Parallel MD agency system for utilizing artist content and diversifying revenue streams

Abstract

The present invention relates to a data-based artist management system, and more specifically, to an integrated management device for global artist discovery and management that supports objective and efficient decision-making by organically integrating and automating the entire process from the discovery of new artists to potential evaluation, overseas market analysis, partner matching, and customized content planning.

Description

Technology Field

[0001] The present invention relates to a data-based artist management system, and more specifically, to an integrated management device for global artist discovery and management that supports objective and efficient decision-making by organically integrating and automating the entire process from the discovery of new artists to potential evaluation, overseas market analysis, partner matching, and customized content planning. Background Technology

[0002] With the recent rapid growth of the global entertainment market, spearheaded by K-POP, the importance of discovering promising new artists and successfully launching them into overseas markets is increasing day by day. However, conventional methods of artist discovery and management have tended to rely heavily on the experience or subjective intuition of a small number of experts.

[0003] This approach lacked consistency in evaluation criteria and left ample room for evaluator bias, posing a constant risk of missing promising talents or making incorrect judgments. In particular, by highly evaluating only specific abilities of an artist without comprehensively considering the balance of their talents, there were frequent instances where artists achieved short-term success but faced limitations in long-term growth.

[0004] Furthermore, the process of deciding to enter overseas markets often relied on the superficial size or recognition of the market rather than objective data analysis. This overlooked the actual loyalty and growth potential of the fandom, becoming a major cause of project failures where companies faced a cold market response despite investing massive capital and time. Moreover, in the process of selecting local partners, reliance on the partners' reputation or personal networks rather than actual collaboration efficiency led to problems that hindered the likelihood of project success due to frequent communication delays and inefficient work processes.

[0005] As such, conventional technology had limitations in that each stage of management—such as artist discovery, market analysis, and partner selection—operated in a fragmented manner, and decision-making at each stage relied on subjective judgment rather than systematic data-based analysis, thereby increasing unnecessary risks and causing inefficiency in resource allocation. The problem to be solved

[0006] The present invention has been devised to solve the problems of the prior art described above. The first objective of the present invention is to provide an integrated management device capable of quantitatively evaluating not only the individual capabilities of artists but also the distribution of talent, i.e., the 'balance,' in order to identify long-term growth risks caused by talent concentration in advance and objectively discover promising talents with comprehensive potential.

[0007] The second task is to provide an integrated management mechanism capable of determining market entry priorities by comprehensively analyzing not only the superficial size of overseas markets but also the actual growth momentum of fandoms and content consumption loyalty, thereby effectively filtering out "hollow" markets and identifying "valuable markets" with high real growth potential.

[0008] The third task is to provide an integrated management mechanism capable of objectively selecting partners with high actual collaboration efficiency, not just nominal reliability, by comprehensively evaluating positive factors such as the collaboration partner's past performance and operational cost factors such as communication delays and physical distance.

[0009] The fourth task is to provide an integrated management mechanism that supports consistent, data-driven decision-making by organically linking the entire process—from artist discovery to market analysis and partner matching—while comprehensively diagnosing complex load factors such as system service quality, request volume, and resource status to promptly respond to performance degradation and ensure stable service operation.

[0010] The problems of the present invention are not limited to those mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0012] According to an embodiment of the present invention for solving the above problem

[0013] A database unit that stores data;

[0014] A control module that performs operations related to artist discovery and management based on information stored in the above database unit and controls the operation of a plurality of function modules; and

[0015] A plurality of functional modules, including a talent data management module, a potential evaluation module, a collaboration network management module, a content creation support module, a fandom analysis module, and a promoter matching module, each performing a function according to the control of the above-mentioned control module;

[0016] Includes,

[0018] The above control module is characterized by collecting a plurality of indicators representing the load status of the system, calculating a comprehensive load index of the system based on the collected indicators, and increasing the resources of the database unit when the calculated comprehensive load index exceeds a preset resource expansion threshold.

[0019] The above talent data management module is,

[0020] The method is characterized by receiving an artist's video clip from an external source, analyzing the received video clip through a pre-trained artificial intelligence model to calculate a quantified evaluation score based on pitch accuracy or choreography matching rate, and storing it in the database unit.

[0021] The above potential evaluation module is,

[0022] Step 1, which receives scores for multiple evaluation items, applies a pre-set weight to each score, adjusts them according to a log-proportional relationship, and then sums them to calculate a basic potential score;

[0023] A second step of calculating the statistical standard deviation of the scores for each of the plurality of evaluation items and calculating a talent balance index having an inverse relationship with the calculated standard deviation; and

[0024] A third step of calculating a final potential evaluation index by adjusting the basic potential score calculated in the first step above in a directly proportional relationship using the talent balance index calculated in the second step above, such that the final potential evaluation index is lowered as the deviation of the scores increases;

[0025] Characterized by performing,

[0026] The above potential evaluation module is,

[0027] The above talent balance index is calculated using an exponential damping formula of e^(-k * standard deviation), wherein k is a damping constant that is statistically corrected so that the standard deviation data of past successful artists is above a preset target threshold on average.

[0028] The above fandom analysis module is,

[0029] A first step of calculating the number of followers, the rate of increase or decrease in followers, and the number of streaming plays collected for a specific overseas region into a fandom size index, a growth momentum index, and a content consumption index, respectively, based on a log-proportional or direct-proportional relationship; and

[0030] A second step of comparing each of the three individual indices calculated in the first step with a preset minimum active threshold, classifying the final overseas market expansion priority index as 'investment deferred' if any of the three individual indices is below the minimum active threshold, and calculating the final overseas market expansion priority index through a comprehensive calculation only if all indices are above the minimum active threshold; thereby selecting only markets where the market size, growth potential, and loyalty have all been verified.

[0031] The above fandom analysis module is,

[0032] The above minimum active threshold can be characterized by setting the threshold based on the minimum requirements of a successful case by applying a percentile (P, 10) calculation formula to the initial data set of artists who have successfully entered overseas markets in the past to calculate a value corresponding to the lower 10% of the data set. Effects of the invention

[0034] The integrated management device according to the present invention has the following effects.

[0035] By calculating a 'Talent Balance Index' based on the statistical standard deviation of individual artist competency scores and adjusting the final potential index accordingly, it is possible to identify potential risks associated with artists who are overly concentrated in specific talents and objectively select promising prospects with high long-term growth potential, thereby improving the success rate of discovering new talent.

[0036] By adopting a multi-gateway filtering method that comprehensively evaluates three indicators—'fandom size,' 'growth momentum,' and 'content consumption'—in overseas markets and assigns high priority only when all indicators pass a minimum activation threshold, this approach effectively filters out inflated but substantial markets and maximizes return on investment (ROI) by concentrating limited resources on markets with verified growth potential.

[0037] By separately evaluating the two contrasting aspects of a collaboration partner—'reliability' and 'operating costs'—and calculating the final suitability based on the ratio between the two, it is possible to objectively identify the partner with the highest actual collaboration efficiency, rather than simply one with a good track record. This effectively minimizes communication and logistics costs that may arise during the project process and increases the likelihood of project success.

[0038] By diagnosing the system's load status from various angles and responding promptly to user-perceived performance degradation, service stability is ensured. At the same time, by controlling potential assessment results, market analysis results, and partner evaluation results to be organically linked as input values ​​for subsequent stages, it supports consistent and rational data-driven decision-making throughout the entire management process, thereby effectively transforming conventional subjective and fragmented work methods into systematic and automated processes.

[0039] The effects according to the present invention are not limited to those exemplified above, and a wider variety of effects are included within the present invention. Brief explanation of the drawing

[0041] Figure 1 illustrates a flowchart between all components according to the present invention. Figure 2 illustrates a flowchart for calculating a potential evaluation index according to the present invention. Figure 3 illustrates a flowchart for calculating the overseas market expansion priority index according to the present invention. Figure 4 illustrates a flowchart for calculating the collaboration partner suitability index according to the present invention. Figure 5 illustrates a flowchart for calculating a system resource expansion trigger index according to the present invention. Specific details for implementing the invention

[0042] Hereinafter, various embodiments are described in more detail with reference to the attached drawings. The embodiments described in this specification may be modified in various ways. Specific embodiments may be depicted in the drawings and described in detail in the detailed description. However, specific embodiments disclosed in the attached drawings are intended only to facilitate understanding of various embodiments. Accordingly, the technical concept is not limited by specific embodiments disclosed in the attached drawings, and it should be understood that it includes all equivalents or substitutions that fall within the spirit and scope of the invention.

