AI-based IP valuation system

Abstract

To provide a system for evaluating an IP value without a subjective evaluation stage of an expert on the basis of objective input data of a general user including a nonexpert.SOLUTION: The IP value evaluation system 10 includes a value evaluation database including conformity information data, patent information data, and economic statistics information data and including statistical data and AI learning data, a collection and purification module configured to calculate and provide statistical data and AI learning data required in a process of generating AI learning data or a core variable, and an AI model for each core variable and configured to learn an AI model for each of the collection and purification module and calculate a corresponding explanatory variable matched with each core variable based on input evaluation target IP information and calculate a corresponding prediction variable value through each explanatory variable value collected or calculated by the collection and purification module and the AI model; The system includes a AI module for respectively calculating core variables and a value evaluation service module for calculating an object IP value based on the core variables and generating a value evaluation report including statistic data.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an AI-based IP valuation system, and concerns a technology that matches patent classification information with industry classification information or import / export codes, links and combines patent statistics information, industry classification information, and company information, collects and processes various data to calculate relevant statistical data and training data from raw data, trains an AI model, estimates key core variables necessary for valuation through the AI ​​model, and provides IP valuation results along with relevant statistical data. [Background technology]

[0002] While the demand for IP (Intellectual Property) valuation is steadily increasing, traditional expert-based valuation methods have faced challenges such as difficulty in ensuring consistency in valuation results due to the subjective evaluations of experts, and the considerable time required for preliminary investigations and report preparation.

[0003] Furthermore, when attempting to value patents held by a business entity, it is generally possible to value a patent portfolio that includes multiple patents related to the business, rather than a single patent.

[0004] At this time, a challenge arose: as the number of patents included in the portfolio increased, so did the time and resources required.

[0005] Furthermore, while there were solutions that received input of evaluation results for individual evaluation elements necessary for value calculation from various experts or users, and calculated patent value and grade based on the input data, these solutions had the drawback that expert evaluations of individual evaluation elements had to be intervened, or that the input data used was limited, resulting in evaluation results that were simple and limited.

[0006] Furthermore, when only the evaluation results for patents were provided, non-experts had difficulty interpreting and utilizing those results. [Prior art documents] [Non-patent literature]

[0007] [Non-Patent Document 1] Managing Technology:The Technology Valuation Approach(IEEE,2007) [Non-Patent Document 2] A Comparative Studyon Methods of Income Approach to Technology Valuation(Journal of Supply Chain and Operations Management,Volume 10,Number 2,September 2012) [Non-Patent Document 3] Internal technology valuation:real world issues(International Journal of Technology Management,Vol.53,No.2-4,2011) [Non-Patent Document 4] Income approach to technology valuation for innovations(International Journal of Technology Management,Vol.88,No.2-4,2022) [Non-Patent Document 5] Review of Methods of New Technology Valuation(IEEE, 2010) [Overview of the project] [Problems that the invention aims to solve]

[0008] Therefore, the technical problem that the present invention aims to solve is to provide a system that can perform IP value evaluation based on objective input data from general users, including non-experts, without requiring a subjective evaluation stage by experts.

[0009] Furthermore, the technical problem that this invention aims to solve is to provide a system for portfolio valuation that objectively and efficiently evaluates the value of the entire portfolio, regardless of the number of individual patents, rather than relying on conventional methods of summarizing individual evaluations of individual patents or evaluating core patents based on the subjective judgment of business entities or experts.

[0010] Furthermore, the technical problem that this invention aims to solve is to estimate core variable values ​​necessary for value calculation via an AI model, train the AI ​​model using various data and input datasets, and improve the reliability of system evaluation results by utilizing the calculated values ​​of the model optimized for each core variable.

[0011] The technical problem that this invention aims to solve is to provide a system that provides relevant statistical data used or processed for core variable estimation and value calculation, which can be used to interpret evaluation results and provide insights into the relevant industrial environment. [Means for solving the problem]

[0012] An IP valuation system according to one aspect of the present invention for solving the aforementioned technical problems is an AI (Artificial Intelligence) based IP (Intellectual Property) valuation system comprising: a valuation database including compliance information data, patent information data, and economic statistical information data as raw information, and statistical data and AI learning data as extracted information processed from the raw information; a collection and purification module that calculates and provides the statistical data and AI learning data necessary in the process of collecting and processing the raw information to generate AI learning data or first to fourth core variables, and stores them in the valuation database; and, for the calculation of the first to fourth core variables, two or more AI models for each core variable are trained via the AI ​​learning data, and descriptions are matched for each core variable based on the input IP information to be evaluated. The AI ​​module includes: an AI module that checks variables and calculates each explanatory variable value collected or calculated by the collection and refinement module, and predictor variable values ​​via the AI ​​model, and calculates core variable values, by calculating the first predictor variable and first core variable value via the first explanatory variable set, the second predictor variable and second core variable value via the second explanatory variable set, the third predictor variable and third core variable value via the third explanatory variable set, and the fourth predictor variable and fourth core variable value via the fourth explanatory variable set; and a valuation service module that calculates the target IP value based on the first to fourth core variable values ​​and generates a valuation report including the IP value and the statistical data.

[0013] In this case, the first to fourth core variables are the IP economic lifespan, royalties, discount rate, and sales revenue, respectively.

[0014] The first to fourth predictor variables are, respectively, IP economic lifespan influencing factors, royalty rate influencing factors, IP commercialization risk premium, and sales growth rate.

[0015] Furthermore, the first to fourth predictor variables are, respectively, used during AI model training, to determine the expert assessment results for the IP's economic lifespan and the TCT (Technology Cycle Time) for the patent classification information to which the target IP belongs. median The difference between these factors, the expert-rated royalty rate, and the industry classification information matched to the patent classification information to which the target IP belongs, are defined and learned as the ratio between standard royalty rates for that industry, the expert-rated IP commercialization risk premium, and the industry-specific sales growth rate.

[0016] The AI ​​module then checks the TCT median of the patent classification information to which the target IP belongs, calculates the first core variable by reflecting the first predictor variable value in the TCT median, checks the standard royalty rate for the industry using industry classification information matched to the patent classification information to which the target IP belongs, calculates the second core variable by reflecting the second predictor variable value in the standard royalty rate, checks the cost of equity and ratio, cost of debt and ratio for the industry using industry classification information matched to the patent classification information to which the target IP belongs, calculates the third core variable by reflecting the third predictor variable value in the cost of equity, and if the past sales of the business entity owning the target IP are confirmed, sets the initial sales through the past sales, calculates the fourth core variable by reflecting the fourth predictor variable value in the initial sales.

[0017] Furthermore, if the AI ​​module cannot confirm the past sales of the business entity that owns the target IP, it will set the initial sales figures through sales statistics of companies of a size already defined in that industry.

[0018] And the first set of explanatory variables for generating the first prediction variable includes the applicant's application number increase / decrease rate (application increase / decrease rate / applicant increase / decrease rate), TCT statistical value, evaluation elements of the grade evaluation system, evaluation index scores of the grade evaluation system, average number of US patent litigations by patent classification information, and average number of trial-related cases by patent classification information. The second set of explanatory variables for generating the second prediction variable includes royalty rate statistics, average number of opinion submissions by patent classification, evaluation elements of the grade evaluation system, evaluation index scores of the grade evaluation system, and number of divisional applications and number of priority claims by patent classification. The third set of explanatory variables for generating the third prediction variable includes, as the optimized explanatory variables for predicting the IP commercialization risk premium, sales growth rate by enterprise scale and industry type, operating profit growth rate by enterprise scale and industry type, evaluation elements of the grade evaluation system, evaluation index scores of the grade evaluation system, and patent concentration. The fourth set of explanatory variables for generating the fourth prediction variable includes the number of applicants, the increase / decrease rate of the number of applications, and the increase / decrease rate of imports and exports.

[0019] Also, the target IP information is the patent registration number of the target IP.

[0020] And the value evaluation service module calculates the target IP value by the royalty exemption method based on the first to fourth core variable values.

[0021] In addition, the compliance information data includes IP evaluation result data of experts and actual IP transaction information data. The patent information data includes, as patent detailed information, citation / cited data by patent, application data, trial data, litigation data, registration data, and family data. As grade evaluation information, it includes grade evaluation element data by patent and score data for the indicators meant by the evaluation elements according to the evaluation results for each element. The economic statistical information data includes, as economic market information, sales growth rate data by industry type, sales statistics data, and economic forecast data. As financial information, it includes stock price data, bond yield data, and corporate financial data. As import / export information, it includes import / export data.

[0022] In this case, the collection and purification module includes: a raw data collection unit for collecting the raw information; a preprocessing unit for performing preprocessing on the collected raw information; a base data calculation unit for generating base data for calculating training data and evaluation criteria data from the preprocessed data; a training data calculation unit for generating AI model training data for calculating the core variables from the base data; a statistical data calculation unit for generating statistical data for one or more of the explanatory variables, predictor variables, and core variables as generated or necessary statistical data in the AI ​​training data or the core variable calculation process; and an evaluation criteria data calculation unit for calculating final evaluation criteria data based on the first to fourth core variable values ​​calculated via the AI ​​module and transmitting it to the value evaluation service module.

[0023] Furthermore, the evaluation criteria data calculation unit calculates the IP economic lifespan as the first core variable value, taking into account one or more of the legal remaining lifespan and commercialization preparation period of the target IP as evaluation criteria data, and calculates the corporate tax rate and corporate tax based on the calculated sales revenue.

[0024] Furthermore, the statistical data used or extracted in the first core variable calculation process includes TCT data, appeal-related statistical data, US litigation data, and market concentration data for the industry sector for the patent classification; the statistical data used or extracted in the second core variable calculation process includes industry-based royalty rate data, the number of opinions submitted, the number of divisional applications, and the number of priority claims for the patent classification; the statistical data used or extracted in the third core variable calculation process includes self / debt capital cost data for the industry and self / debt capital ratio data by industry, patent concentration for the patent classification, and sales / operating profit growth rate data for the industry; and the statistical data used or extracted in the fourth core variable calculation process includes initial sales statistics data according to the industry and size to which the target IP belongs, statistical data on the number of applicants and the rate of increase / decrease in the number of applications for the patent classification, and export / import increase / decrease data for the industry and product.

