[0155] In addition, the technology classification unit 203 may generate (or derive) a technology name, a technology abstract, and main keywords for each specific technology item in at least one specific technology item using at least one clustered document classified according to the content-based clustering.

[0156] Here, the technology classification unit 203 may generate at least one main keyword for each specific technology item in the at least one specific technology item from the clustered documents based on at least a portion of a generative AI and a text mining algorithm.

[0157] At this time, the text mining algorithm can be established based on part of the algorithms in TF-IDF (or c-TF-IDF), Word2Vec, and Doc2Vec.

[0158] In addition, the technology classification unit 203 can generate at least a part of the technology name and technology summary for each specific technology item in at least one specific technology item based on at least one main document and main keyword identified for each specific technology item in the at least one specific technology item according to content-based clustering.

[0159] At this time, the main document is determined based on whether it contains the main keywords. The more main keywords it contains, the closer it is to the center of the cluster, and it can be defined as the main document.

[0160] The technology classification unit 203 may determine a cluster center of each specific technology item in at least one specific technology item based on content-based clustering, and may determine a specified number of clustered documents located at the cluster center as main documents of the corresponding specific technology item.

[0161] More specifically, the technology classification unit 203 may list the at least one cluster document according to the order close to the cluster center, and determine a specified number of cluster documents as main documents according to the order close to the cluster center.

[0162] Reference Figure 7 It can be confirmed that the technology classification unit 203 generates (or exports) a technology summary and / or generates main keywords for the specific technology item of "battery technology", and the technology summary is "Battery technology is a technical field used to improve the type, performance, charging efficiency and life of batteries used in electric vehicles and transportation vehicles, including research and development of using various chemicals and materials to increase battery capacity to achieve longer driving mileage and faster charging speed", and the main keywords include "battery chemical composition", "battery material development", "battery design and modeling", "battery life extension", "battery reuse and environmental protection", "improvement of energy density and charging speed", and "new technologies and research trends". At least part of them.

[0163] Here, the technology classification unit 203 can generate (or export) technology names, technology summaries and main keywords not only for the specific technology item "battery technology", but also for at least part of the specific technology items such as "electric vehicle transmission system", "charging and infrastructure", "vehicle connectivity and autonomous driving", and "renewable energy integration".

[0164] In addition, the technology classification unit 203 can not only generate specific technology items for the "electric vehicles and transportation" technology item, but also generate specific technology items for at least part of multiple technology items such as "chemical substances and electrolytes", "battery design and architecture", "charging and discharging management", "reliability and safety technology", "new materials and production", "energy storage and grid connection", "mobile power supplies and portable devices", and can generate (or export) technology names, technology summaries and main keywords for at least part of the generated specific technology items.

[0165] In addition, the technology classification unit 203 may derive at least one major player (eg, a leading enterprise or a competing enterprise of the specific technology) in each specific technology item.

[0166] In addition, refer to Figure 8 and Figure 9 The technology classification unit 203 may generate specific technology items for the technology items based on at least a portion of the categories of the document.

[0167] More specifically, the technology classification unit 203 may generate at least one specific technology item for each type of technical literature data and network data based on the content-based clustering result and the generative AI model 221 .

[0168] For example, the technology classification unit 203 may generate a specific technology item for a technology item based on technology document data, or may generate a specific technology item for a technology item based on network data.

[0169] Reference Figure 8 , it can be confirmed that the technology classification unit 203 can distinguish the technical literature data and network data of the “charging and discharging management” technical item based on at least a part of the clustering results and the document category classification results and generate at least one specific technical item respectively.

[0170] For example, it can be confirmed that the technology classification unit 203 generates specific technical items such as "charging infrastructure improvement", "automatic charging system", "charging infrastructure connectivity", "battery capacity management", and "energy efficiency optimization" based on the technical literature data of the "charging and discharging management" technical item.

[0171] In addition, it can be confirmed that the technology classification unit 203 generates specific technology items such as "charging and discharging optimization", "wireless charging", "charging speed", "charging management", "analysis and prediction" based on the network data of the "charging and discharging management" technology item to derive specific technology classifications.

[0172] In addition, the technology classification unit 203 may derive at least a portion of technology names, technology summaries, main keywords, and main participants for at least a portion of specific technology items classified based on technology literature data or network data.