[0043] Terms including ordinal numbers, such as first, second, etc., may be used to describe various components, but these components are not limited by the aforementioned terms. The aforementioned terms are used solely for the purpose of distinguishing one component from another.

[0044] Functions related to artificial intelligence according to the present disclosure are operated through a processor and memory. The processor may be composed of one or more processors. In this case, the one or more processors may be general-purpose processors such as CPUs, APs, and DSPs (Digital Signal Processors), graphics-dedicated processors such as GPUs and VPUs (Vision Processing Units), or artificial intelligence-dedicated processors such as NPUs. The one or more processors control the processing of input data according to predefined operation rules or artificial intelligence models stored in memory. Alternatively, if the one or more processors are artificial intelligence-dedicated processors, the artificial intelligence-dedicated processors may be designed with a hardware structure specialized for processing a specific artificial intelligence model.

[0045] The predefined rules of operation or artificial intelligence models are characterized by being created through learning. Here, being created through learning means that a basic artificial intelligence model is trained using a number of training data by a learning algorithm, thereby creating predefined rules of operation or artificial intelligence models configured to perform desired characteristics (or objectives). Such learning may be performed on the device itself where the artificial intelligence according to the present disclosure is executed, or it may be performed through a separate server and / or system. Examples of learning algorithms include supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but are not limited to the examples described above.

[0046] An artificial intelligence model can be composed of multiple neural network layers. Each of the multiple neural network layers has multiple nodes and weight values, and performs neural network operations through calculations between the results of previous layers and the multiple weights. The multiple weights possessed by the multiple neural network layers can be optimized based on the learning results of the artificial intelligence model. For example, multiple weights can be updated so that the loss value or cost value obtained by the artificial intelligence model during the learning process is reduced or minimized. Additionally, to minimize the loss value or cost value, multiple weights can be updated in a direction that minimizes the gradient associated with the loss value or cost value. Artificial neural networks may include deep neural networks (DNNs), such as Convolutional Neural Networks (CNNs), Deep Neural Networks (DNNs), Recurrent Neural Networks (RNNs), Restricted Boltzmann Machines (RBMs), Deep Belief Networks (DBNs), Bidirectional Recurrent Deep Neural Networks (BRDNNs), or Deep Q-Networks, but are not limited to the examples mentioned above.

[0047] A network is a network that serves as a transmission path for web pages; it may be a closed network such as a LAN (Local Area Network) or WAN (Wide Area Network), but it is desirable for it to be an open network such as the Internet. The Internet refers to a global open computer network structure that provides the TCP / IP protocol and various services existing at its upper layers, namely HTTP (HyperText Transfer Protocol), Telnet, FTP (File Transfer Protocol), DNS (Domain Name System), SMTP (Simple Mail Transfer Protocol), SNMP (Simple Network Management Protocol), NFS (Network File Service), and NIS (Network Information Service).

[0048] Terminals can be implemented in various forms. For example, the terminals described in this specification may include mobile terminals such as smartphones, tablet PCs, PDAs, portable multimedia players, and MP3 players, as well as fixed terminals such as smart TVs and desktop computers.

[0049] In this specification, terms such as “comprising” or “having” are intended to specify the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. When a component is described as being “connected” or “connected” to another component, it should be understood that it may be directly connected to or connected to that other component, or that there may be other components in between. On the other hand, when a component is described as being “directly connected” or “directly connected” to another component, it should be understood that there are no other components in between.

[0050] Meanwhile, a "module" or "part" for a component as used in this specification performs at least one function or operation. Furthermore, a "module" or "part" may perform a function or operation by hardware, software, or a combination of hardware and software. Additionally, a plurality of "modules" or a plurality of "parts," excluding a "module" or "part" that must be performed on specific hardware or on at least one processor, may be integrated into at least one module. A singular expression includes a plural expression unless the context clearly indicates otherwise.

[0051] In addition, power, power transmission, and control therefor for the following assembly configurations and embodiments, including "by control," follow conventional technology including terminals, applications, hardware control modules, etc., so they are omitted to avoid redundancy.

[0052] In addition, the operation embodiments and configurations described in a general manner without being explained in detail below follow the prior art and are omitted in order to focus on describing the purpose of the present invention and the resulting effects.

[0053] Furthermore, in describing the present invention, if it is determined that a detailed description of related known functions or configurations may unnecessarily obscure the essence of the invention, such detailed description is abbreviated or omitted.

[0054] The term "normalization" as used in this specification refers to a process of converting the relative position each value occupies within a given data set into a score between 0 and 100 in order to compare raw data having different units and ranges based on the same standard. For example, it is a method of calculating the proportion of the number of followers in a specific city within the entire range based on the minimum and maximum number of followers observed during a certain period, and assigning a scale index in proportion to that proportion.

[0055] Furthermore, the term 'log-proportional' refers to a transformation method in which the score increases as the raw value increases, but the rate of increase becomes progressively flatter. For example, when the number of contracts increases from 1 to 10 and from 10 to 100, the simple proportional method results in the score increasing by the same margin in both cases; however, the log-proportional method relatively reduces the magnitude of the increase in the latter case to prevent extremely large values ​​from dominating the overall evaluation. In this way, by combining normalization and log-proportional transformation, it is possible to fairly compare data with different units and scales while preventing evaluation distortion caused by exceptionally large values.

[0056] An integrated management device for global artist discovery and management according to the present invention is characterized by comprising: a database unit (110) for collecting and storing talent data in various fields; a control module (120) for performing operations related to artist discovery and management based on information stored in the database unit (110) and controlling the operation of each component to be described later; and a plurality of function modules that perform each function according to the control of the control module (120).

[0057] The database unit (110) goes beyond simply storing data to store talent data such as talent profiles, careers, and video materials; evaluation data including scores by evaluation item and final potential index; collaboration data such as partner lists, number of contracts, and response time; fandom data including the number of followers by region, growth rate, and number of streams; and system data including historical system load data and success and failure case data. It also adopts a master-slave replication structure that separates write and read operations to execute commands that efficiently respond to large-scale query requests. Furthermore, the database unit (110) receives a control signal that maintains data consistency and integrity by applying existing data of a similar scale as an initial value when new data is entered, and gradually replacing it with actual data as data accumulates.

[0058] The control module (120) controls the operation of each component, and comprehensively diagnoses the system's response delay load, request volume load, and resource availability load to calculate an internal parameter called the 'comprehensive load index,' and generates active control commands such as determining 'resource expansion' when the system is overloaded by comparing it with a predefined threshold value. In addition, the control module (120) continuously monitors the analysis results of the fandom analysis module (170) and transmits a trigger signal to automatically initiate the operation of the promoter matching module (180) when a pre-set condition is met, such as maintaining the 'highest priority target' grade for two consecutive weeks, where the priority index of a specific city is maintained.

[0059] The above talent data management module (130) performs the function of collecting various types of data, such as profiles, careers, and video materials of potential personnel not limited to specific fields such as broadcast MCs, entertainers, singers, and sports players, classifying them according to predetermined categories, and storing them in the above database unit (110), and executes a data preprocessing command to analyze audition video clips, which are unstructured data, to calculate pitch accuracy, beat matching rate, choreography similarity, etc., for the calculation of a subsequent analysis module, and converts them into a score form, which is structured data.

[0060] The potential evaluation module (140) accesses data of potential personnel stored in the database unit (110) and quantitatively evaluates the potential of potential personnel based on multiple pre-set evaluation items such as star quality, talent, and market suitability specialized for the K-POP and K-Culture markets, and calculates the evaluation results. It performs a calculation that ensures objectivity and reproducibility of the evaluation process by performing a three-stage logical processing procedure using weights and damping constants derived through statistical analysis based on past data and a clear calculation formula. Additionally, the potential evaluation module (140) executes commands that include exception handling logic to ensure reliability, such as identifying outliers in the evaluation data to reduce their influence or suspending the evaluation in the event of data omission.

[0061] The above-mentioned collaboration network management module (150) systematically manages a list of multiple entertainment companies, broadcasters, and investors capable of collaboration, as well as related cooperation relationship data, within the above-mentioned database unit (110), analyzes and identifies cooperation targets with high potential for joint planning and promotion in relation to a specific artist or project, and performs a specific three-stage evaluation algorithm that evaluates the efficiency of partners based on two clear criteria: 'collaboration reliability' and 'operational costs', wherein the weights used in the above-mentioned algorithm are calculated by objectively deriving them through multiple regression analysis of past project success rates.

