Server for dynamically determining optimal selling price of unsold performance ticket and operation method

Abstract

The invention provides a server for dynamically determining the optimal selling price of an unsold performance ticket and an operation method. The server comprises a memory, a communication module and a processor, and the processor is configured to collect first ticket related data and first performance related data of performance tickets of a plurality of performance seats at at least one moment, and input the first ticket related data and the first performance related data into a first artificial intelligence model, predicting the expected sales data of the performance ticket at the sales cut-off time of the performance ticket; and inputting the remaining time, the remaining number, the second ticket related data and the second performance related data from the sales cut-off time of the non-sold performance tickets in the performance tickets into a second artificial intelligence model to determine the optimal selling price of the non-sold performance tickets.

Description

Technical Field

[0001] Various embodiments of the present invention relate to a server and its operation method for dynamically determining the optimal selling price of unsold performance tickets using an artificial intelligence model. Background Technology

[0002] Artificial intelligence (AI) systems are computer systems that achieve human-level intelligence. These machines can learn and make judgments independently, and their recognition rate improves with repeated use. AI technologies include machine learning (deep learning) techniques that use algorithms to classify / learn the features of input data, as well as element technologies that use machine learning algorithms to simulate the cognitive and judgmental functions of the human brain. Element technologies may include at least one of the following: language understanding technology that recognizes human language / text; visual understanding technology that recognizes objects like human vision; reasoning / prediction technology that judges information and performs logical reasoning and prediction; knowledge representation technology that processes human experience information into knowledge data; and motion control technology that controls autonomous driving of vehicles and the movement of robots. Summary of the Invention

[0003] The technical problem that the invention aims to solve

[0004] Various embodiments of the present invention can provide methods for using multiple artificial intelligence models to predict expected sales data for performance tickets, determine whether to dynamically change the price of unsold performance tickets, and determine the optimal selling price of unsold performance tickets.

[0005] Technical solutions for solving the problem

[0006] According to various embodiments, a server that dynamically determines the optimal selling price of unsold performance tickets using an artificial intelligence model may include a memory, a communication module, and a processor. The processor may be configured to: collect first ticket-related data and first performance-related data for multiple performance seats at at least one moment; predict expected sales data for the performance tickets at the sales deadline by inputting the first ticket-related data and the first performance-related data into a first artificial intelligence model; and when the expected sales data is less than a predetermined threshold, input the remaining time, remaining quantity, second ticket-related data, and second performance-related data from the sales deadline of the unsold performance tickets into a second artificial intelligence model to determine the optimal selling price of the unsold performance tickets. The first artificial intelligence model is trained based on the following data:

[0007] (1) Multiple ticket-related data and multiple performance-related data collected before the scheduled period starting from the sales deadline of multiple performance tickets, and (2) Multiple final ticket sales data collected at the sales deadline of the multiple performance tickets, the second artificial intelligence model is trained based on the following data: (1) the remaining time and remaining quantity of multiple unsold performance tickets before the sales deadline of the multiple unsold performance tickets, multiple ticket-related data and multiple performance-related data collected before the remaining time starting from the sales deadline of the multiple unsold performance tickets, and (2) multiple optimal selling prices.

[0008] According to various embodiments, a server operation method for dynamically determining the optimal selling price of unsold performance tickets using an artificial intelligence model includes the following steps: collecting first ticket-related data and first performance-related data for multiple performance seats at at least one moment; predicting expected sales data for the performance tickets at the sales deadline by inputting the first ticket-related data and the first performance-related data into a first artificial intelligence model; when the expected sales data is less than a predetermined threshold, inputting the remaining time, remaining quantity, second ticket-related data, and second performance-related data from the sales deadline of the unsold performance tickets into a second artificial intelligence model to determine the optimal selling price of unsold performance tickets. The first artificial intelligence model is trained based on the following data to determine the optimal selling price of the unsold performance tickets: (1) multiple ticket-related data and multiple performance-related data collected before the predetermined period starting from the sales deadline of the multiple performance tickets, and (2) multiple final ticket sales data collected at the sales deadline of the multiple performance tickets. The second artificial intelligence model is trained based on the following data: (1) the remaining time and remaining quantity of the multiple unsold performance tickets before the sales deadline of the multiple unsold performance tickets, multiple ticket-related data and multiple performance-related data collected before the remaining time starting from the sales deadline of the multiple unsold performance tickets, and (2) multiple optimal selling prices.

[0009] The effects of the invention

[0010] This invention employs a dynamic pricing strategy, where the price of unsold performance tickets fluctuates based on the remaining time before the ticket sales deadline and the number of remaining tickets. This allows buyers to purchase performance tickets at a more reasonable price, and sellers can also adjust ticket price fluctuations based on sales rates, ultimately achieving the effect of increasing total sales and reducing unsold tickets. Attached Figure Description

[0011] Figure 1 This is a block diagram of a user device and a server according to various embodiments of the present invention.

[0012] Figure 2This is a flowchart illustrating the steps of a server using an artificial intelligence model to determine the optimal selling price of unsold performance tickets in various embodiments of the present invention.

[0013] Figure 3 This illustrates an embodiment of a first artificial intelligence model for predicting expected sales data for performance tickets, according to various embodiments of the present invention.

[0014] Figure 4a This illustrates an embodiment of a second artificial intelligence model for determining the optimal selling price of unsold performance tickets according to various embodiments of the present invention.

[0015] Figure 4b The accompanying drawings are for illustrating embodiments of performance seats of the same class with different seat preferences according to various embodiments of the present invention.

[0016] Figure 5 This is a flowchart illustrating the steps of a server using an artificial intelligence model to determine whether to apply dynamic pricing to unsold performance tickets in various embodiments of the present invention. Detailed Implementation

[0017] The various embodiments described herein will now be explained with reference to the accompanying drawings. It should be understood that the examples and terminology used herein are not intended to limit the technology described herein to the particular embodiments, but rather to cover various modifications, equivalents, and / or alternatives to the corresponding embodiments. Similar reference numerals in the drawings may denote similar components. Singular expressions may include plural expressions unless the context clearly indicates otherwise. In this document, expressions such as “A or B” or “at least one of A and / or B” may include all possible combinations of the listed items. Expressions such as “first,” “second,” “firstly,” or “secondarily” may modify components regardless of their order or importance and are used only to distinguish one component from another, without limiting the components. When it is said that a component (e.g., a first) is “(functionally or communicatively) connected” or “integrated” into another component (e.g., a second), said component may be directly connected to said other component, or may be connected via other components (e.g., a third component).