[0173] For example, refer to Figure 8 It can be confirmed that the technology classification unit 203 derives the main keywords such as "charging station improvement", "charging infrastructure innovation", "multiple charging ports", and "city charging" for the specific technology item "charging infrastructure improvement" in the technology item "charging and discharging management" based on the classification of technical literature data, and derives the main participants such as "ChargePoint", "EVBox", "EVgo", "Blink Charging", and "Electrify America".

[0174] In addition, it can be confirmed that the technology classification unit 203 exports the main keywords such as "charging and discharging efficiency", "charging and discharging algorithm", "software optimization" and "fast charging" for the specific technology item of "charging and discharging optimization" based on network data classification in the technology item of "charging and discharging management", and exports the main participants such as "ChargePoint", "Voltaiq", "ABB", "EVgo" and "Blink Charging".

[0175] In addition, if Figure 8 As shown, the technology classification unit 203 can export at least a part of the technology name, technology summary, main keywords and main participants for at least a part of the specific technology items classified as "automatic charging system", "charging infrastructure connectivity", "battery capacity management", "energy efficiency optimization" based on technical literature data and "wireless charging", "charging speed", "charging management", "analysis and prediction" based on network data.

[0176] In addition, the technology classification unit 203 may provide the number of documents classified by document category for each specific technology item of the classification.

[0177] Reference Figure 9 It can be confirmed that when the technology classification unit 203 outputs specific technology classification information based on technical documents, among the 73 patent documents classified as the "charging and discharging management" technology item according to the patent document category, the patent documents corresponding to the specific technology items of "charging infrastructure improvement", "automatic charging system", "charging infrastructure connectivity", "battery capacity management" and "energy efficiency optimization" are 12, 15, 21, 13 and 12 respectively.

[0178] In addition, the technology classification unit 203 may provide a patent document list classified for each specific technology item, and may provide at least one document included in the list or output a link to the document.

[0179] The technical classification unit 203 can not only provide document information, lists and documents (or document links) by specific technical items for patent documents, but also provide document information, lists and documents (or document links) by specific technical items for various document categories such as standard documents, papers, examination regulations, reports, design documents, manuals, educational materials, etc.

[0180] As described above, the technology classification unit 203 may generate at least one specific technology item for at least one technology item based on the generative AI model 221. In addition, the technology classification unit 203 may derive at least a portion of the technology name, technology summary, main keywords, and main participants for at least one specific technology item based on the generative AI model 221 and at least a portion of the text mining algorithm.

[0181] In addition, if the technology classification unit 203 confirms that there are no non-clustered documents after clustering the documents, the technology classification unit 203 can assign specific technology items to the non-clustered documents based on the last clustering result. At this time, if no specific technology items corresponding to the non-clustered documents are confirmed, the technology classification unit 203 can perform content-based clustering again on the documents including the non-clustered documents.

[0182] Afterwards, refer to Figure 10 , the trend analysis unit 205 may perform trend analysis (305) based on the collection results and classification results performed based on the information collection (301) and the technical classification (303). At this time, the trend analysis unit 205 may perform trend analysis based on the generative AI model 221 or the deep learning AI model 223.

[0183] In addition, the trend analysis unit 205 further performs identification of main research subjects and manpower based on the collection results and the classification results.

[0184] More specifically, the trend analysis unit 205 can perform trend analysis on technical analysis keywords, technical items, and at least a part of specific technical items based on various information obtained based on information collection (301) and technical classification (303), such as technical items generated based on meta-framework items, search keywords and search formulas generated for technical items, documents obtained based on at least a part of the search keywords and search formulas, clustering results of documents, and at least one specific technical item classified for technical items, technical names derived for specific technical items, technical summaries, main keywords, main participants, etc.

[0185] For example, refer to Figure 10 It can be confirmed that the trend analysis unit 205 performs trend analysis of proportion analysis by specific technology classification (A), time series trend analysis (B), application frequency analysis by country / region (C), and future growth forecast based on technology life curve (D).

[0186] More specifically, regarding the weight analysis (A) by specific technology category, the trend analysis unit 205 can analyze the weight of each technology in the overall technology trend for each specific technology item. Based on this, the trend analysis unit 205 can determine the growth potential or competitiveness of the corresponding specific technology item.

[0187] Furthermore, regarding time series trend analysis (B), the trend analysis unit 205 can analyze the temporal trends of technology items or specific technology items. Based on this, the trend analysis unit 205 can determine the speed or pattern of technological development and predict the direction of technological development. For example, the trend analysis unit 205 can analyze the changes in the number of applications and the speed of development of specific technology items or specific technology items to predict future technological trends.