[0062] The content production support module (160) analyzes the genre characteristics of a specific artist, fandom preferences, and market trend data, and generates and provides an album planning direction or concert planning solution suitable for the artist based on the analysis results. At this time, it uses fandom preference data and external market trend data received from the fandom analysis module (170) as input values, and generates a control signal that recommends the optimal genre and concept based on the artist's strengths and weaknesses data derived from the analysis results of the potential evaluation module (140).

[0063] The above-mentioned fandom analysis module (170) collects and analyzes domestic and international fandom-related data in real time, such as social media data, album sales, and streaming records. Through the analysis, it identifies specific overseas regions and times where the size and activity of the fandom exceed a predefined threshold and determines them as candidate locations for fan meetings or concerts. It performs an analysis algorithm using a multi-gate filtering method that applies a 'minimum active threshold' based on three criteria: 'fandom size', 'growth momentum', and 'content consumption'. The threshold is calculated by statistically deriving it through percentile analysis of past successful cases.

[0064] The above-mentioned promoter matching module (180) stores and manages information on overseas clients and promoters in the database unit (110). When planning an overseas performance or promotion of a specific artist, it performs the function of connecting a promoter that matches conditions such as the artist's genre, scale, and activity region among the promoter information stored in the database unit (110). It limits the search range based on the 'highest priority target' city information received from the fandom analysis module (170), and transmits a command to perform a multi-stage decision-making process that applies the 'collaboration partner suitability index' calculation logic of the collaboration network management module (150) to the group of promoter candidates within the range and recommends the promoter with the highest index as the highest priority.

[0065] The core information processing processes performed by the integrated management device according to the present invention are described in more detail below. The device of the present invention organically performs the following four core information processing processes to evaluate an artist's potential from various angles, identify the optimal market, select an efficient partner, and maintain system stability.

[0066] First, the potential evaluation process overcomes the limitations of fragmentary evaluations by analyzing not only the artist's individual competency scores but also the statistical deviations of those scores to calculate a comprehensive potential index that takes into account the balance of talent.

[0067] Next, in the market analysis process, three criteria—the current size of the overseas fandom, future growth potential, and actual content consumption loyalty—are evaluated using a multi-gateway method, and investment risk is minimized by selecting only markets that pass all criteria as valid candidates for entry.

[0068] Next, in the partner selection process, positive aspects such as the partner's past performance and operational cost aspects related to communication and movement are separated and quantified, and the partner with the highest actual collaboration efficiency is identified through the ratio relationship between the two.

[0069] Next, in the system load management process, the system overload status is precisely determined by comprehensively diagnosing user-perceived performance, current request volume, and available resource status, and based on this, resources are dynamically increased only when necessary to simultaneously ensure service stability and cost efficiency.

[0071] The judgment criteria used to evaluate the potential of an artist in the present invention include 'quantitative scores for each evaluation item' and 'talent balance index', and the potential evaluation module (140) transmits a control command to perform a logical processing procedure consisting of three specific steps to calculate a final potential evaluation index using the judgment criteria.

[0072] Among the above-mentioned judgment criteria, the 'quantitative score by evaluation item' is a standard for objectively measuring the individual capabilities possessed by an artist, and is specifically defined by multiple items such as star quality, vocals, dance, speech, and market suitability, and each item is characterized by being quantified as a value between 0 and 100 points. In addition, among the above-mentioned judgment criteria, the 'talent balance index' is a unique standard of the present invention for measuring how evenly an artist's capabilities are distributed, and is based on the technical concept that not only is the total sum of individual capabilities high, but that a talent possessing skills above a certain level in various fields is more likely to succeed in the long term. It generates a control signal aimed at solving the problem of the risk of talent bias that was overlooked by conventional technology and predicting the sustainable growth potential of an artist.

[0073] The potential evaluation module (140) first performs a first operation by multiplying each of the input 'quantitative scores by evaluation item' by a weight according to the importance within the industry that is pre-set, and then adjusts each score value based on a log-proportional relationship to reflect the non-linear characteristic in which the influence on potential gradually becomes milder as the score rises above a certain level, and finally calculates the 'basic potential score' by summing the scores of all adjusted items. Next, the potential-talent evaluation module calculates a statistical standard deviation for the set of input 'quantitative scores by evaluation item', and the standard deviation is an objective indicator representing how much the scores are scattered from the mean. The potential evaluation module (140) is designed so that the standard deviation value and the 'talent balance index' to be finally calculated have an inverse relationship, and performs an operation to convert the 'talent balance index' so that the smaller the score deviation, the higher the value of the 'talent balance index' becomes, and the larger the deviation, the lower the value of the 'talent balance index' becomes. Finally, the potential evaluation module (140) applies the 'talent balance index' calculated in the second step to the 'basic potential score' calculated in the first step to adjust the final score. Specifically, the 'basic potential score' and the 'talent balance index' are made to have a directly proportional relationship, so that the final score is lowered for artists with unbalanced talent and a low balance index, and the score is maintained or slightly increased for artists with even talent, and a control signal is automatically transmitted.

[0074] The above 'quantitative scores by evaluation item' are collected from multiple sources to enhance reliability, and scores are automatically calculated by having three or more professional judges directly input scores for each item through the evaluation system, or by a pre-trained artificial intelligence model analyzing audition video clips to calculate pitch accuracy, beat matching rate, choreography similarity, etc. For example, calculations are performed assuming that scores for three items, such as vocal score of 85 points, dance score of 90 points, and star quality score of 70 points, have been obtained for 'Artist X'.

[0075] In one embodiment, when the potential evaluation module (140) has both a score assigned by an expert group and a score calculated by a pre-trained artificial intelligence model for the same evaluation item, it adjusts the two scores with different confidence weights and calculates the average to convert them into a single integrated score. For example, the evaluation by expert judges with extensive field experience in a specific genre may be considered to have relatively high confidence in the initial stage, and the artificial intelligence model may be designed to play an auxiliary role until sufficient training data is accumulated. In this case, the system calculates the integrated score by giving greater weight to the expert score and relatively less weight to the artificial intelligence score. Subsequently, after a certain period has elapsed and the artificial intelligence model has sufficiently learned various actual evaluation cases and feedback data, the system automatically adjusts the weights to gradually increase the weight of the artificial intelligence score and relatively decrease the weight of the expert score.

[0076] According to this structure, the potential evaluation module (140) can balance the advantages of both sources by performing evaluations based on the intuition and industry experience of human experts in the initial stages of evaluation, while increasing the weight of data-based artificial intelligence analysis over time. As a result, it is possible to implement a self-improving evaluation system that simultaneously secures objectivity and explainability in evaluation, while improving the model's predictive power as new data is accumulated. The calculated 'final potential evaluation index' compares the indices of all potential personnel within the system and automatically classifies the top 10% as 'Grade A', the top 30% as 'Grade B', etc., and outputs a screen that supports the manager in intuitively identifying promising talents. Additionally, for artists whose 'talent balance index' falls below a certain threshold during the index calculation process, the evaluation item with the lowest score is automatically designated as an 'area requiring intensive training', and information recommending a training program is provided. Furthermore, the system automatically sends a notification to the manager to prioritize the allocation of investment and promotional budgets to Grade A artists.

[0077] In the potential evaluation scenario of 'Artist X', the potential evaluation module (140), which receives 85 points for vocals, 90 points for dance, and 70 points for star power as input values ​​and 0.4 points for vocals, 0.4 points for dance, and 0.2 points for star power as weights, applies weights to each score, adjusts them in a log-proportional relationship to reflect non-linear influence, and then sums the adjusted values ​​to produce a basic potential score of '83.1'. Subsequently, the potential evaluation module (140) calculates the standard deviation of the set of scores (85, 90, 70) as '8.03' and calculates a talent balance index of '0.85' mapped to the standard deviation of 8.03 according to a predefined judgment rule table. Finally, the potential evaluation module (140) derives a final potential evaluation index of '70.6' through a calculation that applies a talent balance index of '0.85' as a correction factor to a basic potential score of '83.1', and sends a command to display on the administrator screen that the final potential evaluation index of 'Artist X' is '70.6' and corresponds to 'Grade B'.

[0078] The potential evaluation module (140) performs a logic to ensure objectivity in the evaluation by determining that a score given by a specific judge is an outlier and applying a weight reduced by 50% if the score differs from the average of other judges by more than 2 standard deviations when multiple expert evaluation scores are used. Additionally, if score data for a specific evaluation item is missing, it classifies the artist as 'evaluation pending' and generates a notification requesting the administrator to supplement the data, and executes an exception handling command to prevent evaluation if two or more items are missing.