[0025] The AI ​​module includes: a learning data preprocessing unit that performs preprocessing on AI learning data; an AI learning unit that trains one or more AI models for each of the first to fourth core variables via the AI ​​learning data; a learning optimization unit that sets up one or more AI models optimized for each core variable according to performance indicator results based on the verification results for the AI ​​model predictions; and a core variable calculation unit that calculates core variables via explanatory variables matched for each core variable and predictor variables calculated via the AI ​​model.

[0026] Furthermore, the core variable calculation unit, when there are two or more IPs to be evaluated and it is an IP portfolio, sets the average of the median TCTs for each individual patent's patent classification information as the baseline TCT for the target IP portfolio, calculates the IP economic lifespan influence factors for each individual patent, sets the largest influence factor as the first predictor variable for the target portfolio, reflects the value of the first predictor variable for the target portfolio in the baseline TCT for the target portfolio to calculate the first core variable for the target portfolio, sets the industry according to user input information, or sets the industry with the largest median sales of small businesses in the industry category among the industries matched to each individual patent's patent classification information as the representative industry for the target portfolio, sets the baseline royalty rate of the representative industry for the target portfolio as the baseline royalty rate for the target portfolio, calculates the influence factors on the royalty rate for each individual patent, and sets the largest value The influencing factors are set as the second predictor variable for the target portfolio, the second core variable for the target portfolio is calculated by reflecting the value of the second predictor variable for the target portfolio in the target portfolio's standard royalty rate, the IP commercialization risk premium is calculated for each individual patent, the smallest IP commercialization risk premium is set as the third predictor variable for the target portfolio, the third core variable for the target portfolio is calculated by reflecting the value of the third predictor variable in the cost of equity among the weighted average cost of capital of the target portfolio's representative industry, the fourth predictor variable for the target portfolio is generated based on patent classification information and import / export classification information matched to the target portfolio's representative industry, initial sales are set through the business entity's past sales information or sales statistics of the target portfolio's representative industry, and the fourth core variable for the target portfolio is calculated by reflecting the value of the fourth predictor variable in the initial sales. [Effects of the Invention]

[0027] As explained above, according to the present invention, IP value can be assessed based on objective input data from general users, including non-experts, without the need for a subjective evaluation process by experts.

[0028] In particular, according to the present invention, portfolio valuation can be performed quickly, efficiently, and objectively, regardless of the number of individual patents.

[0029] Furthermore, according to the present invention, it is possible to obtain highly reliable value evaluation results by generating various training data and utilizing multiple AI models trained with the said training data.

[0030] Furthermore, according to the present invention, core variables necessary for valuation can be estimated or relevant statistical data used or calculated during the valuation process can be provided together, thereby providing users with insights into the relevant industry environment and IP utilization. [Brief explanation of the drawing]

[0031] [Figure 1] This is an overall diagram of the IP value evaluation system relating to one aspect of the present invention. [Figure 2] This is a detailed configuration diagram of a value assessment service module relating to one aspect of the present invention. [Figure 3] This is a detailed configuration diagram of a collection / purification module relating to one aspect of the present invention. [Figure 4] This is a detailed configuration diagram of an AI module relating to one aspect of the present invention. [Figure 5] This is a detailed configuration diagram of a stacking ensemble model relating to one aspect of the present invention. [Figure 6] This diagram illustrates the explanatory variables and predictive variables used as input to the AI ​​model for calculating the IP economic lifespan, which is the first core variable among the core variables calculated for evaluating the patent value relating to one aspect of the present invention. [Figure 7] Examples of explanatory variables and predictive variables are shown in the diagram, which are input into the AI ​​model for calculating the royalty rate, the second core variable among the core variables calculated for patent valuation relating to one aspect of the present invention. [Figure 8]Examples of explanatory variables and predictive variables are shown in the diagram, which are input into the AI ​​model for calculating the discount rate, the third core variable among the core variables calculated for patent valuation relating to one aspect of the present invention. [Figure 9] Examples of explanatory variables and predictive variables are shown in the diagram, which are input into the AI ​​model to calculate sales revenue, the fourth core variable among the core variables calculated for patent valuation relating to one aspect of the present invention. [Figure 10] This is a detailed configuration diagram of a value valuation management module relating to one aspect of the present invention. [Figure 11] This is a detailed configuration diagram of a value evaluation database relating to one aspect of the present invention. [Figure 12] This section describes the overall process by which the system evaluates IP value based on objective target IP information entered by the user, and generates a report, in accordance with one aspect of the present invention. [Figure 13] In accordance with one aspect of the present invention, the overall process by which the system calculates the final value of a portfolio based on objective target IP portfolio information entered by the user will be explained. [Figure 14] In accordance with one aspect of the present invention, the overall process by which the system calculates the final value of a portfolio based on objective target IP portfolio information entered by the user will be explained. [Figure 15] This is a diagram illustrating an example of a computer device according to one aspect of the present invention. [Modes for carrying out the invention]

[0032] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings, so that they can be easily implemented by a person with ordinary skill in the art to which the present invention pertains. However, the present invention can be implemented in a variety of different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly illustrate the present invention in the drawings, unnecessary parts have been omitted, and similar components throughout the specification have been denoted by similar reference numerals.

[0033] When a specification as a whole states that a part "includes" a certain component, unless otherwise stated, this means that it may include other components rather than excluding them.

[0034] Furthermore, terms such as "...part," "...device," and "...module" as described in the specification refer to a unit that processes at least one function or operation, which may be implemented by hardware, software, or a combination of hardware and software.

[0035] The apparatus described in this invention consists of hardware including at least one processor, a memory device, a communication device, etc., and stores a program that is executed in combination with the hardware at a designated location. The hardware has a configuration and performance that enables the method of the present invention. The program includes instructions that implement the operating method of the present invention as described with reference to the drawings, and executes the present invention in combination with hardware such as the processor and memory device.

[0036] In this specification, “transmission or provision” may include not only direct transmission or provision, but also indirect transmission or provision via other devices or by utilizing alternative routes.

[0037] In this specification, expressions written in the singular may be interpreted as singular or plural unless explicitly stated otherwise, such as "one" or "single."

[0038] In this specification, regardless of the drawings, the same drawing number refers to the same component, and "and / or" includes each of the components mentioned and all combinations of one or more of them.

[0039] Terms including ordinal numbers such as "first," "second," etc., can be used to describe various components, but the components are not limited to those terms. The terms are used solely for the purpose of distinguishing one component from another. For example, without exceeding the scope of the rights of this disclosure, the first component may be named the second component, and similarly, the second component may be named the first component.

[0040] In this specification, in flowcharts explained with reference to the drawings, the order of operations can be changed, various operations can be merged, certain operations can be split, and certain operations do not need to be performed.

[0041] Furthermore, this invention is a system that applies the royalty exemption method, one of the various methodologies used in IP valuation. In this case, the royalty exemption method is a method for estimating the value of the IP by estimating the appropriate royalties that would be incurred if one does not own the rights to the IP being valued. Hereinafter, this specification will explain using the case where the IP being valued is a patent as an example.

[0042] More specifically, in patent valuation, the royalty exemption method is a patent valuation method that estimates the present value of royalties that should be paid as licensing fees during the economic life of the patent being valued.

[0043] In particular, the aforementioned royalty exemption method is suitable for evaluating patents of startups or small and medium-sized enterprises that possess IP but have not yet generated sales, and is also suitable for evaluating R&D-calculated patents that are not easily commercialized.

[0044] When evaluating patent value using the royalty exemption method, the specific calculation formula is as follows:

number

[0045] Therefore, according to the present invention, the IP valuation system can calculate the economic lifespan of the target IP (hereinafter also referred to as the "first core variable"), the royalty rate (hereinafter also referred to as the "second core variable"), the discount rate (hereinafter also referred to as the "third core variable"), and the estimated sales revenue (hereinafter also referred to as the "fourth core variable") as core variables based on objective input data from the user, and can perform a valuation of the patent to be evaluated.

[0046] The IP value evaluation system according to one aspect of the present invention will be described in more detail below with reference to the drawings.

[0047] First, as illustrated in Figure 1, the IP valuation system 10 may include a valuation service module 100, a collection / refinement module 200, an AI module 300, a valuation management module 400, and a valuation database 500.

[0048] The valuation service module 100 can receive objective information data related to the IP to be evaluated from the user, receive evaluation criteria data generated based on the said information data, calculate the IP value, and generate a report.

[0049] Specifically, the valuation service module 100 provides an interface that allows the user to input information about the IP to be evaluated, and can receive such information.

[0050] Furthermore, the valuation service module 100 can also receive, as objective information data, information related to the business entity that owns the IP to be evaluated, such as business scale and industry-related information, based on user selection.

[0051] The valuation service module 100 can manage valuation attributes for the target IP, such as the purpose of evaluation and the evaluation method.

[0052] The valuation service module 100 can perform IP valuation based on evaluation objectives and methods set via evaluation criteria data calculated from the target IP information.

[0053] Furthermore, the valuation service module 100 can generate an IP valuation report that includes the IP valuation results and relevant statistical data calculated during the valuation process.

[0054] The collection / purification module 200 can periodically collect, purify, and process raw information necessary for IP valuation, and extract the required information.

[0055] Specifically, the collection / refinement module 200 can collect, for example, IP subject expert valuation data, university technology transfer data, public research institute technology transfer data, and exchange IP transaction data as reference information.

[0056] Furthermore, the collection / refinement module 200 can collect patent information such as citation / cited data, application data, appeal data, registration data, family data, and rating data for the patent in question.

[0057] The collection / refinement module 200 can collect public data, specifically economic statistical information such as sales growth rate data, sales statistics data, economic forecast data, stock price data, bond interest rate data, corporate financial data, and import / export data.

[0058] The collection / refining module 200 can extract necessary information from the raw information, as well as relevant statistical information and learning information for AI learning.