[0188] In addition, regarding the application frequency analysis by country / region (C), the trend analysis unit 205 can analyze the application frequency of technologies related to a technology item or a specific technology item in a specific country or region. Based on this, the trend analysis unit 205 can determine the technological leadership, competitiveness, major players, etc. of the technology item or specific technology item in the specific country.

[0189] Furthermore, regarding the future growth forecast (D) based on the technology lifecycle curve, the trend analysis unit 205 can analyze data related to the lifecycle of a specific technology to predict future growth. Based on this, the trend analysis unit 205 can predict the growth momentum of the specific technology, the emergence of competing technologies, the possibility of technological substitution, etc.

[0190] Afterwards, refer to Figure 10 and Figure 11 , the candidate technology mining unit 207 can use the trend analysis results to mine promising candidate technologies (307). At this time, the candidate technology mining unit 207 can perform mining of promising candidate technologies based on the generative AI model 221 or the deep learning AI model 223.

[0191] For example, refer to Figure 10 It can be confirmed that the candidate technology mining unit 207 derives the prospect evaluation and candidate technology mining based on impact / latestness / frequency (E), screening by representative patents / papers and major participants of candidate technologies (F), external environment analysis based on competitiveness (G), and a promising technology evaluation map (H) from the perspective of internal and external forces based on the trend analysis results.

[0192] More specifically, with respect to the prospect evaluation and candidate technology mining (E) based on impact / recency / frequency, the candidate technology mining unit 207 can utilize at least a portion of the results of the weight analysis (A) by specific technology classification and the results of the application frequency analysis (C) by country / region to evaluate the impact and recency of a specific technology field.

[0193] At this point, the candidate technology mining unit 207 can determine the application frequency and growth momentum of a specific technology item or a specific specific technology item based on the time series trend analysis (B) results, and estimate promising technology fields based on the frequency. Based on this, the candidate technology mining unit 207 can mine candidate technologies in promising technology fields.

[0194] In addition, regarding the screening of candidate technologies by representative patents / papers and major players (F), the candidate technology mining unit 207 can analyze representative technical documents (e.g., patents, papers, etc.) in the mined candidate technologies to assess the reliability and state of the art of the technologies. The candidate technology mining unit 207 can identify major players such as enterprises, research institutions, and individuals to assess their dominance and competitiveness in the technology field.

[0195] Furthermore, regarding the competitiveness-based external environment analysis (G), the candidate technology mining unit 207 can analyze the technologies (e.g., patents, papers, products) of competing enterprises to assess competitive trends and competitiveness in technology fields. The candidate technology mining unit 207 can determine the preferred fields and application trends of competing enterprises based on at least a portion of the results of the weight analysis (A) by specific technology classification and the application frequency analysis (C) by country / region.

[0196] Furthermore, regarding the Promising Technology Assessment Chart (H) from the perspective of internal and external forces, the candidate technology mining unit 207 can integrate the internal forces and external environment analysis to derive the Promising Technology Assessment Chart based on the analysis results of the promising technology fields and competing companies identified from the weight analysis by specific technology classification (A) and the application frequency analysis by country / region (C). Based on this, the candidate technology mining unit 207 can intuitively display the position of promising technology fields and competing companies in the internal forces and external environment.

[0197] In addition, the candidate technology mining unit 207 can output trend analysis and promising candidate technology mining results. At this time, the candidate technology mining unit 207 can output trend analysis and promising candidate technology mining results based on the generative AI model 221 or the deep learning AI model 223.

[0198] For example, Figure 11As shown, the candidate technology mining unit 207 can output trend analysis and promising candidate technology mining results for at least part of the topics including technology growth analysis (1), core patent classification and abstract information (2), information on major technology owners (3), and internal power and external environment analysis (4).

[0199] More specifically, the candidate technology mining unit 207 may generate technology growth analysis (1) and results based on the time series trend analysis (B) results and the future growth prediction (D) results based on the technology life curve.

[0200] In addition, the candidate technology mining unit 207 can derive core patent classification and abstract information (2) and generate results based on the weight analysis (A) results by specific technology classification and the application frequency analysis (C) results by country / region.

[0201] In addition, the candidate technology mining unit 207 can derive the information of the main participants of the technology (3) and generate results based on the prospect evaluation based on impact / recentness / frequency and the candidate technology mining (E) results and the screening (F) results by representative patents / papers and main participants of the candidate technology.