[0018] In this document, “configured to” may be used interchangeably with “suitable,” “capable,” “modified to,” “made,” “able to,” or “designed to,” depending on the context, such as hardware or software aspects. In some cases, the expression “configured to” may mean that the device is “capable” of doing something together with other devices or components. For example, the phrase “processor configured to perform A, B, and C” can refer to a dedicated processor (e.g., an embedded processor) that performs the respective operations, or a general-purpose processor (e.g., a CPU or application processor) that can perform the respective operations by executing one or more software programs stored in a storage device.

[0019] User devices or electronic devices according to various embodiments of the present invention may include at least one of, for example, smartphones, tablets, desktop computers, laptops, netbooks, workstations, and servers.

[0020] refer to Figure 1 This section describes user device 100 and server 101 in various embodiments. User device 100 may include communication module 110, processor 120, memory 130, and display 140. In some embodiments, user device 100 may omit at least one of the plurality of components or further include other components.

[0021] The communication module 110 can establish communication between, for example, the user device 100 and an external device (e.g., a first external electronic device 102, a second external electronic device 104, or a server 101). For example, the communication module 110 can be connected to the network 180 via wireless or wired communication to communicate with the external device (e.g., the second external electronic device 104 or the server 101).

[0022] Wireless communication may include cellular communication using at least one of, for example, Long Term Evolution (LTE), LTE-A (LTE-A Advanced), CDMA (Code Division Multiple Access), WCDMA (Wideband CDMA), UMTS (Universal Mobile Telecommunications System), WiBro (Wireless Broadband), or GSM (Global System for Mobile Communications). In one embodiment, wireless communication may include at least one of, for example, Wireless Fidelity (WiFi), Bluetooth, Bluetooth Low Energy (BLE), Zigbee, Near Field Communication (NFC), Magnetic Secure Transmission, Radio Frequency (RF), or Body Area Network (BAN). In one embodiment, wireless communication may include GNSS. GNSS can be, for example, GPS (Global Positioning System), GLONASS (Global Navigation Satellite System), BeiDou Navigation Satellite System (hereinafter referred to as "BeiDou"), or the European global satellite-based navigation system Galileo. Hereinafter, "GPS" may be used interchangeably with "GNSS". Wired communication can include at least one of, for example, USB (Universal Serial Bus), HDMI (High Definition Multimedia Interface), RS-232 (Recommended Standard 232), power line communication, or POTS (Plain Old Telephone Service). Network 180 can include at least one of a telecommunications network, such as a computer network (e.g., a LAN or WAN), the Internet, or a telephone network.

[0023] Processor 120 may include one or more of a central processing unit, an application processor, or a communication processor (CP). Processor 120 may, for example, perform operations or data processing related to the control and / or communication of at least one other component of user device 100.

[0024] Memory 130 may include volatile memory and / or non-volatile memory. Memory 130 may store, for example, commands or data associated with at least one other component of user device 100.

[0025] Display 140 may include, for example, a liquid crystal display (LCD), a light-emitting diode (LED) display, an organic light-emitting diode (OLED) display, a microelectromechanical system (MEMS) display, or an electronic paper display. Display 140 may display various content to a user (e.g., text, images, videos, icons, and / or symbols). Display 140 may include a touchscreen and may receive touch, gesture, proximity, or hover input using an electronic pen or a part of the user's body.

[0026] The first external electronic device 102 and the second external electronic device 104 may be devices of the same or different types as the user device 100. According to various embodiments, all or part of the operations performed in the user device 100 may be performed in one or more other electronic devices (e.g., electronic devices 102, 104, or server 101). According to one embodiment, when the user device 100 wants to perform a function or service automatically or on request, the user device 100 may request at least some related functions from another device (e.g., electronic devices 102, 104, or server 101) instead of performing the function or service itself or additionally. The other electronic device (e.g., electronic devices 102, 104, or server 101) may perform the requested function or additional functions and transmit the results to the user device 100. The user device 100 may process the received results as is or additionally to provide the requested function or service. For this purpose, cloud computing, distributed computing, or client-server computing technologies may be used, for example.

[0027] Server 101 may include a communication module 111, a processor 121, and a memory 131. In some embodiments, server 101 may omit at least one component or include other components. The communication module 111, processor 121, and memory 131 may perform the same functions as the communication module 110, processor 120, and memory 130 in user device 100, respectively.

[0028] Figure 2For the purpose of illustrating various embodiments of the present invention, the server (e.g.: Figure 1 The flowchart of the steps for server 101 to determine the optimal selling price of unsold performance tickets using an artificial intelligence model.

[0029] Figure 3 This illustrates an embodiment of a first artificial intelligence model for predicting expected sales data for performance tickets, according to various embodiments of the present invention.

[0030] Figure 4a This illustrates an embodiment of a second artificial intelligence model for determining the optimal selling price of unsold performance tickets according to various embodiments of the present invention.

[0031] Figure 4b The accompanying drawings are for illustrating embodiments of performance seats of the same class with different seat preferences according to various embodiments of the present invention.

[0032] According to various embodiments, server 101 may run an application that allows user devices (e.g., Figure 1 Electronic devices 100, 102, 104 (e.g., PCs, laptops, smartphones, etc.) can purchase various types of performance tickets, communicate with user device 100 via networks 162, 164, process ticket purchase requests received from user device 100 through applications or web pages, and transmit the request information to user device 100. According to one embodiment, server 101 may include multiple devices for performing [operations related to] [other functions]. Figure 1 The components of the electronic device 100 have the same function. According to one embodiment, the component referred to as user device 100 in this invention may represent a user account that can access server 101 through the corresponding user device.