[0059] Specifically, the collection / refinement module 200 can extract statistical information from raw data, such as IPC specific patent citation lifetimes, industry-specific standard royalty rates, industry-specific self / debt capital costs, and industry-specific self / debt capital ratios.

[0060] Furthermore, the collection / refining module 200 can extract data such as patent concentration, number of appeals and lawsuits, number of prior IPs, number of technology transfers, number of licenses, depth of dependent claims of the target IP, number of claim series, number of opinions submitted, number of divisional applications, number of claims, number of citations, CAGR (Compound Annual Growth Rate) by industry, sales statistics by industry, sales growth rate by industry, export and import growth rate, economic forecast, and application growth rate by IPC as AI learning data.

[0061] The AI ​​module 300 may receive AI training data from the collection / purification module 200, preprocess it, and then use the training data to train an AI model for calculating core variables.

[0062] The AI ​​module 300 can calculate the core variables necessary for value assessment based on user input data through each AI model optimized through learning.

[0063] The valuation management module 400 can perform management functions for the IP valuation system 10.

[0064] Specifically, the valuation management module 400 can manage the reference information for performing IP valuation and can manage system users and the system. Furthermore, the valuation management module 400 can handle the payment process for using the IP valuation service.

[0065] The valuation database 500 can store collected raw information and information processed and extracted from that raw information.

[0066] The valuation database 500 can store compliance information data, patent information data, and economic statistics information data collected by the collection / purification module 200, as well as statistical data and AI training data as extracted information processed by the collection / purification module 200.

[0067] The modules 100 to 400 and the valuation database 500 will be described below with reference to Figures 2 to 10.

[0068] As illustrated in Figure 2, the valuation service module 100 may include a user interface unit 110, an attribute management unit 120, a value calculation unit 130, and a report generation unit 140.

[0069] Specifically, the user interface unit 110 provides an interface that allows the user to input information on the IP to be evaluated, and additionally allows the user to select or input information related to the business scale and industry of the business entity.

[0070] The attribute management unit 120 can manage attribute information such as the purpose of IP valuation and the method of IP valuation as IP valuation attributes. For example, the attribute management unit 120 can set up an IP-backed loan as the purpose of valuation and the royalty exemption method as the valuation method.

[0071] The value calculation unit 130 receives evaluation criteria data necessary for calculating value according to the evaluation method (royalty exemption method in this invention), and can calculate the IP value from the evaluation criteria data.

[0072] Specifically, the value calculation unit 130 can calculate the value of the target IP using the specific calculation formula shown in Equation 1.

[0073] The report generation unit 140 can generate an IP valuation report that includes not only the IP valuation results but also relevant statistical data calculated during the valuation process.

[0074] For example, the report generation unit 140 may first include the Technology Cycle Time (TCT) statistics for each IPC (International Patent Classification) of the target IPC (International Patent Classification) calculated in the process of estimating the economic life of the target IP.

[0075] Specifically, the report generation unit 140 can generate a report that includes quantile, median, mean, and tertile information as TCT statistics for the target IP, based on the patent citation lifetimes in Korea and the United States.

[0076] Furthermore, the report generation unit 140 includes statistical analysis information related to the IPC's judging process, which has been extracted and utilized as learning data for calculating the IP's economic lifespan, enabling the understanding of the competitive strength of the IP.

[0077] Similarly, the report generation unit 140 may include initial sales statistics for the business entity's industry, which are extracted and used for calculating sales, and may also include information on the number of applicants and the rate of increase or decrease in the number of applications of the IPC, as well as information on the rate of increase or decrease in exports and imports by item, which are used as training data, allowing for the confirmation of the IPC's industry sector trends, operating profitability, market growth trends, etc.

[0078] Furthermore, the report generation unit 140 can also provide royalty rate statistics for the relevant industry as a benchmark royalty rate used for calculating royalty rates, and can also provide statistics on the number of opinion papers submitted by the IPC, divisional applications, and domestic priority claims that are reflected in the learning data, which can be used to predict the stability of rights, research sustainability, and future development potential for the IP in question.

[0079] Furthermore, the report generation unit 140 can provide, as discount rate-related statistics for the industry, stock information, bond information, and financial information, along with information on the cost of debt and cost of equity, which are used to calculate the discount rate. The report generation unit 140 can also provide, for comparison with the overall industry average, information on the operating profit growth rate by size and sales CAGR for the industry, which are extracted for use as training data in calculating the discount rate, and can be used to forecast the operating profit stability or margin rate of the industry in question relative to the overall industry.

[0080] On the other hand, the aforementioned IPC is an internationally unified patent classification system that indicates the technical field of an invention, and is an example of IP classification information according to the present invention. IP classification information is not limited to IPC, but can also include CPC (Cooperative Patent Classification) and other patent classification systems. In this specification, IPC will also be used as an example to explain IP classification information below.

[0081] On the other hand, as illustrated in Figure 3, the collection / purification module 200 may include a raw data collection unit 210, a preprocessing unit 220, a basic data calculation unit 230, a training data calculation unit 240, a statistical data calculation unit 250, and an evaluation criteria data calculation unit 260.

[0082] The raw data collection unit 210 can collect source data used for IP valuation, but first, it can collect data on previously conducted expert IP valuations and actual IP transaction information as reference information that serves as the basis for valuation.

[0083] Furthermore, the raw data collection unit 210 can collect patent information from existing IP information provision systems, including IP-specific citation / cited data, application / trial / litigation / registration data, design right data, and family data.

[0084] The raw data acquisition unit 210 can then collect grading data from the IP grading system.

[0085] The raw data collection unit 210 can collect economic statistical information such as sales growth rate data, sales statistics data, economic forecast data, stock price data, bond interest rate data, corporate financial data, and import / export data.

[0086] The preprocessing unit 220 can preprocess the preferentially collected source data by passing and purifying it.

[0087] The base data calculation unit 230 can generate base data for AI training data and evaluation criterion data from the collected and pre-processed data, respectively.

[0088] The learning data calculation unit 240 can generate AI model training data from the basic data, which is used to estimate each predictor variable for calculating core variables for IP value evaluation.

[0089] For example, the learning data calculation unit 240 uses the following as learning data to input to the AI ​​model for calculating the economic life of an IP: expert evaluation results for the economic life of an IP, which are reference information, and the TCT for that IP. median The difference between the two (hereinafter also referred to as "factors influencing the economic lifespan of IP") can be used as a predictor variable, and the applicant / application growth rate, patent concentration, TCT statistics, grading factors, grading index scores, US litigation statistics by IPC, and appeal statistics by IPC can be used as explanatory variables to generate training data.

[0090] Furthermore, the learning data calculation unit 240 can generate learning data for the AI ​​model to calculate royalty rates by using the ratio between the expert evaluation royalty rate, which is compliant information, and the standard royalty rate for the industry in question (hereinafter also referred to as "factors influencing the standard royalty rate") as a predictor variable, and using standard royalty rate statistics, the number of opinions submitted, grading evaluation elements, grading evaluation index scores, the number of divisional applications, the number of priority claims, and the average depth of dependent claims as explanatory variables.

[0091] The learning data calculation unit 240 can generate learning data to be input into the AI ​​model for calculating the discount rate, using expert-reviewed IP commercialization risk premium as a predictor variable and explanatory variables such as sales growth rate and operating profit growth rate by company size / industry, grade evaluation elements, grade evaluation index scores, and patent concentration.

[0092] The learning data calculation unit 240 can generate learning data to be input into the AI ​​model for calculating sales, using the sales growth rate as the predictor variable and the increase / decrease rate of exports and imports by industry, the increase / decrease rate of IPC special applications, etc., as explanatory variables.

[0093] The statistical data calculation unit 250 can calculate relevant statistical data required during the training data generation or core variable calculation process.

[0094] The statistical data calculation unit 250 can provide statistical data for reference information, basic statistics for core variables, and statistics for each explanatory variable.

[0095] Specifically, the statistical data calculation unit 250 can calculate TCT statistics by IPC, statistical analysis information related to appeals, initial sales statistics of business entities, number of applicants and application growth / decrease rates by IPC, export / import growth / decrease rates by item obtained through IPC and export / import code matching, self / debt capital costs by industry size, self / debt capital ratios by industry size, patent concentration by IPC, statistics on sales / operating profit growth rates by industry size, industry-specific standard royalty rate statistics, number of opinions submitted by IPC, and statistics on divisional applications and priority claims.

[0096] According to the present invention, statistical information for each core variable can be calculated by reflecting the predictor variable estimation results obtained through two or more AI models for each core variable.

[0097] In this case, the statistical data calculation unit 250 may calculate statistical information for the estimated predictor variables or statistical information for each core variable and provide it to the user through a report.

[0098] Therefore, users can utilize not only the median of the core variable or its related predictor variables, but also quantiles, tertiles, and mean information.

[0099] The evaluation criteria data calculation unit 260 can ultimately calculate the evaluation criteria data to be input into the value evaluation formula based on the calculated statistical data and core variables.

[0100] Specifically, the evaluation criteria data calculation unit 260 can confirm the legal remaining lifespan of the target IP and compare it with the IP economic lifespan calculated via the AI ​​module.

[0101] The evaluation criteria data calculation unit 260 can determine an even shorter remaining lifespan as the final economic lifespan of the target IP based on the comparison results, or, if a commercialization preparation period is required, it can calculate the final economic lifespan of the target IP by reflecting that period.

[0102] Furthermore, the valuation criteria data calculation unit 260 can determine the corporate tax rate and finalize the corporate tax expense. Based on the sales revenue during the final economic lifespan of the target IP, the valuation criteria data calculation unit 260 can apply the determined corporate tax rate and calculate the final corporate tax expense.

[0103] The AI ​​module 300, which calculates the core variables used in the IP valuation, will be explained in detail below with reference to Figures 4 to 9.

[0104] As illustrated in Figure 4, the AI ​​module 300 may include a training data preprocessing unit 310, an AI training unit 320, a training optimization unit 330, and a core variable calculation unit 340.

[0105] The training data preprocessing unit 310 can perform preprocessing on the training data transmitted from the training data calculation unit 240.