[0079] In a scenario comparing the operation results of the conventional technology and the present invention, when evaluating a biased 'Artist B' with 95 points for vocals, 50 points for dance, and 60 points for star power, and a balanced 'Artist C' with 75 points for vocals, 80 points for dance, and 70 points for star power, the conventional technology using a simple weighted summation method yields '70.0' points and '76.0' points, respectively. On the other hand, according to the three-step procedure of the present invention, 'Artist B' has a very high standard deviation of 19.1, so the talent balance index is calculated as low at 0.6, and the final index is significantly lowered to '41.7', whereas 'Artist C' has a very low standard deviation of 4.0, so the talent balance index is calculated as high at 0.98, and the final index is maintained at '74.3'. These results generate data demonstrating the technical effect of the present invention, which enables more precise and reliable decision-making by clearly detecting the risk of talent bias in Artist B and lowering the final index, thereby evaluating Artist C, who has high long-term growth potential, more highly.

[0080] The present invention verifies its effects through 10 simulation examples as shown in the following table.

[0081]

[0082] The present invention stores a judgment rule table as shown in the following table in a database unit (110) and executes a command to refer to the table when calculating the talent balance index.

[0083]

[0084] The judgment rule of the present invention is based on a universal and reproducible statistical measure called 'standard deviation,' thereby ensuring objectivity by excluding the subjectivity or arbitrariness of the evaluator. Furthermore, as the modern entertainment market demands diverse capabilities from artists, adopting the balance of talent as a core judgment criterion from the potential evaluation stage is a technical necessity to meet market demands and minimize investment risk.

[0085] By adopting a judgment criterion called the 'Talent Balance Index' and a corresponding step-by-step processing procedure, the present invention improves the prediction accuracy of selecting artists with sustainable growth potential rather than short-term popularity compared to conventional technology, enhances the reliability of decision-making by transforming the talent discovery process—which previously relied on subjective judgment—into a data-based quantitative process, and provides the effect of specifically reducing the probability of investment failure in new artists by identifying potential risks that may arise from talent bias in advance and reflecting them in the evaluation.

[0086] In this specification, the 'talent balance index' refers to a normalized index between 0 and 1 that indicates how evenly an artist's capabilities are distributed, set to be inversely proportional to the statistical deviation of multiple evaluation item scores. Additionally, the interval threshold of the standard deviation used in the 'judgment rule table' is statistically set by analyzing the distribution of initial evaluation data of 100 artists who successfully debuted in the past during the initial stages of system operation, and a control command is transmitted to automatically correct the data by additionally learning the success and failure data of new artists' debuts quarterly to reflect changing market trends.

[0087] The damping constant k used in the above-mentioned Talent Balance Index calculation formula is objectively determined through simulations based on historical data. Specifically, using the pre-debut evaluation data (standard deviations) of 10 artists who successfully debuted in the past, the optimal value of k is derived to ensure that each artist's 'Talent Balance Index' is at least the target threshold of 0.9. This is performed by finding the k that minimizes the following objective function:

[0088] Objective function J(k) = Σ0.9 - e^(-k * standard deviation_i)))^2

[0089] In the above formula, i represents each successful artist, and the max(0, ...) function calculates the error (penalty) only when the exponent is less than 0.9. The table below shows the results of 10 simulations performed while changing the value of k.

[0090]

[0091] According to the simulation results above, when the damping constant k is set to 0.03 (SIM-K-04), the balance index in all successful artist cases satisfies 0.9 or higher, and the objective function value approaches 0. Therefore, in the present invention, by setting the initial value of the damping constant k to 0.03, objectivity in setting the threshold value and data-based rationale are secured.

[0092] In one embodiment, the 'Talent Balance Index' is designed to become a value close to 1 as the standard deviation between multiple evaluation item scores approaches 0, and to gradually decrease to around 0.5 as the standard deviation increases. Specifically, in the design stage, representative standard deviation values ​​of artists who have successfully debuted in the past are collected, and a damping constant is repeatedly searched for such that the Talent Balance Index corresponding to each standard deviation value becomes greater than a certain threshold value (e.g., 0.9 or higher). At this time, if the Talent Balance Index for each artist is calculated to be smaller than the threshold value, the difference is considered as an error, squared, and the errors for all artists are summed, and the damping constant that minimizes this sum is adopted as the final design value.

[0093] In actual system operation, instead of using the continuous decline curve derived as described above, the range of standard deviations is divided into multiple predetermined intervals (e.g., 0 or greater and less than 5, 5 or greater and less than 10, etc.), and a judgment rule table is generated by selecting a representative value for each interval and associating it with a talent balance index fixed to that interval (e.g., 1.00, 0.85, 0.75, 0.60, 0.50). The judgment rule table is used for the purpose of approximately implementing the continuous decline curve defined by the damping constant without real-time computation; therefore, the expression of the continuous function based on the damping constant and the expression of the judgment rule for each standard deviation interval correspond to the same design result being represented in different forms.

[0094] Furthermore, as new debut cases accumulate, the system periodically repeats the aforementioned procedure to re-evaluate the standard deviation distributions of the success and failure groups, and automatically adjusts the damping constant and the interval boundaries or exponential values ​​of the judgment rule table based on the results. Through this, the Talent Balance Index functions as a dynamic evaluation metric that reflects changing market environments and evaluation criteria in near real-time.

[0095] To verify the effect of the set damping constant k=0.03, data from 10 artists who failed to debut in the past are applied to the same formula and the results are compared. If it is a successful model, the average equilibrium index of the failure group should be significantly lower than that of the success group.

[0096]

[0097] As a result of the verification simulation above, the average talent balance index of the success group was 0.928, while the average of the failure group was significantly lower at 0.685. This demonstrates that the threshold value of k=0.03 functions as an objective criterion that effectively distinguishes between groups with a high probability of success and those with a low probability. Therefore, it is clear that this threshold setting is a reasonable and verified result that reflects the distribution characteristics of the actual data.

[0099] The judgment criteria used to determine the priority of market expansion in a specific overseas region in the present invention include three independent indicators: a 'fandom size index', a 'growth momentum index', and a 'content consumption index'. The fandom analysis module (170) generates a control signal that performs a logical processing procedure consisting of two specific steps to organically combine the three judgment criteria to calculate a final priority index.

[0100] Among the aforementioned criteria, the 'Fandom Size Index' is a standard for measuring the current size of the potential customer base formed within a specific region, calculated based on the number of followers on social media platforms, and aims to objectively identify the current reach of the market. Additionally, among the aforementioned criteria, the 'Growth Momentum Index' is a standard for measuring how rapidly the fandom in a specific region is expanding, calculated based on the rate of increase or decrease in followers over a certain period, and aims to predict the future growth potential of the market and distinguish between a stagnant market and a dynamically growing market. Furthermore, among the aforementioned criteria, the 'Content Consumption Index' is a standard for measuring how actively the fandom in a specific region consumes the artist's content, calculated based on the number of streaming plays of music tracks and music videos, and executes a command aimed at gauging actual fandom loyalty and the potential for purchase conversion in the market, going beyond mere awareness.

[0101] The above-mentioned fandom analysis module (170) first calculates a normalized individual index based on raw data corresponding to each judgment criterion, and the 'fandom size index' is calculated based on the number of collected followers, but the value is adjusted based on a log-proportional relationship to prevent the influence of a specific market with a very large number of followers from being excessively reflected. In addition, the 'growth momentum index' is calculated based on the rate of increase / decrease of collected followers, and the rate of increase / decrease value and the index value are set to have a direct proportional relationship, and the 'content consumption index' is calculated based on the number of collected streaming plays, and the calculation is performed to adjust the value based on a log-proportional relationship for the same technical reasons as the fandom size index. Next, the fandom analysis module (170) calculates the final overseas market expansion priority index by comprehensively calculating the three individual market indices calculated in the first step. At this time, the fandom analysis module (170) is characterized by performing calculations based on a logical AND condition that all three indicators must satisfy a minimum standard, rather than using a simple summation method, and automatically transmits a command to significantly lower the final priority index or classify it as 'investment hold' if any of the three individual market indices falls below a predefined 'minimum active threshold'.

[0102] The 'Overseas Market Priority Index' derived from the above market analysis process is an indicator that expresses a single comprehensive score by combining three distinct indices—fandom size, growth rate, and consumption loyalty—into elements of different importance, rather than simply listing them. First, since the units and scales of the raw data for each index differ, the system analyzes the minimum and maximum values ​​of each index and performs normalization processing to convert all values ​​into a range between 0 and 100. In the next step, if any of the size index, growth index, or consumption index falls below a pre-set minimum active threshold (e.g., 50 points), the city is determined to lack a fandom base or have low growth potential, is immediately classified as 'Investment Pending,' and the comprehensive priority index is not calculated.