[0033] In step 201, according to various embodiments, server 101 (e.g., Figure 1 The processor 121) can collect first ticket-related data and first performance-related data for multiple performance seats at at least at any given time.

[0034] According to one embodiment, the first ticket-related data is all information related to the performance ticket and may include: information about the performance ticket's seat location, seat class (e.g., VIP, R, S, A, etc.), seat preference, ticket price, whether it is on sale, sales method (e.g., on-site purchase, telephone purchase, etc.), distribution method (e.g., QR code, physical ticket), sales data (e.g., sales rate, sales volume, sales amount, etc.), whether a promotional activity is being carried out, or discount data for each level of promotional activity.

[0035] According to one embodiment, the first performance-related data is all information related to the performance and may include at least one of the following: the type of performance (e.g., concert, drama, musical, etc.), the performers of the performance, the schedule of the performance (e.g., date / time, etc.), the location of the performance venue, the size of the performance venue (e.g., large, medium, small, etc.), information on other performances in the same time period (competitive performance information), or the marketing costs of the performance.

[0036] According to one embodiment, server 101 can also collect customer characteristic data for performance tickets. Customer characteristic data refers to customer characteristics that can be considered in the performance ticket sales process, and may include at least one of the following: age group, gender, place of residence, purchase history, preferred performance type, birthday / anniversary, occupation, and income level of the user (customer) who purchased the performance ticket.

[0037] According to one embodiment, at least one moment in which the first ticket-related data and the first performance-related data are collected may correspond to a moment preceding a predetermined period (e.g., 24 / 48 / 72 hours) before the deadline for ticket sales, or may correspond to ticket sales data (e.g., sales rate (n...)).

[0038] The sales figures (%), sales volume (m units), and sales revenue (k Korean Won) are the times when the predetermined sales data are reached. For example, at least one time can be at least one of the following: before the first period (e.g., 72 hours) predetermined from the ticket sales deadline (e.g., December 31, 2024, 18:00) (e.g., December 28, 2024, 18:00); before the second period (e.g., December 29, 2024, 18:00); or before the third period (e.g., December 30, 2024, 18:00). For example, this at least one time can correspond to at least one of the following: the time when the ticket sales data reaches the predetermined first sales data (e.g., sales rate n%), the time when the predetermined second sales data (e.g., sales rate n+5%), or the time when the predetermined third sales data (e.g., sales rate n+10%). The scheduled period or the scheduled sales data can be freely changed and set by the administrator of server 101.

[0039] In step 203, according to various embodiments, server 101 (e.g.: Figure 1 The processor 121 can input the first ticket-related data and the first performance-related data into the first artificial intelligence model to predict the expected sales data of the performance tickets at the sales deadline.

[0040] According to one embodiment, reference Figure 3When inputting first ticket-related data (e.g., seat location, seat class, current sales rate, price, whether it is on sale, sales channel, and distribution method) and first performance-related data (e.g., type of performance, performers, whether there are promotional activities, discount prices by class, performance schedule, performance venue location, competing performance information, and performance venue size) for multiple performance seats, the first artificial intelligence model can output expected sales data for the performance tickets (e.g., at least one of expected sales rate, expected sales volume, or expected sales revenue) as the output value.

[0041] According to various embodiments, the first artificial intelligence model can be trained based on the following data:

[0042] (1) Multiple ticket-related data and multiple performance-related data collected prior to the pre-determined period starting from the sales deadline for multiple performance tickets, and (2) multiple final ticket sales data collected at the sales deadline for multiple performance tickets. According to one embodiment, a first artificial intelligence model can learn the correlation between the multiple ticket-related data and the multiple performance-related data and the multiple final ticket sales data. For example, refer to... Figure 3 Server 101 can mark the multiple final ticket sales data as ground truth for the multiple ticket-related data and the multiple performance-related data in order to train the first artificial intelligence model.

[0043] According to one embodiment, server 101 may further utilize multiple customer feature data collected before a predetermined period starting from the sales deadline of multiple performance tickets to train a first artificial intelligence model, and input the customer feature data collected before the predetermined period starting from the sales deadline of performance tickets, together with first ticket-related data and first performance-related data, into the first artificial intelligence model to predict the expected sales data of performance tickets at the sales deadline of performance tickets.

[0044] According to one embodiment, server 101 obtains result values ​​(output data) using an artificial intelligence model with arbitrary weight values, compares the obtained result values ​​with the labeled data of the training data, and performs backpropagation based on the error to optimize the weight values, thereby training a first artificial intelligence model for predicting expected sales data of performance tickets. Specifically, training the first artificial intelligence model can refer to the process of training the first artificial intelligence model based on training data and labeled data, enabling the first artificial intelligence model to determine output data for input data. In other words, the first artificial intelligence model forms rules and makes judgments based on the data. According to one embodiment, during the training or prediction process of the first artificial intelligence model, server 101 can preprocess the aforementioned input data or training data before inputting each data point into the model to suit the data format used for learning or prediction in the model.

[0045] According to one embodiment, the first artificial intelligence model is a prediction model based on a time series model, and can be implemented in the form of polynomial linear, autoregressive (AR), moving average (MA), ARIMA (Autoregressive Integrated Moving Average), SARIMA (Seasonal ARIMA), SARIMAX (Seasonal ARIMA Exogenous), Prophet, LSTM (Long Short-Term Memory), RNN (Recurrent Neural Network), GRU (Gated Recurrent Unit), exponential smoothing (Transformer), transformer, etc. The implementation of the first artificial intelligence model is merely an example and is not limited thereto; various models feasible to those skilled in the art can be implemented. According to one embodiment, server 101 can use multiple learning algorithms from a variety of learning algorithms used to compute predicted values. For example, an ensemble method can be used on the prediction model, which can achieve better prediction performance than using a single learning algorithm.