[0106] Specifically, the training data preprocessing unit 310 can perform duplicate row processing, outlier processing, and normalization processes for crystalline AI models, and duplicate row processing and outlier processing for generative AI models.

[0107] The AI ​​learning unit 320 includes two or more AI models and can perform training on the AI ​​models via the training data.

[0108] Specifically, the AI ​​learning unit 320 includes one or more crystalline AI models and generative AI models, and can utilize training data to train an AI model for predictor variable estimation for calculating core variables.

[0109] For example, the AI ​​learning unit 320 can use AutoML (Automated Machine Learning) as its crystalline AI model, and may be composed of a stacking ensemble model.

[0110] As illustrated in Figure 5, a stacking ensemble model includes base models (level-0 models) and a metamodel (level-1 model, the final model), and can use the predictions of multiple base models as training data for the final model to make predictions.

[0111] According to the present invention, for example, the base model may include a statistical base model (model 1), a tri-based model (model 2), and a neural network model (model 3).

[0112] In this case, the statistics-based model can include a K-Nearest Neighbors model. The tribe-based model can include Decision Tree, Random Forest, Extra Tree, XGBoost, LightGBM, and CatBoost models. Furthermore, the neural network model can include a Multilayer Perceptron model.

[0113] At this time, the AI ​​learning unit 320 can perform training and prediction for all of the aforementioned models.

[0114] The AI ​​learning unit 320 can train a base model using an input dataset, and then use the predicted values ​​generated through the trained base model as input data for a metamodel to generate final predicted values ​​and perform model training.

[0115] Specifically, the AI ​​learning unit 320 can train each base model with default hyperparameters and evaluate the learning performance of various combinations of hyperparameters through random sampling from a predefined set of hyperparameters.

[0116] The AI ​​learning unit 320 can generate a weighted average of the predicted values ​​of each learned base model, utilize the weighted average predicted value as new input data, and generate a final predicted value through the metamodel.

[0117] At this time, the AI ​​learning unit 320 generates multiple predicted values ​​by differentiating the characteristics of the learning data, the learning evaluation index, and the initial setting value (seed), and the final predicted value can be selected as the median.

[0118] For example, the AI ​​learning unit 320 can generate 27 different predicted values ​​by varying three training datasets, three learning evaluation metrics, and three initial settings, and can select the dual median as the final predicted value.

[0119] In this case, for example, the three initial values ​​can be set to 1, 2, and 3 using arbitrary random values, and the three evaluation metrics can include RMSE (Root Mean Square Error), MAPE (Mean Absolute Percentage Error), and R² (R-Squared).

[0120] Alternatively, the AI ​​learning unit 320 may include generative models and can use Bayesian neural network models (BNNs), sparse Gaussian process models (Sparse GPs), and variational sparse GPs based on marginal inference.

[0121] Specifically, in the case of BNNs, the weights of the hidden layer are defined as latent variables, and in this case, the latent variables are random variables with an arbitrary distribution. The dataset used for training is also a random variable with an arbitrary distribution, but below, the arbitrary distribution of the above-mentioned set of explanatory variables can be described as a joined distribution coupled with the latent variables.

[0122] According to one aspect of the present invention, the joint distribution to be estimated is set to a distribution in which explanatory variables and latent variables are coupled, and learning can be carried out in a way that minimizes the difference between any candidate distributions that are easy to use for estimating this joint distribution. The learning method can, for example, be carried out in a way that maximizes the Evidence Lower Bound (ELBO) and minimizes the Kullback-Leibler (KL) distance. In this case, ELBO represents the expected value of the difference between the joint distribution to be estimated and the candidate distribution. The KL distance represents the conditional distribution for the latent variables, i.e., the difference between the joint distributions, when the candidate distribution and explanatory variables are fixed, and can represent the difference between the distribution value generated by the candidate distribution and the distribution value generated by the latent variable distribution value based on the given learning data.

[0123] For the aforementioned learning, the Monte Carlo method can be used. Distribution values ​​can be generated from the candidate distribution and the connected distribution, respectively. The sum of these distribution values ​​can be defined as the expected value. The parameters of the connected distribution can be fixed, and the parameters of the candidate distribution can be found to maximize the expected value. Next, the parameters of the connected distribution can be fixed, and the parameters of the connected distribution that minimize the KL distance can be found. The process of fixing the parameters of the connected distribution again and retrieving the parameters of the candidate distribution that maximize the ELBO can be repeated, thereby learning the connected distribution.

[0124] Through the process described above, when each parameter no longer changes or a predetermined number of iterations are reached, the learning process can be terminated, and a number of predicted values ​​can be generated that are defined by a combined distribution composed of the learned parameters.

[0125] Next, in the case of a sparse Gaussian process (Sparse GP), GP, as a distribution for the function, means a distribution constructed from any function for the given explanatory variables. The said distribution has the form of a multivariate normal distribution with a mean function and a covariance function, but since the GP is used as a prior distribution for predictor variable inference, the mean function 0 and the covariance function can be assumed to be any function constructed from the explanatory variables. In this case, the covariance function can be assumed to be a radial kernel function. The radial kernel function may consist of a function that shows the relationships between the data and parameters that represent the characteristics of the given data.

[0126] In this case, given explanatory variables, each predictor variable is a distribution generated by a predefined GP, consisting of the GP of the explanatory variable and an arbitrary error. Therefore, in the GP as well, it is necessary to estimate the parameters of the kernel function and an arbitrary error that maximizes the distribution value of the predictor variable. A value is generated by adding an arbitrary error assuming a normal distribution to the GP for the explanatory variable using the Monte Carlo method, and the parameters can be estimated by numerical interpretation of the generated value. In this case, if the size of the data is enormous, a lot of computation is required, so by setting points in the data space where the data is examined, the computational load can be reduced and high-speed learning can be performed.

[0127] Furthermore, the Variational Sparse GP (Variational Sparse GP) is a GP that does not follow the normal distribution assumption for arbitrary errors used in the GP assumptions described above, but instead assumes an arbitrary distribution. In other words, it is a model that assumes that any error of the predictor variable follows an arbitrary distribution.

[0128] Therefore, to assume the aforementioned arbitrary distribution, the Monte Carlo method, which maximizes ELBO and minimizes KL distance—a method used during BNN training—can be used to estimate the arbitrary distribution. Furthermore, as explained in Sparse GP for inference on massive amounts of data, the computational complexity can be reduced and training can be performed quickly by reducing the number of computation points.

[0129] According to one aspect of the present invention, the AI ​​learning unit 320 includes all of the above-mentioned Bayesian neural network models (BNN), rare Gaussian process models (Sparse GP), and variational inference-based rare Gaussian process models (Variational Sparse GP) as generative AI models, and can generate 10 predicted values ​​for each model, for a total of 30 predicted values.

[0130] In this case, the AI ​​learning unit 320 includes, for example, the stacking ensemble model and the generative AI model described above, and can generate 27 predicted values ​​via the stacking ensemble model and 30 predicted values ​​via the generative model, for a total of 57 predicted values, of which the median can be used as the final predicted value.

[0131] The AI ​​learning unit 320 can then train and validate all of the stacking ensemble models and generative AI models on the overall training dataset in a ratio of, for example, 9:1, and proceed with learning.

[0132] On the other hand, the AI ​​learning unit 320 can learn the stacking ensemble model or generative AI model as a crystal model based on learning data that have differences in their input variables and are different from each other.

[0133] For example, according to one aspect of the present invention, the AI ​​model can be trained using a set of explanatory variables A that correlates with the predictor variable with a Pearson correlation coefficient significance level of less than 0.05, a set of explanatory variables B that also includes statistical values ​​of the correlated variables, and a set of explanatory variables C that reflects variables that do not correlate and includes all grade evaluation elements of the grade evaluation system.

[0134] The more information reflected in model training, the more complex the model becomes, potentially leading to decreased predictive performance. Conversely, insufficient information reflected in model training can also lead to decreased predictive performance. Therefore, in this invention, both information loss and parsimony are taken into consideration, and variable sets can be divided and compared.

[0135] More specifically, in calculating the first core variable, the AI ​​learning unit 320 can utilize the explanatory variables shown in Figure 6 to predict the factors influencing the economic lifespan of the target IP, which is the first predictor variable. More specifically, it can utilize three sets of explanatory variables.

[0136] In this case, each set of explanatory variables may commonly include the applicant application growth rate, industry-specific HHI (Herfindahl-Hirschman index), grade evaluation system scores, TCT statistics, and US litigation or appeal statistics by IPC.

[0137] Furthermore, explanatory variable set A may include, among prior art, the number of articles, the number of foreign patents, the total number of citations, the length of independent clauses, and the number of independent clauses; explanatory variable set B may further include, among prior art by IPC, the number of foreign patents, the total number of citations, the length of independent clauses, the number of independent clauses, and among prior art by IPC, the average number of foreign patents, the average total number of citations by IPC, the average length of independent clauses by IPC, and the average number of independent clauses by IPC; and explanatory variable set C may further include all evaluation elements of the grading system.

[0138] At this time, the aforementioned grading system (for example, SMART5) uses the following information about the target IP: specification information (number of independent claims, length of independent claims, average depth of dependent claims, length of description of the invention, number of dependent claims, number of claim sequences), surge information (number of IPCs, number of drawings, number of divisional applications / priority claims, number of inventors), examination information (whether or not early publication was made, whether or not priority examination was requested, number of information provided, number of opinions submitted), and post-registration administrative information (number of annual registrations, number of changes in rights holders, number of overseas family countries, number of licensees, number of pledges established by financial institutions, and registration of extension of term). This is a grading system based on the presence or absence of a patent, litigation trial information (number of invalidation trials dismissed, number of invalidation trials cited / withdrawn / rejected, number of negative scope confirmation trials dismissed, number of negative scope confirmation trials cited / withdrawn / rejected, number of positive scope confirmation trials dismissed / withdrawn / rejected, number of positive scope confirmation trials cited, number of appeals against rejection decisions, and correction trials), and citation information (total number of citations, number of non-patent literature / foreign patents among the cited documents for cited patents, difference between filing date and cited date, and number of papers / foreign patents in prior art).

