Device for adjusting sales price on basis of reservation probability and control method thereof

Abstract

A server device disclosed in the present document comprises: a communication unit for receiving room data for training a machine learning model related to room sales price adjustment; a memory for storing the room data; and a control unit for extracting a feature for room sales price adjustment from the room data to generate preprocessing data, generating a room reservation probability determination model and a room sales price adjustment model using the preprocessing data as training data, deriving a room reservation probability as an output value by inputting the room data into the room reservation probability determination model, deriving room sales price adjustment data as an output value by inputting the room reservation probability into the room sales price adjustment model, and transmitting the room sales price adjustment data to an external device.

Description

Sales price adjustment device based on reservation probability and its control method

[0001] Cross-citation with related applications

[0002] This application claims the benefit of priority to Republic of Korea Patent Application No. 10-2023-0101831, filed August 3, 2023, and Republic of Korea Patent Application No. 10-2023-0153927, filed November 8, 2023, the entire contents of which are incorporated herein by reference.

[0003] Technology field

[0004] The embodiments disclosed in this document relate to a server device and a control method of the server device for dynamically adjusting a room sales price.

[0005] Accommodation products are characterized by their inability to increase or decrease supply in the short term and by a fixed sales deadline. Sellers of accommodation products aim to maximize profits by optimizing their selling prices. In particular, for accommodation products where demand fluctuates significantly depending on the date or time, adjusting prices accordingly is crucial for generating higher sales.

[0006] For example, during low-demand periods, lowering the selling price increases the likelihood of a sale, while during high-demand periods, raising the selling price does not reduce the likelihood of a sale. However, typical selling price suggestion systems have limitations in accurately measuring or predicting rapidly changing demand, making them incapable of setting optimal selling prices.

[0007] Additionally, for competitive room models that dynamically adjust the prices of accommodation products, the proposed price differs significantly from the actual sales price, and there are limitations to the advanced work to improve the model's performance, so the introduction of a new model is necessary.

[0008] According to one embodiment disclosed in this document, a server device and a control method of the server device are provided for dynamically adjusting the sales price of a room using a machine learning algorithm based on room data.

[0009] The technical problems of the embodiments disclosed in this document are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the descriptions below.

[0010] According to one embodiment, a server device includes a communication unit that receives room data for training a machine learning model related to room sales price adjustment, a memory that stores the room data, a control unit that extracts features for room sales price adjustment from the room data to generate preprocessed data, generates a room reservation probability determination model and a room sales price adjustment model using the preprocessed data as training data, inputs the room data into the room reservation probability determination model to derive a room reservation probability as an output value, and inputs the derived room reservation probability into the room sales price adjustment model to derive room sales price adjustment data as an output value.

[0011] The control unit may determine a product demand index and a regional demand index based on the room reservation probability, and input the product demand index and the regional demand index into the room sales price adjustment model to derive the room sales price adjustment data as an output value.

[0012] The above control unit can determine the average value of the business demand index and the room demand index including the room forecast reservation rate (sale rate) and the room pickup rate as the product demand index.

[0013] The above room reservation probability determination model may include a reservation occurrence probability prediction model that predicts the probability of a reservation occurring on a specific stay date for each room, and a reservation occurrence count prediction model that predicts the number of reservations occurring on a specific stay date for each business.

[0014] The above control unit can input the room data into the reservation occurrence number prediction model to output the predicted number of reservations for the room, and determine the predicted reservation rate for the room by dividing the predicted number of reservations for the room by the remaining number of rooms.

[0015] The above control unit can determine the room pickup rate by dividing the number of rooms actually sold from the room data by the total number of rooms.

[0016] The above control unit can input the room data into the reservation occurrence probability prediction model to output the reservation probability of the room, and determine the reservation probability of the room as the room demand index.

[0017] The above room sales price adjustment model may include an AEEF (Asymmetric Exponent-Exponential Form) model having the product demand index and the regional demand index as variables.

[0018] The control unit can modify the room sales price adjustment data based on receiving at least one of the lowest room price line and the price priority between rooms through the communication unit.

[0019] The above control unit can modify the room sales price adjustment data based on an amount obtained by excluding the user's coupon usage amount from the room sales price adjustment data.

[0020] The above control unit can transmit the room sales adjustment data to an external device.

[0021] A control method of a server device according to one embodiment may include receiving room data for training a machine learning model related to room sales price adjustment, extracting features for room sales price adjustment from the room data to generate preprocessed data, generating a room reservation probability determination model and a room sales price adjustment model using the preprocessed data as training data, inputting the room data into the room reservation probability determination model to derive a room reservation probability as an output value, inputting the room reservation probability into the room sales price adjustment model to derive room sales price adjustment data as an output value, and transmitting the room sales price adjustment data to an external device.

[0022] The above room sales price adjustment data can be derived as an output value by determining a product demand index and a regional demand index based on the room reservation probability, and inputting the product demand index and the regional demand index into the room sales price adjustment model to derive the room sales price adjustment data as an output value.

[0023] In determining the above product demand index, the average value of the business demand index and the room demand index, including the room forecast reservation rate (sale rate) and the room pickup rate, can be determined as the above product demand index.

[0024] The above room reservation probability determination model may include a reservation occurrence probability prediction model that predicts the probability of a reservation occurring on a specific stay date for each room, and a reservation occurrence count prediction model that predicts the number of reservations occurring on a specific stay date for each business.

[0025] Predicting the above-mentioned room prediction reservation rate can be accomplished by inputting the room data into the reservation occurrence number prediction model to output the predicted number of reservations for the room, and dividing the predicted number of reservations for the room by the remaining number of rooms to determine the predicted reservation rate for the room.