[0046] According to one embodiment, the first artificial intelligence model may include multiple sub-artificial intelligence models trained by setting different predetermined periods starting from the sales cutoff times of multiple performance tickets. As another example, the first artificial intelligence model may include a first sub-artificial intelligence model trained based on data such as:

[0047] (1) Multiple ticket-related data and multiple performance-related data collected prior to the first period (e.g., 72 hours) starting from the sales deadline of multiple performance tickets, and (2) multiple final ticket sales data collected at the sales deadline of multiple performance tickets. Alternatively, the first artificial intelligence model may include a second sub-artificial intelligence model trained based on the following data: (1) multiple ticket-related data and multiple performance-related data collected prior to the second period (e.g., 48 hours) starting from the sales deadline of multiple performance tickets, and (2) multiple final ticket sales data collected at the sales deadline of multiple performance tickets. This example is merely one illustration; the number of multiple sub-artificial intelligence models can be freely created by the administrator of server 101.

[0048] According to one embodiment, when the first artificial intelligence model includes multiple sub-artificial intelligence models, the server 101 can apply weight values ​​to multiple expected sales data output by the multiple sub-artificial intelligence models according to mathematical formula 1 to calculate the final expected sales data.

[0049] Mathematical formula 1:

[0050]

[0051] Where Y can be the final expected sales data, w i It can be a weight value, a i It can be multiple sub-AI models.

[0052] According to one embodiment, server 101 can assign higher weight values ​​to sub-AI models trained for shorter predetermined periods starting from the sales deadlines of multiple performance tickets. For example, in this case, server 101 can apply a first weight value (e.g., 0.3) to a first sub-AI model trained on data collected before the sales deadlines of multiple performance tickets (e.g., 72 hours prior), and apply a second weight value (e.g., 0.7) greater than the first weight value to a second sub-AI model trained on data collected before the sales deadlines of a second predetermined period (e.g., 24 hours prior). This is because the shorter the period from the sales deadline, the higher the accuracy of the collected data in predicting the expected sales data of the performance tickets. According to one embodiment, the weight values ​​can be freely set by the administrator of server 101.

[0053] According to one embodiment, server 101 can use evaluation methods such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared to evaluate the result value of the first artificial intelligence model and train the first artificial intelligence model to minimize the loss function.

[0054] In step 205, according to various embodiments, server 101 (e.g.: Figure 1 The processor 121 can determine whether the expected sales data for the performance tickets is less than a predetermined threshold.

[0055] According to one embodiment, server 101 can determine whether at least one of the expected sales rate, expected sales volume, or expected sales revenue of performance tickets is less than a predetermined threshold as expected sales data for performance tickets. According to one embodiment, the predetermined threshold can be freely changed and set by the administrator of server 101.

[0056] According to one embodiment, server 101 can determine whether the final expected sales data calculated in step 203 is less than a predetermined threshold.

[0057] In step 207, according to various embodiments, server 101 (e.g.: Figure 1 The processor 121 can maintain the existing price for unsold tickets.

[0058] According to one embodiment, if the expected sales data for performance tickets are higher than a predetermined threshold, the server 101 can maintain the existing price of unsold performance tickets.

[0059] According to one embodiment, server 101 can transmit product information related to unsold performance tickets to user device via communication module 111 while maintaining the existing selling price of unsold performance tickets (e.g., ...). Figure 1 User device 100).

[0060] In step 209, according to various embodiments, server 101 (e.g.: Figure 1The processor 121) can input (1) the remaining time until the sales deadline, (2) the remaining quantity, (3) second ticket-related data, and (4) second performance-related data of unsold performance tickets into a second artificial intelligence model to determine the optimal selling price of the unsold performance tickets. According to one embodiment, when the expected sales data of performance tickets is less than a predetermined threshold, the server 101 can input (1) the remaining time until the sales deadline, (2) the remaining quantity, (3) the second ticket-related data, and (4) the second performance-related data into the second artificial intelligence model to determine the optimal selling price of the unsold performance tickets. According to one embodiment, the type of the second ticket-related data and the type of the second performance-related data can be at least some of the types of the first ticket-related data described in step 201, or at least some of the types of the first performance-related data.

[0061] According to one embodiment, server 101 may, starting from a time (e.g., 06:00 on December 31, 2024) before the sales deadline for unsold performance tickets (e.g., 18:00 on December 31, 2024), at predetermined time intervals (e.g., 1-hour intervals), determine the remaining time and quantity until the sales deadline, and collect second ticket-related data and second performance-related data.

[0062] In one embodiment, reference Figure 4a When inputting (1) the remaining time until the sales deadline, (2) the remaining quantity, (3) second ticket-related data (e.g., sales channel, price, seat class, seat preference), and (4) second performance-related data (e.g., performers, venue location, venue size, competing performance information, performance schedule, marketing costs) of unsold performance tickets, the second artificial intelligence model can output the optimal selling price of the unsold performance tickets as the output value. According to one embodiment, the seat preference, as the second ticket-related data, can be directly specified by the administrator of server 101, or it can correspond to the seat sales rate level of the corresponding seats. According to one embodiment, server 101 can set different seat preferences for each seat class. For example, refer to... Figure 4bEven if the first performance seat (e.g., seat 9) and the second performance seat (e.g., seat 14) have the same seat class (e.g., VIP), the seat preference of the first performance seat can be set higher than that of the second performance seat. In one embodiment, the optimal selling price of unsold performance tickets can be either the lowest first selling price or a second selling price that is considered appropriate. According to one embodiment, the administrator of server 101 can choose whether the second artificial intelligence model outputs the optimal selling price of unsold performance tickets as the lowest price or an appropriate price. In this case, the lowest price can refer to the price at which selling performance tickets below this price would result in a loss for the performance (profit price), while the appropriate price can refer to the price determined according to the pricing strategy required by the performance planning company, or the price determined by the business manager considering factors such as operating profit margin for each seat.