[0139] The grading system assigns points to the target IP based on the values ​​for each of the 32 evaluation elements, and assigns a grade according to the score interval.

[0140] Furthermore, in calculating the second core variable, the AI ​​learning unit 320 can utilize the explanatory variables shown in Figure 7 to predict the factors influencing the second predictor variable, the standard loyalty rate. More specifically, it can utilize three sets of explanatory variables.

[0141] In this case, each set of explanatory variables may commonly include the number of divisional applications / priority claims, the number of opinions submitted, sales growth rates by industry, royalty rate statistics, IPC-specific appeal statistics, and evaluation scores from the grading system.

[0142] Furthermore, explanatory variable set A may further include the difference in filing dates of cited applications, average depth of dependent claims, number of independent claims, length of independent claims, and TCT statistics. Explanatory variable set B may further include the difference in filing dates of cited applications, average depth of dependent claims, number of independent claims, length of independent claims, TCT statistics, average number of priority claims for divisional applications per IPC, average number of opinions submitted per IPC, average difference in filing dates of cited applications per IPC, average average depth of dependent claims per IPC, average number of independent claims per IPC, and average length of independent claims per IPC. Furthermore, explanatory variable set C may further include all evaluation elements of the grading system.

[0143] Furthermore, in calculating the third core variable, the AI ​​learning unit 320 can utilize the explanatory variables shown in Figure 8 as the third predictor variable for predicting the IP commercialization risk premium, and more specifically, it can utilize three sets of explanatory variables.

[0144] In this case, each set of explanatory variables can commonly include sales growth rates by company size and industry, operating profit growth rates by company size and industry, patent concentration, and evaluation scores from a grading system.

[0145] Furthermore, explanatory variable set A may further include the length of independent terms, the number of overseas patent citations of the cited patent, industry-specific business activity indices, and IPC-specific appeal statistics. Explanatory variable set B may further include the length of independent terms, the number of overseas patent citations of the cited patent, industry-specific business activity indices, the average length of independent terms by IPC, the average number of overseas patent citations of the cited patent by IPC, and IPC-specific appeal statistics. Furthermore, explanatory variable set C may further include all evaluation elements of the grading system.

[0146] On the other hand, the AI ​​learning unit 320 can include either an ARIMA (Autoregressive Integrated Moving Average) model or an ETS (Exponential Smoothing) model in the calculation of the fourth core variable, and can perform learning for estimating the sales growth rate, which is the fourth predictor variable.

[0147] The AI ​​learning unit 320 can utilize the explanatory variables shown in Figure 9 when calculating the fourth core variable. More specifically, it can incorporate information on the number of applicants and application growth / decrease rates by IPC, sales growth rates by IPC, and export / import growth / decrease rates into the learning data to train a model for estimating sales growth rates. Furthermore, the training set and validation set can be divided into an 8:2 ratio for training.

[0148] In this case, the ARIMA model can include ARIMA, SARIMA (Seasonal ARIMA), ARIMAX (Autoregressive Integrated Moving Average Exogenous), and SARIMAX (Seasonal Autoregressive Integrated Moving Average Exogenous) models.

[0149] Furthermore, the ETS model can include Holt-Winter's seasonal technique, Holt-Winters damped technique, damped trend technique, and SES (Simple Exponential Smoothing).

[0150] The AI ​​learning unit 320 can evaluate the model's performance through the absolute value of the CAGR difference between predicted and actual values ​​obtained via the validation set, the mean absolute error (MAE), and the mean squared error (MSE).

[0151] The learning optimization unit 330 can set the optimal parameter combination for each AI model for calculating core variables based on the learning results for the AI ​​model and the validation results for the model predictions of the AI ​​learning unit 320.

[0152] The learning optimization unit 330 can also remove low-importance training data features based on the performance evaluation of each base model in the stacking ensemble model, and perform additional hyperparameter tuning on the best-performing model.

[0153] Furthermore, the learning optimization unit 330, while each model parameter in the ARIMA model has AR order p, MA order q, difference d, seasonal AR order P, seasonal MA order Q, and seasonal difference D, can derive the combination that minimizes AICc from among the possible parameter combinations using the modified AKAIKE information criterion (AICc), and select the optimal parameter combination.

[0154] Furthermore, the learning optimization unit 330 has parameters for each model in the ETS model, such as alpha (smoothing_level), beta (smoothing_trend), initial_level (l_0), initial_trend (b_0), gamma (seasonality), phi (damping_trend), s_0, s1, s2, and s3 (initial_seasons). The L-BFGS-B method (quasi-Newton method) can be used to fit these parameters to the training dataset.

[0155] Specifically, the learning optimization unit 330 sets the following as optimized explanatory variables for predicting factors influencing the economic lifespan of the target IP, which is the first core variable: the rate of increase or decrease in the number of applicant applications (rate of increase or decrease in applications / rate of increase or decrease in applicants), TCT statistics, evaluation elements of the grading system, evaluation index scores of the grading system, the average number of US patent lawsuits by IPC, and the average number of appeal-related cases by IPC, as the first set of explanatory variables, and can set up two or more AI models optimized according to the performance index results.

[0156] Furthermore, for example, the learning optimization unit 330 can set royalty rate statistics, the average number of opinion submissions per IPC, evaluation elements of the grading system, evaluation index scores of the grading system, and the number of priority claims for divisional applications per IPC as the second set of explanatory variables in order to calculate the royalty rate of the target IP, which is the second core variable, and set up two or more AI models optimized according to the performance index results.

[0157] Furthermore, the learning optimization unit 330 can set up two or more AI models optimized according to the performance indicator results as third explanatory variables for predicting the IP commercialization risk premium, such as the sales growth rate by company size and industry, the operating profit growth rate by company size and industry, the evaluation elements of the grading evaluation system, the evaluation index scores of the grading evaluation system, and the patent concentration, as third explanatory variables for calculating the discount rate, which is the third core variable.

[0158] Furthermore, the learning optimization unit 330 can set the number of applicants, the rate of increase or decrease in the number of applications, and the rate of increase or decrease in exports and imports as the fourth set of explanatory variables for estimating the sales growth rate, which is the fourth core variable, and can set up two or more optimized AI models according to the performance indicator results.

[0159] The core variable calculation unit 340 can calculate each core variable for valuing the target IP.

[0160] The core variable calculation unit 340 can generate predictor variables through two or more AI models optimized for each core variable, and calculate each core variable based on those predictor variables.

[0161] Specifically, the core variable calculation unit 340 can calculate the values ​​of the first to fourth core variables, but it can prioritize the calculation of the first to fourth predictor variable values ​​for each of the core variables.

[0162] In other words, the core variable calculation unit 340 can check the first set of explanatory variables and calculate the first predictor variable value, and then calculate the first core variable value based on the first predictor variable value.

[0163] Using the same method, the core variable calculation unit 340 can sequentially calculate the second predictor variable and the second core variable value via the second explanatory variable set, and calculate the third predictor variable and the third core variable value via the third explanatory variable set. Then, the core variable calculation unit 340 can check the fourth explanatory variable set and calculate the fourth predictor variable and the fourth core variable value.

[0164] On the other hand, if the core variable calculation unit 340 obtains two or more first to fourth predictor variable values ​​for each of the first to fourth core variables through two or more AI models, it may select specific values ​​according to the first to fourth criteria for each, or obtain statistically processed values ​​for each result.

[0165] More specifically, the first core variable relating to one aspect of the present invention is the economic lifespan of the target IP.

[0166] The core variable calculation unit 340 can calculate the factors influencing the IP economic lifespan as a first predictor variable via the first explanatory variable set and AI model set by the learning optimization unit 330.

[0167] The core variable calculation unit 340 can check the TCT median of the IPC to which the target IP belongs, and by reflecting the first predictor variable value in the TCT median, it can calculate the first core variable, i.e., the economic lifespan of the target IP.

[0168] Furthermore, a second core variable relating to one aspect of the present invention is the royalty rate.

[0169] The core variable calculation unit 340 can calculate the factors influencing the loyalty rate as a second predictor variable via the second explanatory variable set and AI model set by the learning optimization unit 330.

[0170] The core variable calculation unit 340 can set the industry based on industry classification information matched to the IP classification information of the target IP, confirm the standard royalty rate value for that industry, and calculate the second core variable, i.e., the final royalty rate of the target IP, by reflecting the second predictive variable value in the standard royalty rate value.

[0171] Furthermore, a third core variable relating to one aspect of the present invention is the discount rate.

[0172] The core variable calculation unit 340 can calculate the IP commercialization risk premium as a third predictor variable via the third explanatory variable set and AI model set by the learning optimization unit 330.

[0173] The core variable calculation unit 340 can calculate the weighted average cost of capital (WACC) by reflecting the IP commercialization risk premium, which is the third predictor variable value, in the cost of equity capital, and can then calculate the discount rate, which is the third core variable, from the weighted average cost of capital. The specific formula is as follows:

number

[0174] At this time, K d This is the cost of external capital, K e T is cost of equity, T is corporate tax rate, E is equity, D is debt.

number

number

[0175] On the one hand, the cost of equity can be calculated through the listed company's CAPM (Capital Asset Pricing Model). However, in the case of unlisted companies, the size risk premium can be added. The core variable calculation unit 340 adds the calculated IP commercialization risk premium to calculate the cost of equity (K e ) as follows and can calculate the final discount rate according to the formula 2.

Formula

[0176] At this time, the cost of equity of a listed company, that is, CAPM, is as follows.

Formula

[0177] The statistical data calculation unit 250 can calculate the market risk premium, etc. based on financial information such as stock price data, bond interest rate data, and corporate financial data. For example, recently, the arithmetic average of the stock price index return rate in the past year is calculated to obtain the market expected return rate (E(R m )), and the average of the 5-year maturity Treasury bond yield is calculated as the risk-free interest rate (R f ) from the expected return rate.

[0178] The statistical data calculation unit 250 can calculate the market risk premium by taking the difference between the market expected return rate (E(R m )) and the risk-free interest rate (R f ).