[0103] Only for cities where all three indices exceed a threshold, the system calculates a comprehensive priority index by multiplying each index by a different importance coefficient and summing them. The importance coefficients are determined by analyzing data on revenue, audience numbers, and contract renewal rates of overseas projects conducted over the past three years, ensuring that higher weight is assigned to indices that have a greater impact on actual performance. For example, in certain genres or periods, consumer loyalty may show a closer correlation with actual revenue than the size of the fandom; in such cases, the system calculates the comprehensive index by assigning a higher importance coefficient to the consumption index. Furthermore, as the performance of new projects accumulates, the system periodically re-estimates the aforementioned importance coefficients to automatically adjust the influence of each index in line with actual market response.

[0104] Therefore, the overseas market priority index features a weighted summation structure in which the influence of each index varies according to actual contributions verified in the market, rather than a simple average; this supports multidimensional decision-making that considers growth potential and consumer loyalty, rather than relying solely on the size of the short-term fandom.

[0105] Data regarding 'fandom size', 'growth momentum', and 'content consumption' are periodically collected via APIs from external platforms, such as by collecting the number of followers by region on a daily basis through official APIs of major social media platforms like Twitter and Instagram, calculating the weekly growth rate by comparing the number of followers at the current time with the number of followers 7 days ago, and collecting the number of streaming plays by region on a daily basis through artist dashboard APIs provided by global music streaming services like Spotify and YouTube Music. For example, calculations are performed assuming that data for the past week for 'City A' includes 1,000,000 followers, a follower growth rate of +5%, and 2,500,000 streaming plays.

[0106] The calculated 'Overseas Market Expansion Priority Index' visually generates a 'Global Priority Map' within the system, sorting major cities around the world in index order, and provides this to marketing managers. Additionally, the system generates information that automatically recommends differential marketing budget sizes for cities classified by the index into 'Top Priority Target' and 'Major Target' grades. Furthermore, if the priority index of a specific city maintains the 'Top Priority Target' grade for two consecutive weeks, a control signal is automatically transmitted to a promoter matching module (180) to search for and attempt to match active promoters in that region.

[0107] In the scenario for calculating the priority index of 'City A', the fandom analysis module (170), which receives 1 million followers, a follower growth rate of +5%, and 2.5 million streaming plays as input values, calculates the 'fandom size index' as '90' by converting the 1 million followers according to a log-proportional relationship, calculates the 'growth momentum index' as '85' by converting the follower growth rate of +5% according to a direct-proportional relationship, and calculates the 'content consumption index' as '88' by converting the 2.5 million streaming plays according to a log-proportional relationship. Subsequently, the fandom analysis module (170) determines whether all three calculated indices exceed the 'minimum active threshold' of '50', and since all conditions are satisfied, performs a calculation to combine the three indices to derive the final priority index '87.5'. Finally, execute a command to update the global priority map with the final priority index of 'City A' being '87.5', which corresponds to the 'Top Priority Target' grade, and to send a notification to the marketing manager.

[0108] The above-mentioned fandom analysis module (170) performs a logic to apply a 7-day moving average to the follower growth rate and streaming playback count to smooth the data and then use it to calculate an index, in order to prevent data distortion caused by temporary viral phenomena or bot activity. Additionally, if data reception from an external platform's API fails for more than 24 hours, the city is classified as 'unable to collect data' and temporarily excluded from analysis, and an exception handling command is executed to automatically send a warning including an error log to the system administrator.

[0109] In a scenario comparing the operational results of the prior art and the present invention, when evaluating 'City B,' a massive stagnant market with 5 million followers, a growth rate of 0.1%, and 1 million streams, and 'City C,' a high-growth market with 800,000 followers, a growth rate of 15%, and 3 million streams, the prior art based solely on the number of followers judges 'City B' as the first priority and 'City C' as the second priority. On the other hand, the present invention detects that the growth momentum index of 'City B' is '15,' which falls short of the 'minimum active threshold,' and downgrades its final priority to 'Investment Deferred,' whereas 'City C' satisfies all conditions with a size index of '88,' a growth index of '99,' and a consumption index of '92,' and is selected as the 'Top Priority Target' for final priority. This generates data proving the technical effect of the present invention in fundamentally blocking the risk of incorrect resource allocation.

[0110] The present invention stores a judgment rule table as shown in the following table in a database unit (110) and executes a command that refers to the table when calculating the comprehensive priority index.

[0111]

[0112] The judgment rules of the present invention are based on objective and verifiable data, such as the number of followers, growth rate, and number of views, collected through an external API, thereby ensuring objectivity by excluding the subjectivity or arbitrariness of the analyst. Furthermore, since there are frequent cases of failure when entering the global market solely because of a large number of fans, adopting a multi-gateway evaluation model that simultaneously verifies the market's 'present value,' 'future value,' and 'actual value' is a technical necessity for managing investment risk and maximizing the probability of success.

[0113] By adopting three judgment criteria—'fandom size,' 'growth momentum,' and 'content consumption'—and combination rules based on logical AND conditions, the present invention effectively filters out fictitious markets and significantly improves prediction accuracy compared to conventional technology in discovering markets with actual growth potential; it induces the concentration of marketing resources on a small number of markets with verified growth potential, thereby maximizing the return on investment of a limited budget; and specifically reduces the risk of project failure by preventing hasty market entry decisions through multifaceted verification procedures and supporting systematic decision-making based on data.

[0114] In this specification, the 'minimum active threshold' refers to a numerical value defined as the minimum threshold that an individual market index must exceed in order for a specific market to be considered a significant investment target. Additionally, the 'minimum active threshold' is statistically set based on the bottom 10 percentile of the data from the first six months of entry for 50 teams of artists who have successfully entered overseas markets over the past three years, and transmits a control command that is automatically corrected once a year by reflecting the average growth rate data of emerging markets published in the annual global music market report.

[0115] The aforementioned 'Minimum Activation Threshold' is derived through a statistical formula based on data from past success stories. Specifically, it utilizes a dataset P of normalized individual market indices (size, growth, consumption) from the initial six months of entry for 10 artist teams that successfully entered overseas markets over the past three years. The threshold is set to the value corresponding to the bottom 10% of this dataset, which is calculated using the following percentile formula:

[0116] Minimum activation threshold = percentile(P, 10)

[0117] This is a conservative and objective method that uses the case satisfying the minimum conditions among the success stories as a standard. The table below shows the simulation process of deriving a threshold value using a portion of the data from 10 actual past successful artist teams.

[0118]

[0119] Based on the simulation results above, when considering all indicators comprehensively, it can be seen that the bottom 10% of successful cases are distributed around 50 points. Therefore, setting the 'minimum active threshold' to 50 is a reasonable decision based on data in order to establish a standard that all individual indices must exceed at least this level.

[0120] To verify the effect of the set 'minimum active threshold = 50', the same rule is applied to data from 10 cities that failed to enter the market or had poor performance in the past, thereby verifying how effectively the logic of the present invention filters out these 'trap markets'.

[0121]

[0122] As a result of the verification simulation above, all cities in the success group had three indices exceeding the threshold of 50, whereas cities in the failure / underperforming group had at least one index below the threshold and were accurately filtered as 'investment deferred'. This clearly demonstrates that the threshold of 50 acts as a significant criterion distinguishing success from failure in market entry, and proves that the logic of the present invention effectively manages risk based on actual data patterns.

[0124] The judgment criteria used to determine the suitability of collaboration with a specific potential partner in the present invention include two complementary indicators, namely a ‘collaboration reliability index’ and an ‘operation cost index,’ and the collaboration network management module (150) executes a control command that performs a logical processing procedure consisting of three specific steps to calculate the final suitability index by combining the two judgment criteria.

[0125] Among the aforementioned evaluation criteria, the 'Collaboration Reliability Index' is a standard for measuring the reliability of a partner's experience and past track record. It is calculated based on the number of contracts concluded in the past and aims to objectively quantify the partner's industry network, project execution capabilities, and reliability. Additionally, among the aforementioned criteria, the 'Operational Cost Index' is a standard for measuring tangible and intangible costs and obstacles that may arise when collaborating with a partner. It is calculated based on the average time required from proposal to response and the average physical distance between the artist's primary activity area and the partner company, and generates a control signal intended to evaluate operational efficiency and the potential for smooth communication during the actual collaboration process, rather than merely the partner's reputation.