[0063] According to various embodiments, the second artificial intelligence model can be trained based on the following data: (1) multiple ticket-related data collected before the sales deadline of multiple unsold performance tickets, including the remaining time, remaining quantity, and remaining time before the sales deadline of multiple unsold performance tickets, as training input data; and (2) multiple optimal selling prices as ground truth data for labeling the multiple performance-related data, and training is performed using the multiple performance-related data as training input data. According to one embodiment, the second artificial intelligence model can learn the correlation between the training input data and the ground truth data. For example, refer to... Figure 4a Server 101 can use multiple ticket-related data and multiple performance-related data collected before the sales deadline for multiple unsold performance tickets, including the remaining time, remaining quantity, and remaining time before the sales deadline for multiple unsold performance tickets, to mark multiple optimal selling prices as ground truth, in order to train a second artificial intelligence model. According to one embodiment, the multiple optimal selling prices serving as ground truth data can be multiple minimum selling prices selected by the administrator of server 101, or appropriate selling prices obtained by adding tiered pre-sale prices to multiple minimum selling prices. That is, operators can choose minimum selling prices as prices that do not impair profitability, or choose appropriate selling prices considering the pricing strategy or operating profit margin of the event planning company. For example, server 101 can train the second artificial intelligence model using data such as (i) using the minimum selling prices selected by the operator as the optimal selling prices for unsold performance tickets (ground truth data), or (ii) using appropriate selling prices with tiered pre-sale prices added to the minimum selling prices, or appropriate selling prices selected by the operator, as the optimal selling prices for unsold performance tickets (ground truth data). The methods described above for determining the minimum or appropriate selling price are merely examples and are not limited to these methods. A variety of methods can be used to determine the minimum or appropriate selling price.

[0064] According to one embodiment, server 101 may further utilize multiple customer feature data collected before the remaining time before the sales deadline of multiple unsold performance tickets to train a second artificial intelligence model, and input the customer feature data collected before the remaining time before the sales deadline of multiple unsold performance tickets, the remaining time before the sales deadline of unsold performance tickets, the remaining quantity, second ticket-related data, and second performance-related data into the second artificial intelligence model to determine the optimal selling price of unsold performance tickets.

[0065] According to one embodiment, server 101 obtains result values ​​(output data) using an artificial intelligence model with arbitrary weight values, compares the obtained result values ​​with the labeled data of the training data, and performs backpropagation based on the error to optimize the weight values, thereby performing training of a second artificial intelligence model for determining the optimal selling price of unsold performance tickets. Specifically, training the second artificial intelligence model can refer to the process of training the second artificial intelligence model based on training data and labeled data, enabling the second artificial intelligence model to determine output data for input data. In other words, the second artificial intelligence model forms rules and makes judgments based on the data. According to one embodiment, during the training or prediction process of the second artificial intelligence model, server 101 can preprocess the aforementioned input data or training data before inputting each data point into the model to suit the data format used for learning or prediction in the model.

[0066] According to one embodiment, the second artificial intelligence model can be a machine learning-based predictive model, which may include a classification model and / or a regression model. The classification model can be implemented using decision trees, support vector machines, random forests, logistic regression, k-nearest neighbors, DNNs, etc., while the regression model can be implemented using linear regression, regression trees, support vector regression, kernel regression, etc. This second implementation of the artificial intelligence model is merely an example and is not limited to it; it can be implemented as various models feasible to those skilled in the art.

[0067] According to one embodiment, server 101 can use evaluation methods such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared to evaluate the result value of the second artificial intelligence model and train the second artificial intelligence model to minimize the loss function.

[0068] According to one embodiment, server 101 can train a second artificial intelligence model such that (1) as the remaining time before the sales deadline for unsold performance tickets decreases, the optimal selling price of unsold performance tickets decreases; and (2) as the remaining quantity of unsold performance tickets before the sales deadline decreases, the optimal selling price of unsold performance tickets increases. According to one embodiment, server 101 can train a second artificial intelligence model such that (1) as the remaining time before the sales deadline for unsold performance tickets decreases, the optimal selling price of unsold performance tickets increases; and

[0069] (2) As the number of unsold tickets decreases before the sales deadline, the optimal selling price of the unsold tickets decreases. The above are various embodiments of the present invention and are not limited thereto. The second artificial intelligence model can be trained by various combinations of conditions, such as remaining time and remaining quantity, selected by the administrator of server 101.

[0070] According to one embodiment, server 101 can train a second artificial intelligence model using second ticket-related data, such that (1) the higher the seat preference of unsold performance tickets, the smaller the prediction range of the optimal selling price of unsold performance tickets; and (2) the lower the seat preference of unsold performance tickets, the larger the prediction range of the optimal selling price of unsold performance tickets. According to one embodiment, server 101 can train a second artificial intelligence model using second ticket-related data, such that (1) the higher the seat preference of unsold performance tickets, the larger the prediction range of the optimal selling price of unsold performance tickets; and (2) the lower the seat preference of unsold performance tickets, the smaller the prediction range of the optimal selling price of unsold performance tickets. The above are various embodiments of the present invention and are not limited thereto. Furthermore, various combinations of seat preference conditions and optimal selling price conditions selected by the administrator of server 101 can be used to train the second artificial intelligence model.

[0071] According to one embodiment, server 101 can set an overall minimum price and an overall maximum price for the predicted range of optimal selling prices for unsold performance tickets, and train a second artificial intelligence model so that the optimal selling price of unsold performance tickets is determined to be between the overall minimum price and the overall maximum price. For example, server 101 can set an overall minimum price (e.g., 70,000 Korean won) and an overall maximum price (e.g., 150,000 Korean won) for the predicted range of optimal selling prices for unsold performance tickets.