[0179] The statistical data calculation unit 250 can then calculate the correlation coefficient between stock index returns and individual stocks, calculate it by industry, and calculate beta.

[0180] On the other hand, the statistical data calculation unit 250 can calculate the cost of debt capital by adding an additional risk spread to the cost of debt capital of listed companies in the relevant industry.

[0181] Specifically, the statistical data calculation unit 250 can calculate the financial costs of listed companies based on financial information such as stock price data, bond interest rate data, and corporate financial data, and can calculate the additional risk spread of unlisted companies compared to the average credit grade of listed companies by utilizing the average credit grade spread of unsecured corporate bonds, and can calculate the cost of debt capital.

[0182] A fourth core variable relating to one aspect of this invention is sales revenue.

[0183] The core variable calculation unit 340 can calculate the sales growth rate as a fourth predictor variable via the fourth explanatory variable set and AI model set by the learning optimization unit 330.

[0184] The core variable calculation unit 340 can set initial sales figures according to the industry, and can calculate sales figures during the economic lifespan by reflecting the fourth predictive variable in the initial sales figures.

[0185] Specifically, the core variable calculation unit 340 can set initial sales revenue through the average past sales of a business entity if it has identified one. If the core variable calculation unit 340 cannot identify a business entity, it can set initial sales revenue through sales statistics corresponding to the industry, size, or product category.

[0186] The core variable calculation unit 340 can calculate the sales flow for the period by reflecting the sales growth rate calculated in the initial sales.

[0187] On the other hand, as illustrated in Figure 10, the valuation management module 400 may include a reference information management unit 410, a user management unit 420, a settlement management unit 430, and a valuation system management unit 440.

[0188] The standard information management unit 410 can manage evaluation standard information necessary for performing IP valuation, such as relevant laws and regulations including tax rates.

[0189] The user management unit 420 can manage basic information and historical information for users of the system according to the present invention, and the payment management unit 430 can manage the payment of expenses for using the system according to the present invention. Furthermore, the value evaluation system management unit 440 can manage the resources of the system according to the present invention.

[0190] Figure 11 is a detailed configuration diagram of a valuation database 500 according to one aspect of the present invention. As shown in the figure, the valuation database 500 can include raw information and extracted information collected and processed by the collection / purification module 200.

[0191] First, raw information may include compliance information data, patent information data, and economic statistics information data.

[0192] Compliance information data is information that is used as a standard in the valuation process of the IP valuation system or as a verification standard for system evaluation results using AI models, and may include IP evaluation information and transaction information that has already been completed.

[0193] In this case, the valuation information may include, for example, expert-valued IP economic lifespan, expert-valued royalty rate, and expert-valued IP commercialization risk premium as IP valuation data completed by experts. The transaction information may include, for example, actual IP transaction data such as technology transfer data from universities, technology transfer data from public research institutes, and IP transaction data from IP exchanges.

[0194] On the other hand, patent information data may include patent detail information and grading information as objective data related to patents.

[0195] Patent detail information may include patent-specific citation / cited data, application data, appeal data, litigation data, registration data, and family data. The grading information may include patent-specific grading data, such as the grading element data described above, and score data for the indicators that each grading element represents, according to the evaluation results for each element.

[0196] Economic statistical information data may include economic market information, financial information, and import / export information necessary for IP valuation, among the public data generated or acquired and managed by public institutions.

[0197] In this case, economic market information can include industry-specific sales growth rate data, sales statistics data, and economic forecast data. Financial information can include stock price data, bond interest rate data, and corporate financial data. And import / export information can include import / export data.

[0198] Next, the extracted information is information extracted by processing raw information and may include statistical data calculated in the IP valuation process and provided along with the final valuation result for the target IP, as well as AI training data used for AI model training.

[0199] In this case, the statistical data may include information related to the economic lifespan of intellectual property (IP), royalty rates, discount rates, and sales revenue.

[0200] Specifically, statistical information used or extracted in the process of calculating the IP economic life may include, for example, patent citation lifetime (TCT) data by IPC, appeal-related statistical data, US litigation data, and market concentration data for the relevant industry sector (sector).

[0201] Furthermore, the statistical information used or extracted in the royalty rate calculation process may include industry-specific standard royalty rate data, the number of opinions submitted by each IPC, the number of divisional applications, and the number of priority claims.

[0202] Furthermore, statistical information used or extracted in the discount rate calculation process may include industry-specific self / debt capital cost data and industry-specific self / debt capital ratio data, patent concentration by target IPC, and the most recent 5-year sales / operating profit CAGR data by company size and industry for the target IP business sector.

[0203] Furthermore, statistical information used or extracted in the sales calculation process may include initial sales statistics data corresponding to the industry and size to which the IP of the business entity belongs, statistical data on the number of applicants and the rate of increase or decrease in the number of applications of the IPC, and statistical data on the rate of increase or decrease in exports and imports for the industry and product.

[0204] On the other hand, the AI ​​training data used to train the AI ​​model can include training data input to the AI ​​model for calculating the first core variable, the IP economic lifespan; training data input to the AI ​​model for calculating the second core variable, the royalty rate; training data input to the AI ​​model for calculating the third core variable, the discount rate; and training data input to the AI ​​model for calculating the fourth core variable, the sales revenue.

[0205] The specific examples of the training data used to calculate the first to fourth core variables are the same as those previously described, so we will omit further details.

[0206] The following outlines and explains the overall flow of the process executed by the IP valuation system according to one aspect of the present invention, with reference to Figure 12.

[0207] First, the user can input information about the IP to be evaluated via the valuation service module 100. For example, the user can input the registration number of the patent to be evaluated.

[0208] The AI ​​module 300 can calculate the first to fourth core variables for valuing the target IP, and based on the registration number that is entered preferentially, it can check the patent classification information of the patent from the patent information in the valuation database 500 and check the TCT statistics of the patent classification information from the statistical information in the valuation database 500.

[0209] The AI ​​module 300 can retrieve the first set of explanatory variables for the target IP from the valuation database 500, and calculate the factors influencing the first predictor variable, the IP's economic lifespan, through the configured first set of explanatory variables and the AI ​​model.

[0210] The AI ​​module 300 can check the TCT median of the patent classification information and calculate the first core variable, the IP economic life, by reflecting the first predictor variable value in the TCT median. At this time, the collection / refinement module 200 can finally calculate the shorter of the two as the final IP economic life (S100).

[0211] The AI ​​module 300 can identify the industry by checking the industry classification information that matches the patent classification information. Then, the AI ​​module 300 can check the sales statistics for that industry from the economic statistics information of the valuation database 500 (S101, S103, S105).

[0212] When the AI ​​module 300 identifies the size and industry of a business entity from user input information and economic statistics information from the valuation database 500, it can proceed with calculating core variables based on the identified size and industry (S107, S113).

[0213] The AI ​​module 300 can retrieve the second explanatory variable set values ​​for the target IP from the valuation database 500, and calculate the influencing factors for the second predictor variable, the loyalty rate, through the configured second explanatory variable set and AI model.

[0214] The AI ​​module 300 can check the standard royalty rate for the industry in question from the valuation database 500, and calculate the final royalty rate, which is the second core variable, by reflecting the second predictor variable value in the standard royalty rate (S117).

[0215] The AI ​​module 300 can then check the values ​​of the third explanatory variable set for the target IP from the valuation database 500, and calculate the IP commercialization premium, which is the third predictor variable, through the configured third explanatory variable set and the AI ​​model.

[0216] The AI ​​module 300 can calculate the weighted average cost of capital (WACC) as the final discount rate by reflecting the value of the third predictor variable in the cost of equity for the industry or size to which the target IP belongs, calculated from the economic statistical information of the valuation database 500 (S115).

[0217] The AI ​​module 300 then checks the values ​​of the fourth explanatory variable set for the target IP from the valuation database 500 based on the mapping between IP classification information and import / export item classification information, and calculates the sales growth rate, which is the fourth predictor variable, through the configured fourth explanatory variable set and AI model.

[0218] When a business entity is identified and sales are confirmed, the AI ​​module 300 can set an initial sales amount using the average past sales amount of the business entity, and then calculate the sales amount for the period by reflecting the fourth predictive variable in the initial sales amount (S109, S111, S121).

[0219] On the other hand, if the AI ​​module 300 cannot identify sales, it may set initial sales using sales statistics of a specific size (e.g., small-scale) company already established in the industry, and calculate sales reflecting the fourth predictor variable (S109, S119, S121).

[0220] The collection / refinement module 200 can then calculate the final corporate tax expense based on the estimated sales during the economic life of the target IP, the final royalty rate, and the corporate tax rate based on sales (S123).

[0221] As previously described, the valuation service module 100 can perform a valuation of the subject IP according to the aforementioned formula 1, based on the final IP economic lifespan, final royalties, final discount rate, estimated sales revenue, and final corporate tax value (S125).

[0222] The valuation service module 100 can then generate a report that includes the valuation results, the reference information extracted or utilized in the valuation process, and statistical information for the core variables or various explanatory variables used to estimate the core variables (S127).

[0223] On the other hand, with reference to Figures 13 and 14, the portfolio valuation process of the IP valuation system according to one aspect of the present invention will be summarized and explained.

[0224] First, the user can input information about the IP portfolio to be evaluated via the valuation service module 100. For example, the user can input the registration number of the patent to be evaluated.

[0225] The AI ​​module 300 can calculate the first to fourth core variables for a portfolio in order to value the target IP portfolio, and based on the preferentially entered registration number, it can check the patent classification information for each individual patent from the patent information in the valuation database 500, and check the TCT statistics for the patent classification information of the individual patent from the statistical information in the valuation database 500.

[0226] The AI ​​module 300 can set the average of the median TCT values ​​for each individual patent as the baseline TCT value for the portfolio (S201).

[0227] At this time, the AI ​​module 300 can calculate the factors influencing the IP economic lifespan for each individual patent constituting the portfolio in the manner described above. The AI ​​module 300 can then estimate the factor with the largest value among the calculated values ​​as the first predictor variable for the target portfolio, i.e., the factor influencing the economic lifespan of the target portfolio (S203).