[0126] The collaboration network management module (150) first extracts the cumulative number of contracts concluded from a specific partner's past database and calculates a 'collaboration reliability index' by applying a log-proportional relationship to the extracted number of contracts concluded to reflect the non-linear characteristic that a higher number of contracts indicates greater experience of the partner, but the rate of increase in reliability gradually slows down. Next, the collaboration network management module (150) extracts 'average response time' and 'physical distance' data from communication logs with the partner and a geographic information system, applies a pre-set weight to each of these two elements, and sums them to calculate an 'operation cost index'. At this time, the module performs an operation to design the index so that its value increases as the response time lengthens or the distance increases. Finally, the collaboration network management module (150) determines a final suitability index by combining the 'collaboration reliability index' calculated in the first step and the 'operation cost index' calculated in the second step, and automatically transmits a command to calculate the final index such that it is in a directly proportional relationship with the 'collaboration reliability index' and inversely proportional relationship with the 'operation cost index'.

[0127] Data regarding 'collaboration reliability' and 'operating costs' is collected from internal and external systems, such as by querying the internal contract management system database to aggregate the history of past contract signings with specific partners, analyzing logs of official communication channels with partners (email, etc.) to calculate the average time difference between the time a proposal is sent and the time the first response is received, and calculating the shortest path distance between the two locations via a map API by geocoding the artist's main activity base and the partner company's headquarters address. For example, calculations are performed assuming that data has been obtained for 'Partner Company A' with 20 past contract signings, an average response time of 2.5 days, and a physical distance of 350 km.

[0128] The calculated 'Collaboration Partner Suitability Index' automatically sorts a list of potential partner candidates for a specific project within the system in order of suitability index, providing the person in charge with a recommended ranking. Additionally, for partners whose index falls below a pre-set 'Collaboration Caution Threshold,' a warning notification is displayed to the person in charge, specifying concrete risk reasons such as 'communication delays' or 'geographical inefficiency.' Furthermore, when negotiating with a partner who has a high suitability index but high specific cost factors, the system automatically proposes additional contract terms to offset those costs, thereby generating information to support the negotiation.

[0129] In the scenario for calculating the suitability index of 'Partner Company A', the collaboration network management module (150), which receives 20 contracts, a response time of 2.5 days, and a physical distance of 350 km as input values, converts the 20 contracts according to a log-proportional relationship to calculate the 'collaboration reliability index' as '85'. Subsequently, the response time of 2.5 days and the physical distance of 350 km are each converted into cost scores, and by applying weights and summing them, the 'operation cost index' is calculated as '65'. Finally, the collaboration network management module (150) performs a final calculation through the ratio relationship between the reliability index '85' and the cost index '65' to derive the 'final collaboration suitability index' as '78', and transmits a command to update and display the partner ranking list on the manager screen that the final suitability index of 'Partner Company A' is '78' and corresponds to the 'Recommended Partner' grade.

[0130] The above-mentioned collaboration network management module (150) performs logic to mitigate cold start problems by applying industry average data or data from a partner of similar size as an initial value in cases where historical data is insufficient, such as with a new partner, and controls the initial value to be gradually replaced with actual data as collaboration data with the partner accumulates. Additionally, when a change in key information is detected, such as a change in the partner's address or communication manager, it executes an exception handling command that automatically sets the relevant cost index to a 'recalculation required' state and creates a task to request the manager to verify and update the information.

[0131] In a scenario comparing the operational results of the prior art and the present invention, when evaluating 'Partner B,' a high-reliability / high-cost partner with 50 contracts, a response time of 7 days, and a distance of 2,000 km, and 'Partner C,' a medium-reliability / low-cost partner with 15 contracts, a response time of 1 day, and a distance of 50 km, the prior art, based on the single criterion of the number of contracts, judges 'Partner B' as the first priority and 'Partner C' as the second priority. On the other hand, the present invention downgrades the final suitability of 'Partner B' to a 'Collaboration Caution' grade because its reliability index is very high at '98' but its operating cost index is very high at '95,' while selecting 'Partner C' as the 'Top Recommendation' target because its reliability index is '75' and its operating cost index is very low at '20.' This generates data that demonstrates the technical effect of the present invention in identifying actual collaboration risks hidden behind outwardly visible performance results.

[0132] The present invention stores a judgment rule table as shown in the following table in a database unit (110) and executes a command to classify grades by referring to the table when calculating the final collaboration suitability index.

[0133]

[0134] The judgment rules of the present invention are based on objective data automatically collected and calculated by the system, such as internal system logs and external APIs, thereby ensuring objectivity by fundamentally blocking any possibility of subjective preferences from the person in charge interfering. Furthermore, since a successful project is largely determined not only by the partner's reputation but also by practical factors such as smooth communication and close cooperation, adopting the model of the present invention, which evaluates the positive aspects and cost aspects of collaboration in a balanced manner, has a technical necessity for increasing the probability of project success.

[0135] The present invention supports more rational partner selection decisions by providing a balance between nominal performance and actual operational efficiency by adopting evaluation rules based on two criteria, 'collaboration reliability' and 'operational costs,' and by prioritizing partners who can minimize unnecessary time and costs associated with communication and travel, thereby improving the overall operational efficiency of a project. Furthermore, it provides the effect of preventing and managing potential problems that may arise during the collaboration process by quantitatively evaluating and identifying 'communication risk' and 'geographical risk' in advance, which are easily overlooked by conventional technologies.

[0136] In this specification, the 'collaboration caution threshold' refers to a numerical value defined as a standard by which the system determines that additional review or caution is required before proceeding with collaboration if the final collaboration suitability index is below this value. In addition, the classification criteria including the 'collaboration caution threshold' are statistically set based on the distribution of suitability indices of partners associated with failed projects by analyzing 100 successful and failed projects performed over the past three years, and the system automatically learns the results of partner satisfaction surveys and project performance data each year to transmit control commands that are automatically corrected to continuously improve the accuracy of the partner evaluation criteria.

[0137] The 'collaboration partner suitability index' calculated by the above-mentioned collaboration network management module (150) is expressed as a linear combination structure in which the collaboration reliability index and the operating cost index exert influence in opposite directions. More specifically, the system first evaluates the extent to which each partner contributed to the success of past projects based on the collaboration reliability index, and applies a correction item designed so that the penalty increases as the operating cost index for the same partner increases. This correction item is defined as the value obtained by multiplying the operating cost index by a constant damping coefficient; since the operating cost index naturally increases for partners with long response times or long physical distances, the amount deducted from the final suitability index also increases through multiplication with the damping coefficient.

[0138] As a result, the final fit index has a structure in which it increases as the collaboration reliability index increases and decreases as the operating cost index increases; the damping coefficient used in this process is automatically derived through regression analysis of past project data. In the regression analysis process, the success rate or satisfaction evaluation score of past projects is set as the dependent variable, and cost-related indicators such as response time and distance are set as independent variables to quantitatively calculate the impact of changes in each variable on the success rate. If the analysis results indicate that response time has a greater impact on the success rate than distance, the system sets the damping coefficient so that the weight corresponding to response time is reflected more significantly.

[0139] Under this linear combination structure, even if two partners have the same level of collaboration reliability, the partner with faster response times and closer proximity, resulting in a lower operational cost index, is consistently assigned a higher final suitability index. Conversely, even if a partner has a high reliability index due to a large number of contracts, if their average response time is long and they are far away, they may be significantly penalized by the operational cost index and consequently classified as 'Collaboration Caution' or 'Collaboration Pending.' Unlike conventional technology that evaluates partners based solely on the number of past contracts, this enables more realistic collaboration decision-making that considers actual operational efficiency and costs.

[0140] The weights W_time and W_dist used to calculate the aforementioned operating cost index are derived through multiple regression analysis of historical project data. This analysis quantitatively calculates how much average response time and physical distance (independent variables), respectively, affect the project success rate (dependent variable).

[0141] Project Success Rate = β0 + β_time * Normalized(Response Time) + β * Normalized(Distance)

[0142] The final weights are set using the ratio of the absolute values ​​of the standardized regression coefficients β and β derived from the above regression equation. For example, assume that the simulation results for the following 10 project data are as follows.

[0143]

[0144] Based on the simulation results above, the absolute value of the standardization factor for response time (0.65) is approximately 2.6 times greater than the absolute value of the factor for distance (0.25). This implies that response time has a greater impact on project success. Therefore, if the sum of the two weights is normalized to 1, W_time 0.72, W_dist You can set objective weights based on data, such as 0.28.

[0145] To verify the effect of the set weights and formulas, the final fit index is calculated using data from 10 partners whose past collaboration satisfaction was rated as 'very high' and 10 partners whose satisfaction was rated as 'very low', and the index distribution between the two groups is compared.