[0072] According to one embodiment, server 101 can set a minimum and a maximum price for each seat class for the predicted range of the optimal selling price of unsold performance tickets, and train a second artificial intelligence model so that the optimal selling price of unsold performance tickets is determined between the minimum price and the maximum price of each seat class. For example, server 101 can determine the predicted range of the optimal selling price for unsold performance tickets of the first seating class (e.g., VIP) between the first minimum price (e.g., 150,000 KRW) and the first maximum price (e.g., 200,000 KRW), the predicted range of the optimal selling price for unsold performance tickets of the second seating class (e.g., R) between the second minimum price (e.g., 120,000 KRW) and the second maximum price (e.g., 145,000 KRW), the predicted range of the optimal selling price for unsold performance tickets of the third seating class (e.g., S) between the third minimum price (e.g., 90,000 KRW) and the third maximum price (e.g., 115,000 KRW), and the predicted range of the optimal selling price for unsold performance tickets of the fourth seating class (e.g., A) between the fourth minimum price (e.g., 40,000 KRW) and the fourth maximum price (e.g., 85,000 KRW). In this scenario, the highest optimal selling price (e.g., 145,000 KRW) for unsold tickets in the lower tier (e.g., R) can be set lower than the lowest optimal selling price (e.g., 150,000 KRW) for unsold tickets in the higher tier (e.g., VIP). The example above regarding the predicted range for optimal selling prices is merely illustrative and not limiting, and the optimal selling price can be determined by training a second artificial intelligence model using various methods. For instance, depending on the event planning company's pricing strategy, there may not be a predicted optimal selling price range by seat class; in such cases, the appropriate selling price for unsold tickets in the higher tier (e.g., VIP) may be lower than the appropriate selling price for unsold tickets in the lower tier (e.g., R).

[0073] According to one embodiment, when the expected sales rate, expected sales volume, and expected sales revenue of performance tickets are all less than predetermined thresholds, server 101 can input the remaining time before the sales deadline for unsold performance tickets, the remaining quantity, second ticket-related data, and second performance-related data into a second artificial intelligence model to determine the optimal selling price of the unsold performance tickets. For example, when the expected sales rate of performance tickets is less than a predetermined first threshold, the expected sales volume is less than a predetermined second threshold, and the expected sales revenue is less than a predetermined third threshold, server 101 can use the second artificial intelligence model to determine the optimal selling price of the unsold performance tickets.

[0074] According to one embodiment, when the final expected sales data calculated in step 203 is less than a predetermined threshold, the server 101 inputs the remaining time, remaining quantity, second ticket-related data, and second performance-related data from the sales deadline for unsold performance tickets into the second artificial intelligence model to determine the optimal selling price of the unsold performance tickets.

[0075] According to one embodiment, server 101 can transmit product information related to unsold performance tickets to user device 100 via communication module 111 when the selling price of unsold performance tickets is changed from the existing selling price to the optimal selling price.

[0076] In step 211, according to various embodiments, server 101 (e.g.: Figure 1 The processor 121 can update the optimal selling price of unsold tickets at predetermined time intervals until the sales deadline for unsold tickets.

[0077] According to one embodiment, server 101 can confirm the remaining time and quantity before the sales deadline at predetermined time intervals (e.g., 1-hour intervals), and collect second ticket-related data and second performance-related data until the sales deadline for unsold performance tickets.

[0078] According to one embodiment, server 101 can input (1) the remaining time until the sales deadline, (2) the remaining quantity, (3) second ticket-related data and (4) second performance-related data into a second artificial intelligence model at predetermined time intervals for unsold performance tickets, determine the subsequent optimal selling price of unsold performance tickets, and update the optimal selling price of unsold performance tickets at predetermined time intervals.

[0079] Figure 5 To illustrate servers according to various embodiments (e.g.: Figure 1 The flowchart of the steps for server 101 to use an artificial intelligence model to determine whether to apply dynamic pricing to unsold performance tickets.

[0080] In step 501, according to various embodiments, server 101 (e.g.: Figure 1 The processor 121) can collect first ticket-related data and first performance-related data for multiple performance seats at at least at any given time. According to one embodiment, the server 101 can use... Figure 2 The method described in step 201 is executed in step 501.

[0081] In step 503, according to various embodiments, server 101 (e.g.: Figure 1 The processor 121) can input the first ticket-related data and the first performance-related data into a prediction model based on a classification model to determine whether to apply a dynamic pricing strategy to unsold performance tickets at the end of the performance ticket sales period.

[0082] According to one embodiment, the prediction model based on the classification model can output the dynamic pricing strategy of unsold performance tickets at the sales deadline as the output value when inputting the first ticket-related data and the first performance-related data of multiple performance seats. (e.g., "Applied (1)" or "Not Applied (0)").

[0083] According to various embodiments, the classification-based predictive model can be trained based on (1) multiple ticket-related data and multiple performance-related data collected prior to the pre-booking period starting from the sales cutoff time of multiple performance tickets, and (2) whether the administrator of server 101 applies a dynamic pricing strategy to multiple unsold performance tickets. According to one embodiment, the classification-based predictive model can learn the correlation between the multiple ticket-related data and the multiple performance-related data and whether the dynamic pricing strategy is applied. As another example, server 101 can mark whether the dynamic pricing strategy is applied as ground truth for the multiple ticket-related data and the multiple performance-related data to train the classification-based predictive model.

[0084] According to one embodiment, server 101 may further utilize multiple customer feature data collected prior to the pre-scheduled period starting from the sales deadline of multiple performance tickets to train a classification-based prediction model, and input the customer feature data collected prior to the pre-scheduled period starting from the sales deadline of performance tickets, along with first ticket-related data and first performance-related data, into the prediction model to determine whether to apply a dynamic pricing strategy to unsold performance tickets.

[0085] According to one embodiment, server 101 can train a classification-based predictive model for determining whether to apply a dynamic pricing strategy by using an artificial intelligence model with arbitrary weights to obtain result values ​​(output data), comparing the obtained result values ​​with the labeled data of the training data, and backpropagating based on the error to optimize the weight values. Specifically, training the predictive model can refer to training the predictive model based on the training data and labeled data, enabling the predictive model to determine the output data for the input data.

[0086] In step 505, according to various embodiments, when a dynamic pricing strategy is not applied to unsold performance tickets, then server 101 (e.g.: Figure 1 The processor 121 can maintain the existing price for unsold performance tickets. According to one embodiment, the server 101 can use... Figure 2 The method described in step 207 is executed in step 505.