[0228] The AI ​​module 300 can reflect the value of the first predictor variable for the portfolio in the reference TCT value for the portfolio and calculate the first core variable for the portfolio, i.e., the IP economic lifetime for the portfolio (S205).

[0229] The collection / refinement module 300 can then calculate the final IP economic life of the portfolio as the shorter of the remaining legal lifespan of each individual patent (S207).

[0230] In other words, if the IP economic life calculated in S205 is shorter than the shortest remaining life, the calculated IP economic life is reflected as is. If the IP economic life is longer than the shortest remaining life, the shortest remaining life replaces the IP economic life calculated in S205.

[0231] On the other hand, if the AI ​​module 300 identifies the size and industry of a business entity from user input information or economic statistics information from the valuation database 500, it can proceed with calculating core variables for the portfolio based on the identified size and industry (S301, S303).

[0232] Furthermore, if the business entity cannot be identified, the AI ​​module 300 can identify each industry using industry classification information matched to each IP classification information of individual IPs, assume the company size is small, and check and compare the median sales of small companies in each industry from the valuation database 500 (S301, S305, S307). The AI ​​module 300 can then set the industry with the largest median sales as the representative industry of the IP portfolio (S309).

[0233] Furthermore, the AI ​​module 300 can check the standard royalty rate for representative industries of the IP portfolio from the valuation database 500 and set it as the standard royalty rate for the IP portfolio (S401).

[0234] The AI ​​module 300 may calculate the factors influencing the royalty rate for each individual patent constituting the portfolio using the method described above. The AI ​​module 300 can then estimate the factor with the largest value among the calculated values ​​as the second predictor variable for the target portfolio, i.e., the factor influencing the royalty rate of the target portfolio (S403).

[0235] The AI ​​module 300 can calculate the final loyalty rate of the portfolio by reflecting the portfolio's loyalty rate influencing factors in the portfolio's baseline loyalty rate (S405).

[0236] On the other hand, the AI ​​module 300 can calculate the weighted average cost of capital for a portfolio using the cost of equity / ratio and cost of debt / ratio for representative industries of the portfolio, which are calculated from the economic statistical information in the valuation database 500 (S501).

[0237] At this time, the AI ​​module 300 can calculate the IP commercialization premium for each individual IP in the manner described above, and the smallest of these IP commercialization premiums can be estimated as the third predictor variable for the target portfolio, i.e., the IP commercialization premium of the target portfolio (S503).

[0238] The AI ​​module 300 can calculate the final discount rate of the portfolio by reflecting the estimated IP commercialization premium of the portfolio in the cost of equity, which is part of the weighted average cost of capital for the portfolio industries (S505).

[0239] Furthermore, while the AI ​​module 300 can set initial sales, if a business entity is identified and its past sales are confirmed from the economic statistics information in the valuation database 500, the average of past sales (for example, the average of past sales over the last 3 to 5 years) can be assumed as the initial sales for the year immediately preceding the valuation date (S601).

[0240] Furthermore, the AI ​​module 300 can estimate the sales growth rate of the target portfolio, which is the fourth predictor variable for the target portfolio, in the manner described above, based on IP classification information and import / export item classification information mapping based on the representative industry of the portfolio (S603).

[0241] On the other hand, if there is no business entity or sales information, the AI ​​module 300 can set initial sales figures using sales statistics of a specific size (e.g., small-scale) company pre-configured in the portfolio's representative industry.

[0242] The AI ​​module 300 can calculate sales for a set period, reflecting the portfolio's sales growth rate in the initial sales (S605).

[0243] On the other hand, if the sales growth rate is estimated to be a quarterly growth rate, the AI ​​module 300 can convert the quarterly sales growth rate to an annual growth rate using a geometric mean, and similarly, the overall period is limited to the economic lifespan of the IP.

[0244] The collection and refinement module 200 can then determine the corporate tax rate based on the sales revenue over the economic lifespan of the IP portfolio, for example, based on the average sales revenue over the entire period (S701).

[0245] The data collection and refinement module 200 can reflect the portfolio's estimated sales, final royalty rate, and corporate tax rate, and calculate corporate tax expense (S703).

[0246] The valuation service module 100 can calculate the final value of an IP portfolio according to the aforementioned formula 1, based on the final IP economic lifespan, final royalties, final discount rate, estimated sales revenue, and final corporate tax value for the portfolio (S705).

[0247] Furthermore, the valuation service module 100 can generate a report that includes not only the valuation results, but also compliance information extracted or utilized in the valuation process for the IP portfolio, and statistical information for the core variables or various explanatory variables used to estimate the core variables.

[0248] Therefore, according to the present invention, it is possible to perform a rapid, objective, and efficient valuation of the entire portfolio without being affected by the number of individual IPs included in the portfolio.

[0249] Furthermore, according to the present invention, since IP valuation and statistical information provision can be performed by utilizing patent information and economic statistical information in an integrated manner, the objectivity and reliability of the valuation results can be improved.

[0250] Furthermore, according to the present invention, since the user is provided not only with direct IP valuation results but also with related patents and industry statistics, it can provide significant usefulness to the user's understanding of the target IP-related industry, and to the interpretation and utilization of the valuation results.

[0251] In particular, according to the present invention, since statistical information is provided not only for the reference information but also for the calculated core variables and various explanatory variables used in estimating the core variables, the reliability and applicability of the evaluation results can be enhanced.

[0252] Furthermore, according to the present invention, in calculating each core variable for IP valuation, instead of qualitative evaluation indicators conventionally used in expert evaluations, objective statistical data based on patent information and economic statistics is discovered and utilized, thereby improving the objectivity of the evaluation results.

[0253] Furthermore, according to the present invention, raw information can be continuously collected, processed to generate new statistical information and AI learning information, and utilized as training data for AI models for calculating core variables, thereby efficiently ensuring the timeliness and relevance of relevant information used in the IP valuation process.

[0254] Furthermore, according to the present invention, when calculating core variables, the estimation results of two or more AI models optimized for each core variable can be utilized, thereby increasing the objectivity and reliability of the evaluation results.

[0255] Furthermore, according to the present invention, in generating training data for AI model learning, it is possible to reflect value assessment data and actual transaction data accumulated over a long period of time from multiple experts, thereby further improving the reliability of the evaluation results.

[0256] On the other hand, Figure 15 is a diagram illustrating an example of a computer device according to an embodiment of the present invention. The valuation process of the IP valuation system according to one aspect of the present invention described above can be carried out by the computer device 600 shown in Figure 15.

[0257] Such a computer device 600, as illustrated in Figure 15, may include memory 610, a processor 620, a communication interface 630, and an input / output interface 640. Memory 610 may include a computer-readable recording medium such as RAM (random access memory), ROM (read-only memory), and a permanent mass storage device such as a disk drive. Here, the ROM and the permanent mass storage device such as a disk drive may be included in the computer device 600 as separate permanent storage devices distinct from memory 610. Memory 610 may also store an operating system and at least one program code. Such software components can be loaded into memory 610 from a computer-readable recording medium separate from memory 610. Such a separate computer-readable recording medium may include a floppy drive, disk, tape, DVD / CD-ROM drive, or memory card. In other embodiments, the software components may be loaded into memory 610 via the communication interface 630 rather than via a computer-readable recording medium. For example, software components may be loaded into the memory 610 of the computer device 600 based on a computer program installed by a file received via the network 700.

[0258] The processor 620 can be configured to process computer program instructions by performing basic arithmetic, logic, and input / output operations. Instructions may be provided to the processor 620 by memory 610 or a communication interface 630. For example, the processor 620 may be configured to execute received instructions according to program code stored in a recording device such as memory 610.

[0259] The communication interface 630 can provide a function for the computer device 600 to communicate with other devices (for example, the storage device described above) via the network 700. For example, the processor 620 of the computer device 600 can transmit requests, instructions, data, files, etc., generated by program code stored in a recording device such as memory 610 to other devices via the network 700 under the control of the communication interface 630. Conversely, signals, instructions, data, files, etc., from other devices may be received by the computer device 600 via the network 700 and the communication interface 630 of the computer device 600. Signals, instructions, data, etc., received via the communication interface 630 can be transmitted to the processor 620 and memory 610, and files, etc., can be stored in a storage medium (the permanent storage device described above) that the computer device 600 may further include.

[0260] The input / output interface 640 may also be a means for interface with the input / output device 650. For example, the input device may include a microphone, keyboard, or mouse, and the output device may include a display, speaker, or other device. In another example, the input / output interface 640 may be a means for interface with a device that integrates input and output functions into one, such as a touchscreen. The input / output device 650 may consist of the computer device 600 and one other device.

[0261] Furthermore, in other embodiments, the computer device 600 may include fewer or more components than those shown in Figure 14. However, it is not necessary to explicitly show most of the components of the prior art. For example, the computer device 600 may be implemented to include at least some of the input / output devices 650 described above, or it may further include other components such as transceivers and databases.

[0262] The foregoing embodiments can be implemented in the form of a computer program that can be executed on a computer via various components, and such a computer program may be recorded on a computer-readable medium. At this time, the medium may include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program instruction words such as ROMs, RAMs, and flash memories.

[0263] For the steps constituting the method according to the embodiments of the present invention, when there is no clear description of the order or contrary description, the steps can be performed in an appropriate order. The present invention is not necessarily limited by the described order of the steps. The use of all examples or exemplary terms (e.g., "such as") in the present invention is merely for explaining the present invention in detail, and thus the scope of the present invention is not limited thereby. It is also understood that those of ordinary skill in the art can make various modifications, combinations, and changes within the scope of the claims or their equivalents.

[0264] The embodiments of the present disclosure described above are not only realized by the device and method, but may also be realized by a program that realizes the functions corresponding to the configurations of the embodiments of the present disclosure or a recording medium on which the program is recorded.

[0265] Although the embodiments of the present disclosure have been described in detail above, the scope of the rights of the present disclosure is not limited thereto, and various modifications and improvements made by those skilled in the art using the basic concepts of the present disclosure defined in the following claims also belong to the scope of the rights of the present disclosure.