[0146]

[0147] As a result of the verification simulation above, the average final fit index of the group with high collaboration satisfaction was very high at 85.5, whereas the average of the group with low satisfaction was significantly low at 41.2. This demonstrates that the evaluation model of the present invention predicts and distinguishes the quality of actual collaboration and the probability of success with high accuracy. Therefore, it is clear that rules such as setting the 'collaboration attention threshold' to 50 are reasonable decisions based on the actual data distribution.

[0149] In the present invention, the judgment criteria used to measure the load of the database unit (110) and to determine whether to increase resources include three independent technical indicators: 'response delay load', 'request amount load', and 'resource availability load'. The control module (120) generates a control signal that performs a logical processing procedure consisting of two specific steps to determine whether to increase resources by combining the three judgment criteria.

[0150] Among the above criteria, 'response latency load' is a key criterion for directly measuring the degradation of service quality perceived by end users and is calculated based on the system's p95 query response latency; its purpose is to objectively identify the state of overall system performance degradation, excluding temporary delays. Additionally, among the above criteria, 'request load' is a criterion for measuring the amount of work currently applied to the system and is calculated based on the number of queries processed per second; its purpose is to quantify the system's current traffic level and predict potential overload conditions. Furthermore, among the above criteria, 'resource availability load' is a criterion for evaluating the status of resources handling the current load and is calculated based on the number of currently operating read-only database replicas; it executes a command aimed at more accurately diagnosing the load status by reflecting the technical fact that the burden placed on the system is greater when resources are scarce, even with the same request volume.

[0151] The control module (120) first calculates a normalized individual load factor based on system monitoring data corresponding to each judgment criterion, and the 'response delay load' is calculated based on the ratio between the collected current response delay time and the predefined 'target response time,' and the two values ​​have a directly proportional relationship. In addition, the 'request volume load' is calculated based on the number of queries processed per second collected, but the value is adjusted by applying a log-proportional relationship to prevent the load index from increasing excessively when the request volume explodes and to ensure stable control. Furthermore, the 'resource availability load' is calculated based on the number of database replicas currently in operation, and an operation is performed to set an inverse relationship so that the load is evaluated as higher as the number of replicas decreases. Next, the control module (120) calculates a 'comprehensive load index' by applying a preset importance weight to each of the three individual load elements calculated in the first step and summing them, and finally compares the calculated 'comprehensive load index' with a preset 'resource expansion threshold' and automatically generates a control signal to expand the resources of the database unit (110) only when the comprehensive load index exceeds the threshold.

[0152] Data regarding 'response latency', 'request volume', and 'resource availability' are collected in real-time through system monitoring tools, such as collecting p95 DB query response times at 1-minute intervals through an application performance monitoring system, collecting the number of queries processed per second at 1-minute intervals through the database server's internal performance counters, and querying the number of currently active read-only replica instances in real-time through the cloud service provider's management API. For example, calculations are performed assuming that at a specific point in time, a system status was measured with a p95 response latency of 500ms, 3,000 req / s of queries processed per second, and 2 active replicas.

[0153] If the calculated 'overall load index' exceeds the 'resource expansion threshold,' the control module (120) automatically executes a command to add one read-only database replica instance by calling the API of the cloud service provider. Additionally, until the system state stabilizes after resource expansion, it controls the monitoring data collection cycle to be shortened from 1 minute to 10 seconds to receive quick feedback. Furthermore, it executes a command to record detailed logs including the values ​​of each individual load factor and the overall load index at the time of resource expansion, to be used as data for future threshold optimization.

[0154] In a scenario involving a surge in traffic due to the release of a new album, the control module (120), having received input values ​​of a response delay of 500ms, 3,000 queries per second, 2 replicas, and weights of a delay of 0.5, a request volume of 0.3, and a resource of 0.2, calculates the 'response delay load' as '80' based on the fact that the response delay time is 2.5 times the target time, calculates the 'request volume load' as '75' by converting the number of queries per second according to a log-proportional relationship, and calculates the 'resource availability load' as '70' based on the number of active replicas according to an inverse relationship. Subsequently, the control module (120) calculates the 'overall load index' of 76.5 by applying weights to each individual load element and summing them, and compares this with the 'resource expansion threshold' of '70'. Finally, since the overall load index exceeds the threshold, it executes a resource expansion command that calls an API to add one DB replica instance and records the relevant details in the operation log.

[0155] The control module (120) ensures reliability by including a time-based filtering logic that executes actual resource expansion only when the 'comprehensive load index' exceeds the 'resource expansion threshold' for 3 minutes or longer, in order to prevent unnecessary resource expansion due to a short-term surge in traffic. Additionally, if the resource expansion command is not executed normally because the API call from the cloud service provider fails, it executes an exception handling command that retryes the same command after 5 minutes and automatically sends an emergency warning notification to the system administrator if it fails 3 consecutive times.

[0156] In a scenario of performance degradation due to slow queries comparing the operation results of the conventional technology and the present invention, when the system state is a CPU usage of 30%, a p95 response delay of 800ms, 500 req / s queries per second, and 4 replicas, the conventional technology based on a CPU usage of 80% or more takes no action and leaves the user's service delay unaddressed. On the other hand, the present invention generates data demonstrating a technical effect that dramatically improves service stability by automatically sending a 'slow query optimization needed' warning to relevant departments or increasing resources when the response delay load is calculated to be very high at 95 and the overall load index exceeds a threshold value.

[0157] The present invention stores a judgment rule table as shown in the following table in a database unit (110) and executes a command to classify the system state by referring to the table when calculating the comprehensive load index.

[0158]

[0159] The judgment rules of the present invention are based on objective performance indicators automatically measured by the system, thereby ensuring objectivity by not relying on the operator's subjective judgment or intuition. Furthermore, since static resource management methods relying on a single indicator in modern cloud-based service environments inevitably lead to service failures or wasted costs, adopting the model of the present invention, which dynamically manages resources by comprehensively considering service quality, load, and resource status, is a technical necessity for providing stable and efficient services.

[0160] The present invention adopts three judgment criteria—'response delay,' 'request volume,' and 'resource availability'—and a weighted summation-based comprehensive judgment rule to directly detect and rapidly respond to user-perceived performance degradation, thereby improving service stability, reducing unnecessary infrastructure costs by increasing resources only when actual load occurs, and providing the effect of reducing the average failure resolution time through an automated detection and response process.

[0161] In this specification, the 'resource expansion threshold' refers to a numerical value defined as a criterion for determining that the system is in an overloaded state and initiating a resource expansion procedure when the calculated comprehensive load index exceeds this value. Additionally, the 'resource expansion threshold' is statistically set based on the top 99.9th percentile of system load data over the past three months, and in the event that a major event is scheduled, a control command is transmitted to allow the administrator to set the threshold to 'aggressive mode' in advance and temporarily lower it by 15%.

[0162] The aforementioned 'resource expansion threshold' is set through statistical analysis of historical system load data. Specifically, it utilizes a dataset Q of 'overall load index' collected at one-minute intervals over the past three months, and determines the threshold value corresponding to the top 0.1% of this data distribution. This is calculated using the following percentile formula:

[0163] Resource expansion threshold = percentile(Q, 99.9)

[0164] This method provides objective criteria that do not respond to loads during normal peak times, but instead trigger resource expansion only in statistically exceptional overload situations. The table below shows the simulation results for 10,000 hypothetical load index data points.

[0165]

[0166] As a result of the simulation above, it can be seen that in 99.9% of the past data, the overall load index was less than 70. Therefore, setting the 'resource expansion threshold' to 70 is a statistically significant decision based on the data distribution and proves that it is not an arbitrary judgment by the operator.

[0167] To verify the effect of the set threshold of 70, load data immediately prior to 10 past events where service failures actually occurred is applied to the logic of the present invention to verify how effectively the 'overload' state is detected in advance before a failure occurs.

[0168]

[0169] As a result of the verification simulation above, it can be confirmed that in all events where actual failures occurred, the overall load index exceeded the threshold of 70 five minutes prior to the failure, successfully pre-detecting the 'overload' state. On the other hand, during normal peak time situations that did not lead to failures, the threshold was not exceeded, preventing unnecessary resource expansion (false positives). This clearly demonstrates that the threshold of 70 acts as an effective criterion for distinguishing between actual failure situations and normal overloads, and shows that the logic of the present invention substantially contributes to improving system stability.

[0171] Hereinafter, the entire operation process of the 'integrated management device for global artist discovery and management' according to one embodiment of the present invention is explained step-by-step through a hypothetical scenario, and the goal of the scenario is to discover a new artist with high potential, 'NVA', and to establish and execute an optimal overseas expansion strategy.