[0087] In step 507, according to various embodiments, when a dynamic pricing strategy is applied to unsold performance tickets, server 101 (e.g.: Figure 1 The processor 121) can input (1) the remaining time until the sales deadline, (2) the remaining quantity, (3) second ticket-related data, and (4) second performance-related data of unsold performance tickets into a second artificial intelligence model to determine the optimal selling price of the unsold performance tickets. According to one embodiment, the server 101 can use... Figure 2 The methods described in steps 209 and 211 execute the method in step 507 for determining the optimal selling price of unsold performance tickets.

[0088] According to various embodiments, a server that dynamically determines the optimal selling price of unsold performance tickets using an artificial intelligence model includes a memory, a communication module, and a processor. The processor can be configured to: collect first ticket-related data and first performance-related data for multiple performance seats at at least one moment; input the first ticket-related data and the first performance-related data into a first artificial intelligence model to predict the expected sales data of the performance tickets at the sales deadline; and when the expected sales data is less than a predetermined threshold, input the remaining time, remaining quantity, second ticket-related data, and second performance-related data remaining until the sales deadline of the unsold performance tickets into a second artificial intelligence model. An intelligent model is used to determine the optimal selling price of the unsold performance tickets. The first artificial intelligence model is trained based on the following data: (1) multiple ticket-related data and multiple performance-related data collected before the predetermined period starting from the sales deadline of the multiple performance tickets, and (2) multiple final ticket sales data collected at the sales deadline of the multiple performance tickets. The second artificial intelligence model is trained based on the following data: (1) the remaining time and remaining quantity of the multiple unsold performance tickets before the sales deadline of the multiple unsold performance tickets, multiple ticket-related data and multiple performance-related data collected before the remaining time starting from the sales deadline of the multiple unsold performance tickets, and (2) multiple optimal selling prices.

[0089] According to various embodiments, the at least one moment may be a moment preceding a predetermined period from the sales cutoff time.

[0090] According to various embodiments, the first artificial intelligence model can be a prediction model based on a time series model, and the second artificial intelligence model can be a prediction model based on machine learning. The first ticket-related data and the second ticket-related data respectively include: information about at least one of the following: seat location, class, price, whether it is on sale, sales channel, distribution method, sales data, whether a promotional activity is being carried out, or discount data for each level of the promotional activity. The first performance-related data and the second performance-related data respectively include: the type of performance corresponding to the performance ticket, the performers of the performance, the schedule of the performance, the location of the performance venue, the size of the performance venue, information on other performances in the same time period as the performance, or the marketing cost of the performance.

[0091] According to various embodiments, the processor can be configured to: when the first artificial intelligence model includes multiple sub-artificial intelligence models trained by setting different predetermined periods starting from the sales deadline of the multiple performance tickets, apply weight values ​​to multiple expected sales data output from the multiple sub-artificial intelligence models according to Mathematical Formula 1, calculate final expected sales data, and when the final expected sales data is less than the predetermined threshold, input the remaining time, remaining quantity, second ticket-related data, and second performance-related data from the sales deadline of the unsold performance tickets into the second artificial intelligence model to determine the optimal selling price of the unsold performance tickets, wherein the shorter the predetermined period from the sales deadline of the multiple performance tickets, the higher the weight value of the trained sub-artificial intelligence model.

[0092] Mathematical formula 1:

[0093]

[0094] Where Y represents the final expected sales data, w i Let a be the weight value. i These are the multiple sub-artificial intelligence models.

[0095] According to various embodiments, the processor can be configured to: train the second artificial intelligence model to: (1) determine the optimal selling price of the unsold performance tickets in a first direction as the remaining time before the sales deadline of the unsold performance tickets is less, and (2) determine the optimal selling price of the unsold performance tickets in a second direction as the remaining quantity before the sales deadline of the unsold performance tickets is less, and train the second artificial intelligence model to: as second ticket-related data, (1) change the prediction range of the optimal selling price of the unsold performance tickets in a first state as the seat preference of the unsold performance tickets is higher, and (2) change the prediction range of the optimal selling price of the unsold performance tickets in a state opposite to the first state as the seat preference of the unsold performance tickets is lower.

[0096] According to various embodiments, the processor can be configured to: set a minimum price and a maximum price for each seat class for the predicted range of the optimal selling price of the unsold performance tickets, determine the optimal selling price of the unsold performance tickets between the minimum price and the maximum price of each seat class, and the maximum price of the optimal selling price of the unsold performance tickets in the lower class is lower than the minimum price of the optimal selling price of the unsold performance tickets in the higher class.

[0097] According to various embodiments, a method for operating a server that dynamically determines the optimal selling price of unsold performance tickets using an artificial intelligence model may include the following steps: collecting first ticket-related data and first performance-related data for multiple performance seats at at least one moment; predicting expected sales data for the performance tickets at the sales deadline by inputting the first ticket-related data and the first performance-related data into a first artificial intelligence model; when the expected sales data is less than a predetermined threshold, inputting the remaining time, remaining quantity, second ticket-related data, and second performance-related data from the sales deadline of the unsold performance tickets into a second artificial intelligence model, so as to... To determine the optimal selling price of the unsold performance tickets, the first artificial intelligence model is trained based on the following data: (1) multiple ticket-related data and multiple performance-related data collected before the predetermined period starting from the sales deadline of the multiple performance tickets, and (2) multiple final ticket sales data collected at the sales deadline of the multiple performance tickets. The second artificial intelligence model is trained based on the following data: (1) the remaining time and remaining quantity until the sales deadline of the multiple unsold performance tickets, multiple ticket-related data and multiple performance-related data collected before the remaining time starting from the sales deadline of the multiple unsold performance tickets, and (2) multiple optimal selling prices.