Description of Reference Numerals

[0266] 10: IP Value Evaluation System 100: Value Evaluation Service Module 110: User Interface Unit 120: Attribute Management Unit 130: Value calculation unit 140: Report generation unit 200: Data Collection and Refinement Module 210: RAW Data Collection Unit 220: Preprocessing unit 230: Basic data calculation unit 240: Training data calculation unit 250: Statistical data calculation unit 260: Evaluation criteria data calculation unit; 300: AI module 310: Training data preprocessing unit 320: AI training unit 330: Learning Optimization Unit 340: Core Variable Calculation Unit 400: Value assessment management module 410: Standard information management unit 420: User management unit 430: Payment management unit 440: Valuation System Management Unit 500: Valuation Database 600: Computer device 610: Memory 620: Processor 630: Communication interface 640: Input / Output Interface 650: Input / Output Device

Claims

1. An AI (Artificial Intelligence) based IP (Intellectual Property) valuation system, A value assessment database that includes reference information data, patent information data, and economic statistics information data as raw information, and includes statistical data and AI learning data as extracted information processed from the said raw information; A collection and refinement module that collects and processes the raw information to calculate the statistical data necessary for generating AI learning data or the first to fourth core variables, and the AI ​​learning data, and stores them in the value evaluation database; In order to calculate the first to fourth core variables, two or more AI models are trained for each core variable via the AI ​​training data, the explanatory variables matched for each core variable are identified based on the input IP information to be evaluated, the explanatory variable values ​​collected or calculated by the collection and purification module and the predictor variable values ​​are calculated via the AI ​​models, and the core variable values ​​are calculated accordingly. An AI module that calculates a first predictor variable through a first set of explanatory variables including statistical data showing the technical life cycle characteristics and competitive intensity associated with legal disputes of the IP under evaluation, calculates the value of a first core variable based on the first predictor variable, calculates a second predictor variable through a second set of explanatory variables including statistical data showing the rights stability of the IP under evaluation and the market royalty characteristics of the relevant industry, calculates the value of a second core variable based on the second predictor variable, calculates a third predictor variable through a third set of explanatory variables including statistical data showing the financial growth potential of the company and industry to which the IP under evaluation belongs, calculates the value of a third core variable based on the third predictor variable, calculates a fourth predictor variable through a fourth set of explanatory variables including statistical data showing the market size fluctuations and trade balance fluctuation trends of the relevant industry, and calculates the value of a fourth core variable based on the fourth predictor variable; and, A valuation service module that calculates the value of the target IP using the royalty exemption method based on the first to fourth core variable values, and generates a valuation report including the IP value and the statistical data; The AI ​​module confirms the patent classification information to which the target IP belongs from the input IP information to be evaluated, and confirms the matching industry classification information from the confirmed patent classification information. The median TCT (Technology Cycle Time) of the patent classification information to which the target IP belongs is checked, and the first core variable is calculated by adding the first predictor variable value to the median TCT. The standard royalty rate for the relevant industry is confirmed through industry classification information matched to the patent classification information to which the target IP belongs, and the second core variable is calculated by multiplying the standard royalty rate by the second predictor variable value. The cost of equity, equity ratio, cost of debt, and cost of debt ratio of listed companies in the relevant industry are confirmed through industry classification information matched to the patent classification information to which the target IP belongs, and the third core variable is calculated by adding the value of the third predictive variable to the cost of equity of the listed companies. This system calculates the fourth core variable by determining the sales revenue for each period over the economic lifespan of the IP, by applying the sales growth rate (the fourth predictive variable) to the initial sales revenue, which is calculated by applying the sales growth rate (the fourth predictive variable) to the initial sales revenue, if the past sales revenue of the business entity owning the target IP can be confirmed, setting the initial sales revenue based on the past sales revenue, and if the past sales revenue of the business entity owning the target IP cannot be confirmed, setting the initial sales revenue based on the sales statistics by the business entity's existing company size-based sales statistics for that industry. The first to fourth core variables are the IP economic lifespan, royalty rate, discount rate, and revenue, respectively. The first to fourth predictor variables are, The factors influencing the economic lifespan of the IP, the royalty rate, the IP commercialization risk premium, and the sales growth rate were, respectively, considered during the training of the AI ​​model. The difference between the expert assessment result for the economic lifespan of the IP and the median TCT for the patent classification information to which the IP belongs. The ratio between the expert-assessed royalty rate and the standard royalty rate for that industry, based on industry classification information matched to the patent classification information to which the target IP belongs. Expert-evaluated IP commercialization risk premium, and, Defined and learned by industry-specific sales growth rate, The aforementioned AI module is If the number of IPs to be evaluated is two or more, and it is an IP portfolio, A baseline TCT value for the target IP portfolio is set from the TCT statistics for the patent classification information of each individual patent, a first predictor variable for the target portfolio is set from the IP economic lifespan influence factors calculated for each individual patent, and the first core variable for the target portfolio is calculated by adding the value of the first predictor variable for the target portfolio to the baseline TCT of the target portfolio. The system sets industry categories based on user input, or selects a representative industry category for the target portfolio from among the industry categories matched to the patent classification information of each individual patent. The benchmark royalty rate for the representative industry of the target portfolio is set as the benchmark royalty rate for the target portfolio. A second predictor variable for the target portfolio is set from the royalty rate influencing factors calculated for each individual patent. The benchmark royalty rate for the target portfolio is multiplied by the value of the second predictor variable for the target portfolio to calculate the second core variable for the target portfolio. A third predictor variable for the target portfolio is set from the IP commercialization risk premium calculated for each individual patent, and the third core variable for the target portfolio is calculated by adding the value of the third predictor variable to the cost of equity of listed companies among the weighted average cost of capital of the representative industry of the target portfolio. A system that calculates the fourth core variable for a target portfolio by calculating the sales growth rate from the representative industry of the target portfolio, generating a fourth predictor variable for the target portfolio, setting initial sales through the past sales information of the business entity or sales statistics of the representative industry of the target portfolio, applying the sales growth rate (the value of the fourth predictor variable) to the initial sales, and calculating the sales for each period over the IP economic lifespan, which is the first core variable for the target portfolio.

2. The system according to claim 1, wherein the target IP information is the patent registration number of the target IP.

3. The aforementioned compliance information data includes expert IP evaluation result data and actual IP transaction information data. The aforementioned patent information data includes, as patent detail information, patent-specific citation / cited data, application data, appeal data, litigation data, registration data, and family data, and as grading information, patent-specific grading element data and score data for the indicators that the grading element represents according to the evaluation results for each element. The system according to claim 2, wherein the economic statistical information data includes, as economic market information, industry-specific sales growth rate data, sales statistics data, and business forecast data; as financial information, stock price data, bond interest rate data, and corporate financial data; and as export and import information, export and import data.

4. The aforementioned collection and purification module is A raw data acquisition unit that collects the aforementioned raw information; A preprocessing unit that performs preprocessing on the collected raw information; A base data calculation unit that generates base data for calculating training data and evaluation criterion data from preprocessed data; A training data calculation unit that generates AI model training data for calculating the core variables from the aforementioned basic data; A statistical data calculation unit that generates statistical data for one or more of the explanatory variables, predictor variables, and core variables as AI learning data or the core variable calculation process, or as necessary statistical data; and The system according to claim 2, further comprising an evaluation criteria data calculation unit that calculates final evaluation criteria data based on first to fourth core variable values ​​calculated via the AI ​​module and transmits it to the value evaluation service module.

5. The aforementioned evaluation criteria data calculation unit is: The system according to claim 4, which calculates the economic lifespan of an IP as the first core variable value, taking into account one or more of the legal remaining lifespan and commercialization preparation period of the target IP as evaluation criterion data, and calculates the corporate tax rate and corporate tax based on the calculated sales revenue.

6. The aforementioned statistical data The statistical information used or extracted in the calculation process of the first core variable includes TCT data for the patent classification, appeal-related statistical data, US litigation data, and market concentration data for the industry sector. The statistical information used or extracted in the calculation process of the second core variable includes industry-specific royalty rate data, the number of opinions submitted for the patent classification, the number of divisional applications, and the number of priority claims. The statistical information used or extracted in the calculation process of the third core variable includes, but is not limited to, data on the cost of equity / debt for the industry and the equity / debt ratio by industry, the patent concentration for the patent classification, and sales / operating profit growth rate data for the industry. The system according to claim 4, wherein the statistical information used or extracted in the calculation process of the fourth core variable includes initial sales statistics data corresponding to the industry and scale to which the target IP belongs, statistical data on the number of applicants and the rate of increase or decrease in the number of applications for the patent classification, and export and import rate data for the industry and product.

7. The aforementioned AI module is A training data preprocessing unit that performs preprocessing on AI training data; An AI learning unit that trains two or more AI models for each of the first to fourth core variables via the aforementioned AI learning data; A learning optimization unit that sets up two or more AI models optimized according to the performance indicator results for each core variable based on the verification results of the AI ​​model predictions; and The system according to claim 2, comprising a core variable calculation unit that calculates core variables using explanatory variables matched for each core variable and predictor variables calculated via the AI ​​model.

8. The first set of explanatory variables for generating the first predictor variable is: This includes the applicant application growth rate (application growth rate / applicant growth rate), TCT statistics, evaluation elements of the grading system, evaluation index scores of the grading system, the average number of US patent litigation cases by patent classification information, and the average number of appeal-related cases by patent classification information. The second set of explanatory variables for generating the second predictor variable is: This includes royalty rate statistics, average number of opinions submitted by patent classification, evaluation elements of the grading system, evaluation index scores of the grading system, number of divisional applications by patent classification, and number of priority claims by patent classification. The third set of explanatory variables for generating the third predictor variable is: As optimized explanatory variables for predicting IP commercialization risk premiums, we include: revenue growth rate by company size and industry, operating profit growth rate by company size and industry, evaluation elements of the grading system, evaluation index scores of the grading system, and patent concentration. The system according to any one of claims 1 to 7, wherein the fourth explanatory variable set for generating the fourth predictor variable includes the number of applicants and the rate of increase or decrease in the number of applications, and the rate of increase or decrease in exports and imports.

Images