[0172] The device of the present invention begins operation by having a control module (120) monitor the status of the system in real time while processing user requests that have increased due to a new album promotion. The control module (120) receives data from an application performance monitoring system, a database server, and a cloud API, respectively, including a p95 response delay time of 500ms, a number of queries processed per second of 3,000 req / s, and a number of active replicas of 2. Based on the received data, the control module (120) initiates a logical processing procedure to calculate the 'response delay load', 'request volume load', and 'resource availability load', respectively, and then applies a preset weight to derive a 'comprehensive load index' of 76.5. Since the calculated index exceeds 70, which is the 'resource expansion threshold', the control module (120) calls the API of the cloud service provider to automatically execute a command to add one read-only database replica, thereby transmitting a command to secure a system environment in which complex analysis operations in subsequent steps can be performed stably.

[0173] After system stability is secured, the talent data management module (130) receives an audition video clip along with the profile of the newcomer 'NVA' from an external audition platform. The talent data management module (130) analyzes the received video clip through an artificial intelligence model to calculate a vocal score of 85 points based on pitch accuracy and a dance score of 90 points based on choreography matching rate, and simultaneously executes a command to combine the average star quality score of 70 points entered by three professional judges through the system and store them in the database unit (110). Next, the potential evaluation module (140) initiates a logical processing procedure to receive the evaluation score of 'NVA' and calculates a 'basic potential score' of 83.1 in the first step. Subsequently, in the second step, the potential evaluation module (140) calculates the standard deviation of the scores of 8.03 and converts it into a 'talent balance index' of 0.85 according to the 'judgment rule table'. In the final third step, a balance index is applied to the basic score to derive a 'final potential evaluation index' of 70.6, and this index is classified as 'Grade B' within the system, outputting information that clearly indicates to the person in charge that 'NVA' is a target for intensive development.

[0174] When the potential of 'NVA' is confirmed, the fandom analysis module (170) collects global fandom data of senior artists similar in genre to 'NVA' through external social media and streaming platform APIs. When data for a specific 'City A' is collected, such as 1 million followers, a weekly follower growth rate of +5%, and 2.5 million weekly streaming plays, the fandom analysis module (170) initiates a logical processing procedure. In the first step, the collected data is normalized to calculate a 'fandom size index' of 90, a 'growth momentum index' of 85, and a 'content consumption index' of 88. In the second step, it is confirmed that all three calculated indices exceed the 'minimum active threshold' of 50, and a 'overseas market expansion priority index' of 87.5 is derived through comprehensive calculation. As a result, 'City A' is classified as a 'top priority target' grade on the system's 'global priority map' and a screen is created to visually highlight it.

[0175] When 'City A' is determined as the top priority target, the control module (120) recognizes the analysis result of the fandom analysis module (170) as a trigger signal and transmits a command to initiate the operation of the promoter matching module (180). The promoter matching module (180) queries the database unit (110) to first filter a list of promoter candidates whose activity area is 'City A'. Next, the collaboration network management module (150) initiates a logical processing procedure for the filtered list of candidates, and for 'Partner Company A', a 'Final Collaboration Suitability Index' of 78, i.e., a 'Recommended Partner' grade, is calculated based on the number of past contracts, average response time, and physical distance data. Additionally, for 'Partner Company B', the final index is calculated as a 'Collaboration Caution' grade because although the number of past contracts is high, the response time and distance are long, and for 'Partner Company C', the final index is calculated as a 'Top Priority Recommendation' grade because although the number of contracts is low, the response time and distance are excellent. Finally, the promoter matching module (180) receives the suitability indices calculated from the collaboration network management module (150), sorts them, and outputs a screen that provides a ranking recommending 'Partner Company C' as 1st priority and 'Partner Company A' as 2nd priority to the person in charge.

[0176] At the same time as the person in charge begins consultation with 'Partner Company C', which was recommended as the first priority, the content production support module (160) automatically generates a customized content plan for 'NVA's' entry into 'City A'. The content production support module (160) receives an analysis result from the potential evaluation module (140) that 'NVA's' strengths lie in 'vocals' and 'dance', and receives an analysis result from the fandom analysis module (170) that the fandom of 'City A' shows a high preference for keywords such as 'lyrical melody' and 'powerful choreography'. The content production support module (160) synthesizes the received results, proposes a 'dance ballad' genre as the main concept that can maximize 'NVA's' vocal and dance capabilities, and generates a control signal to create and provide a related planning template and a list of demo track candidates.

[0177] Through the above steps, the present invention is characterized by organically performing a series of processes that dynamically respond to system load, discover and evaluate promising talents based on data, identify optimal overseas markets, recommend the most efficient partners in those markets, and further propose customized content strategies.

[0178] Although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above. It is understood that various modifications can be made by those skilled in the art without departing from the essence of the invention as claimed in the claims, and such modifications should not be understood individually from the technical spirit or perspective of the present invention. Explanation of the symbols

[0179] Database unit (110) control module (120) Talent data management module (130) Potential evaluation module (140) Collaboration network management module (150) Content creation support module (160) Fandom analysis module (170) Promoter matching module (180)

Claims

Claim 1 In a parallel MD agency device for utilizing artist content and diversifying revenue, the device comprises: a database unit; a control module that controls the operation of a plurality of function modules based on information stored in the database unit; a potential evaluation module that calculates a potential evaluation result including data on an artist's strengths and weaknesses based on quantitative scores for each evaluation item and a talent balance index; a collaboration network management module that stores information on multiple collaboration partners capable of collaboration, such as entertainment companies, broadcasters, investment firms, and promoters, in the database unit, and calculates a collaboration partner suitability index by quantifying the past project performance, response time, and physical distance of each collaboration partner; a content production support module that analyzes the genre characteristics, fandom preferences, and market trend data of a specific artist, and generates a control signal recommending the optimal genre and concept based on the strengths and weaknesses data of the artist derived from the evaluation result of the potential evaluation module; and a content production support module that calculates a fandom size index, a growth momentum index, and a content consumption index based on log-proportional or direct-proportional relationships, respectively, for a specific overseas region, and the fandom size index, the growth momentum A fandom analysis module that compares each of the index and the content consumption index with a preset minimum active threshold to classify the specific overseas region as a valid entry candidate only when the fandom size index, the growth momentum index, and the content consumption index are all above the minimum active threshold, calculates an overseas market expansion priority index for the valid entry candidates, and generates a global priority map sorted in index order;A promoter matching module that performs a multi-stage decision-making process of storing and managing information on overseas clients and promoters in the database unit, ranking multiple promoters using a collaboration partner suitability index calculated by the collaboration network management module, and recommending the promoter with the highest index as the top priority according to the ranking results; wherein the control module, when there is an overseas region classified as the top priority target grade in the global priority map generated by the fandom analysis module, automatically transmits a control signal to the promoter matching module to search for active promoters corresponding to the top priority target overseas region, and provides the promoter information recommended by the promoter matching module together with the potential evaluation results of the potential evaluation module to the content production support module, thereby triggering the content production support module to generate a content production strategy that reflects the artist's strengths and weaknesses data and the fandom characteristics of the top priority target overseas region, thereby causing the content production support module to recommend the optimal genre and concept corresponding to the content production strategy, and generate and provide a list of related planning templates and demo track candidates that can be utilized for MD product planning, performance planning, or digital content planning based on the recommendation results. MD agency parallel device for artist content utilization and revenue diversification, characterized by being configured to output as a module. Claim 2 In claim 1, the fandom analysis module analyzes the distribution of follower count, follower growth rate, and streaming playback count for an initial data set of artists who have successfully entered overseas markets in the past, statistically derives and sets the minimum active threshold by percentile analysis based on values ​​corresponding to a predetermined lower ratio in the said distribution, and displays regions classified into a predetermined upper grade among overseas regions satisfying or greater than the minimum active threshold in the said global priority map as the highest priority target grade; and the control module generates a trigger signal to recalculate the collaboration partner suitability index of the corresponding region-based partners recorded in the collaboration network management module whenever the overseas region displayed as the highest priority target grade is updated, and provides the recalculated result to the promoter matching module to update the partner ranking for the highest priority target overseas region, and at the same time causes the content production support module to synthesize the artist's strengths and weaknesses data derived from the updated global priority map, the partner ranking, and the evaluation results of the potential evaluation module, and to [develop] the artist's content into multiple types of businesses such as performances, campaigns, and collaboration projects in the highest priority target overseas region. MD agency parallel device for artist content utilization and revenue diversification, characterized by controlling to generate a content production strategy including multiple scenarios for deployment.

Images