[0098] As used herein, the term "module" or "part" includes a unit comprised of hardware, software, or firmware, and is used interchangeably with terms such as logic, logic block, component, or circuit. A "module" or "part" can be a component as a whole or a smallest unit or part performing one or more functions. A "module" or "part" can be implemented mechanically or electronically and may include, for example, known or future-developed application-specific integrated circuit (ASIC) chips, field-programmable gate arrays (FPGAs), or programmable logic devices that perform certain operations and can be executed by processor 120. At least a portion of an apparatus (e.g., a module or its function) or method (e.g., an operation) according to various embodiments can be implemented as instructions stored as program modules in a computer-readable recording medium (e.g., memory 130). When the instructions are executed by a processor (e.g., processor 120), the processor can perform the function corresponding to the instruction. Computer-readable recording media may include hard disks, floppy disks, magnetic media (e.g., magnetic tape), optical recording media (e.g., CD-ROMs, DVDs), magneto-optical media (e.g., optical discs), internal memory, etc. Instructions may include compiler-generated code or interpreter-executable code. Modules or program modules according to various embodiments may include at least one or more of the aforementioned components, some of which may be omitted, or other components may be included. Operations performed by modules, program modules, or other components according to various embodiments may be performed sequentially, in parallel, iteratively, or heuristically, or at least some operations may be performed in a different order, these operations may be omitted, or other operations may be added.

[0099] Furthermore, the embodiments disclosed herein are provided for the purpose of explaining and understanding the disclosed technical content and do not limit the scope of the invention. Therefore, the scope of the invention should be interpreted to include all modifications or various other embodiments based on the technical concept of the invention.

Claims

1. A server for dynamically determining the optimal selling price of unsold performance tickets, characterized in that, include: memory, Communication module; processor, The processor is configured to: Collect first ticket-related data and first performance-related data for multiple performance seats at at least one time. By inputting the first ticket-related data and the first performance-related data into a first artificial intelligence model, the expected sales data of the performance tickets at the sales deadline are predicted. When the expected sales data is less than a predetermined threshold, the remaining time and quantity of unsold tickets, second ticket-related data, and second performance-related data are input into a second artificial intelligence model to determine the optimal selling price of the unsold tickets. The first artificial intelligence model is trained based on the following data: (1) multiple ticket-related data and multiple performance-related data collected before the predetermined period starting from the sales deadline of multiple performance tickets, and (2) multiple final ticket sales data collected at the sales deadline of the multiple performance tickets. The second artificial intelligence model is trained based on the following data: (1) the remaining time and remaining quantity of multiple unsold performance tickets before the sales deadline, multiple ticket-related data and multiple performance-related data collected before the remaining time begins from the sales deadline of the multiple unsold performance tickets, and (2) multiple optimal selling prices.

2. The server according to claim 1, characterized in that, The at least one moment is a moment preceding a predetermined period from the sales deadline.

3. The server according to claim 1, characterized in that, The first artificial intelligence model is a prediction model based on time series models. The second artificial intelligence model is a prediction model based on machine learning. The first ticket-related data and the second ticket-related data respectively include: information on at least one of the following: seat location, class, price, whether it is on sale, sales channel, distribution method, sales data, whether a promotional activity is being carried out, or discount data for each level of promotional activity. The first performance-related data and the second performance-related data respectively include at least one of the following: the type of performance corresponding to the performance ticket, the performers of the performance, the schedule of the performance, the location of the performance venue, the size of the performance venue, information on other performances in the same time period as the performance, or the marketing cost of the performance.

4. The server according to claim 1, characterized in that, The processor is configured to: When the first artificial intelligence model includes multiple sub-artificial intelligence models trained by setting different predetermined periods starting from the sales cutoff time of the multiple performance tickets, weight values ​​are applied to multiple expected sales data output from the multiple sub-artificial intelligence models according to mathematical formula 1 to calculate the final expected sales data. When the final expected sales data is less than the predetermined threshold, the remaining time, remaining quantity, second ticket-related data, and second performance-related data before the sales deadline for the unsold performance tickets are input into the second artificial intelligence model to determine the optimal selling price of the unsold performance tickets. The shorter the predetermined period from the sales deadline for the multiple performance tickets, the higher the weight value of the trained sub-AI model. Mathematical formula 1: Where Y represents the final expected sales data, w i Let a be the weight value. i These are the multiple sub-artificial intelligence models.

5. The server according to claim 1, characterized in that, The processor is configured to: The second artificial intelligence model is trained to: (1) determine the optimal selling price of the unsold performance tickets in the first direction as the remaining time before the sales deadline of the unsold performance tickets is shorter; and (2) determine the optimal selling price of the unsold performance tickets in the second direction as the remaining quantity before the sales deadline of the unsold performance tickets is shorter. The second artificial intelligence model is trained to: (1) the higher the seat preference of the unsold performance ticket, the more the prediction range of the optimal selling price of the unsold performance ticket will change towards the first state, and (2) the lower the seat preference of the unsold performance ticket, the more the prediction range of the optimal selling price of the unsold performance ticket will change towards the opposite state to the first state.

6. The server according to claim 1, characterized in that, The processor is configured to: To predict the optimal selling price range for the unsold performance tickets, minimum and maximum prices are set for each seating class. The second artificial intelligence model is trained to determine the optimal selling price of the unsold performance tickets between the lowest and highest prices for each seating class. The highest price of the optimal selling price for unsold tickets in the lower tier is lower than the lowest price of the optimal selling price for unsold tickets in the higher tier.

7. A method for operating a server to dynamically determine the optimal selling price of unsold performance tickets, comprising the following steps: Collect first ticket-related data and first performance-related data for multiple performance seats at at least one time. By inputting the first ticket-related data and the first performance-related data into the first artificial intelligence model, the expected sales data of the performance tickets at the sales deadline are predicted; When the expected sales data is less than a predetermined threshold, the remaining time and quantity of unsold tickets, second ticket-related data, and second performance-related data are input into a second artificial intelligence model to determine the optimal selling price of the unsold tickets. The first artificial intelligence model is trained based on the following data: (1) multiple ticket-related data and multiple performance-related data collected before the predetermined period starting from the sales deadline of multiple performance tickets, and (2) multiple final ticket sales data collected at the sales deadline of the multiple performance tickets. The second artificial intelligence model is trained based on the following data: (1) the remaining time and remaining quantity of multiple unsold performance tickets before the sales deadline, multiple ticket-related data and multiple performance-related data collected before the remaining time begins from the sales deadline of the multiple unsold performance tickets, and (2) multiple optimal selling prices.

Images