[0202] In addition, the candidate technology mining unit 207 can generate internal force and external environment analysis (4) and results based on the external environment analysis (G) result based on competitiveness and the promising technology evaluation map (H) result from the internal and external forces.

[0203] In addition, the candidate technology mining unit 207 can generate a customized report (or customized technology analysis report) based on the trend analysis and promising candidate technology mining results. In this case, the candidate technology mining unit 207 can generate a customized report based on the generative AI model 221 or the deep learning AI model 223, based on the trend analysis and promising candidate technology mining results.

[0204] For example, refer to Figure 11 It can be confirmed that the candidate technology mining unit 207 can generate a technology perception and operations and intelligence (O / I) candidate mining report or a research and development (R&D) topic mining / promotion report based on the technology growth analysis results (1), core patent classification and abstract information (2), technology ownership information of major participants (3), and internal power and external environment analysis results (4).

[0205] In addition, it can be confirmed that the candidate technology mining unit 207 formulates a new business promotion strategy and generates a merger and acquisition (M&A) candidate mining report based on the technology growth analysis (1) results and the technology ownership main player information (3).

[0206] In addition, it can be confirmed that the candidate technology mining unit 207 generates an intellectual property (IP) planning / strategy formulation report based on the technology growth analysis (1) results and the core patent classification and abstract information (2).

[0207] As described above, according to various embodiments, through the technical analysis method and apparatus, the scope of various technical analyses can be expanded while reducing the time required for technical analysis.

[0208] According to various embodiments, the technical analysis method and apparatus can reduce the cost and time required for basic technical analysis based on raw data, thereby allowing data experts to focus on the analysis of advanced information, thereby improving work efficiency.

[0209] According to various embodiments, a technology analysis method and apparatus based on a generative AI model are provided, thereby enabling the mining of analysis indicators and methods reflecting the multifaceted nature of technology.

[0210] According to the detailed description, the functions of various embodiments described as being performed by the device 100 are operations processed by the processing unit 110 of the device 100 and may be performed by components organically connected to the device 100 and / or devices connected to the device 100.

[0211] The functions performed by the device 100 are described in various embodiments in this disclosure. However, if a component performing the function is not described as a component of the device 100, it should be understood that the device 100 includes and / or is connected to a conventionally known component performing the function.

[0212] As described above, although the embodiments have been described with reference to the limited drawings, a person skilled in the art can make various technical modifications and variations based on the various embodiments.

[0213] For example, even if the described techniques are performed in an order different from that of the described methods, and / or the described systems, structures, devices, circuits, etc. components are combined or combined in a form different from that of the described methods, or are replaced or substituted with other components or equivalents, appropriate results can still be achieved.

[0214] In particular, when describing with reference to flowcharts, although a plurality of steps are described as being included and the steps are sequentially executed in a specified order, the steps are not necessarily limited to the described order.

[0215] In other words, among the steps recorded in the flowchart, modifying and deleting at least a part to execute, or adding at least one step can all be embodiments, and executing more than one step in parallel can also be an embodiment, that is, these steps are not necessarily limited to operating in chronological order, but are included in the embodiments.

[0216] Accordingly, other implementations, other examples, and equivalents are within the scope of the claims.

Claims

1. A technical analysis method comprising: An information collection step, designing a meta-framework structure for the analysis object keywords, using a generative AI model to derive a technical structure based on the meta-framework structure, generating search keywords based on the derived technical structure, and searching and acquiring multiple documents based on the search keywords; as well as The technology classification step refines the plurality of documents according to the similarities between the documents, performs preprocessing for utilizing the contents of the plurality of documents, and then performs specific technology classification based on the result of performing content-based clustering.

2. The technical analysis method according to claim 1, wherein: In the information collection step, the meta-framework structure is designed, and the meta-framework structure includes a plurality of meta-framework items derived from public reports of at least one enterprise or at least one institution registered in the technical field corresponding to the analysis target keyword.

3. The technical analysis method according to claim 1, wherein: The plurality of meta-framework structures include at least one meta-framework item among lower-level technologies, technical elements, process elements, application industries, social impacts, and related technologies.

4. The technical analysis method according to claim 1, wherein: In the information collection step, the generative AI model is used to generate multiple technical items related to the analysis object keyword in the technical field corresponding to the analysis object keyword, and the technical structure is exported by associating at least a part of the multiple technical items with at least a corresponding part of the multiple meta-framework items included in the meta-framework structure.

5. The technical analysis method according to claim 1, wherein: In the information collection step, at least one search keyword corresponding to each of the plurality of technical items included in the technical structure is generated using the generative AI model, and At least one search formula is generated for each of the plurality of technical items using at least a portion of the at least one search keyword generated for each of the plurality of technical items.

6. The technical analysis method according to claim 5, wherein: In the information collection step, based on the search keyword or the at least one search formula, documents including texts related to the analysis target keyword are searched and acquired from a designated database or network.

7. The technical analysis method according to claim 1, wherein: In the technology classification step, for each of the multiple technology items included in the technology structure, the extraction, the preprocessing and the content-based clustering are performed on at least one acquired document based on a deep learning AI model.

8. The technical analysis method according to claim 7, wherein: In the technology classification step, based on the content-based clustering result and the generative AI model, at least one specific technology item is classified for each of the multiple technology items, Using at least one clustered document classified according to the content-based clustering, a technical name, a technical summary and main keywords are generated for each specific technical item in the at least one specific technical item.

9. The technical analysis method according to claim 8, wherein: In the technical classification step, based on at least a part of the generative AI and text mining algorithms, at least one main keyword is generated from the clustered documents for each specific technical item in the at least one specific technical item.

10. The technical analysis method according to claim 8, wherein: In the technology classification step, based on at least one main document and the main keywords identified for each of the at least one specific technical item according to the content-based clustering, at least a part of the technology name and the technology abstract are generated for each of the at least one specific technical item.

11. The technical analysis method according to claim 10, wherein: In the technical classification step, the cluster center of each specific technical item in the at least one specific technical item is determined based on the content-based clustering, and a specified number of cluster documents located at the cluster center are determined as the main documents of the corresponding specific technical item.

12. The technical analysis method according to claim 6, wherein: In the technology classification step, the content-based clustering is performed by distinguishing the categories of technical documents including patents or papers and network data for at least one acquired document of each of the plurality of technical items, Based on the content-based clustering results and the generative AI model, classify at least one specific technical item according to the categories of the technical documents and network data, A main keyword is generated for each specific technical item in the at least one specific technical item by using at least one clustered document classified according to the content-based clustering.

13. The technical analysis method according to claim 1, further comprising: a trend analysis step of performing trend analysis based on the collection results and classification results performed in the information collection step and the technology classification step; as well as The candidate technology mining step utilizes the trend analysis results to mine promising candidate technologies.

14. The technical analysis method according to claim 13, wherein: In the trend analysis step, identification of main research subjects and manpower is further performed based on the collection results and the classification results.

15. A technical analysis device comprising: A storage unit storing a generative AI model; An information collection unit designs a meta-frame structure for the analysis object keywords, uses the generative AI model to derive a technical structure based on the meta-frame structure, generates search keywords based on the derived technical structure, and searches and obtains multiple documents based on the search keywords; as well as The technology classification unit refines the plurality of documents according to similarities between the documents, performs preprocessing for utilizing the contents of the plurality of documents, and then performs specific technology classification based on a result of performing content-based clustering.

16. The technical analysis device according to claim 15, wherein: The information collection unit designs the meta-framework structure, which includes a plurality of meta-framework items derived based on a public report of at least one enterprise or at least one institution registered in the technical field corresponding to the analysis target keyword.

17. The technical analysis device according to claim 15, wherein: The information collection unit uses the generative AI model to generate multiple technical items related to the analysis object keyword in the technical field corresponding to the analysis object keyword, and derives the technical structure by associating at least a part of the multiple technical items with at least a corresponding part of the multiple meta-framework items included in the meta-framework structure.

18. The technical analysis device according to claim 15, wherein: The information collection unit generates at least one search keyword corresponding to each of the plurality of technical items included in the technical structure using the generative AI model, and At least one search formula is generated for each of the plurality of technical items using at least a portion of the at least one search keyword generated for each of the plurality of technical items.

19. The technical analysis device according to claim 15, wherein: The technology classification unit performs the extraction, the preprocessing and the content-based clustering on at least one acquired document based on a deep learning AI model for each of the multiple technology items included in the technology structure.

20. The technical analysis device according to claim 15, further comprising: a trend analysis unit that performs trend analysis based on the collection results and classification results performed in the information collection unit and the technology classification unit; as well as The candidate technology mining unit mines promising candidate technologies using the trend analysis results.

Images