[0026] The above room pick-up rate can be determined by dividing the number of rooms actually sold from the above room data by the total number of rooms.

[0027] Determining the above room demand index can be accomplished by inputting the room data into the reservation occurrence probability prediction model to output the reservation probability of the room, and determining the reservation probability of the room as the room demand index.

[0028] The above room sales price adjustment model may include an AEEF (Asymmetric Exponent-Exponential Form) model having the product demand index and the regional demand index as variables.

[0029] A control method of a server device according to one embodiment may further include modifying the room sales price adjustment data based on receiving at least one of a lowest room price line and a price priority between rooms through a communication unit.

[0030] A control method of a server device according to one embodiment may further include lowering the room sales price adjustment data based on an amount excluding the user's coupon usage amount from the room sales price adjustment data.

[0031] A control method of a server device according to one embodiment can transmit the room sales adjustment data to an external device.

[0032] According to one embodiment of the present invention, a server device can reduce the cost of human resources of a seller and maximize sales by automating sales price adjustments in response to changes in commercial demand.

[0033] FIG. 1 illustrates a control block diagram of a server device according to one embodiment.

[0034] FIG. 2 illustrates input and output values ​​of a machine learning model utilized in a server device according to one embodiment.

[0035] Figures 3 and 4 illustrate a schematic flowchart of a server device adjusting a room sales price according to one embodiment.

[0036] FIG. 5 illustrates features utilized in a server device according to one embodiment.

[0037] FIG. 6 illustrates factors utilized by a control unit of a server device according to one embodiment to adjust a room sales price.

[0038] FIG. 7 illustrates the performance of a room sales adjustment model of a server device according to one embodiment.

[0039] FIG. 8 illustrates an example in which the sales price of a room is adjusted in a server device according to one embodiment.

[0040] FIG. 9 illustrates a method for verifying a room sales adjustment result in a server device according to one embodiment.

[0041] FIG. 10 illustrates a graph comparing the results of room sales adjustment in a server device according to one embodiment with those of an existing model.

[0042] Fig. 11 is a flowchart of a method for controlling a server device according to one embodiment.

[0043] FIG. 12 is a drawing that continues the flowchart of a method for controlling a server device according to one embodiment of FIG. 11.

[0044] Hereinafter, various embodiments disclosed in this document will be described in detail with reference to the attached drawings. In this document, identical components in the drawings are designated by the same reference numerals, and redundant descriptions of identical components are omitted.

[0045] With respect to the various embodiments disclosed in this document, specific structural and functional descriptions are merely illustrative for the purpose of explaining the embodiments, and the various embodiments disclosed in this document may be implemented in various forms and should not be construed as being limited to the embodiments described in this document.

[0046] The expressions "first," "second," "first," or "second" used in various embodiments may describe various components, regardless of order and / or importance, and do not limit the components. For example, without departing from the scope of the embodiments disclosed herein, a first component may be renamed a second component, and similarly, a second component may also be renamed a first component.

[0047] The terms used in this document are intended solely to describe specific embodiments and may not be intended to limit the scope of other embodiments. Singular expressions may include plural expressions unless the context clearly dictates otherwise.

[0048] All terms used herein, including technical or scientific terms, have the same meaning as commonly understood by those of ordinary skill in the art of the embodiments disclosed herein. Terms defined in commonly used dictionaries may be interpreted as having the same or similar meaning in the context of the relevant technology, and unless explicitly defined herein, they shall not be interpreted in an idealized or overly formal sense. In some cases, even if a term is defined herein, it cannot be interpreted to exclude the embodiments disclosed herein.

[0049] The operating principle and embodiments of the present invention will be described with reference to the attached drawings below.

[0050] FIG. 1 illustrates a control block diagram of a server device according to one embodiment.

[0051] Referring to FIG. 1, a server device (1) according to one embodiment includes a control unit (100) including at least one processor (110) and a memory (120) and a communication unit (200), and can dynamically adjust a room price by communicating with an external device (2) through the communication unit (200).

[0052] A server device (1) according to one embodiment may be implemented as various computing devices such as a workstation, a cloud, a data drive, a data station, etc. In addition, the server device (1) may be implemented as one or more server devices (1) that are physically or logically separated based on function, detailed configuration of function, data, etc., and data may be transmitted and received and the transmitted and received data may be processed through communication between each server device (1).

[0053] An external device (2) communicating with a server device (1) may include a user terminal and an external server device that transmits a room price adjusted by the server device (1).

[0054] Specifically, when the external device (2) is a user terminal, the control unit (100) of the server device (1) can transmit the result of the room price adjustment to the user terminal so that the user can check it. At this time, the user terminal may include, but is not limited to, a personal computer, a terminal, a portable telephone, a smart phone, a handheld device, a wearable device, etc.

[0055] In addition, when the external device (2) is an external server device, the external server device can be implemented as various computing devices, and the external server device can be implemented as one or more external server devices, and data can be transmitted and received and the transmitted and received data can be processed through communication between each external server device.

[0056] A server device according to one embodiment may mean any electronic device including a processor (110) and memory (120). Each component of the server device is described in detail below.

[0057] The communication unit (200) may include a wireless communication unit (210) and a wired communication unit (220) to communicate with an external device (2). The communication unit (200) may receive room data to be used as learning data from a separately provided external device (2), or may transmit and receive programs and various data for machine learning.

[0058] The wireless communication unit (210) may include at least one of a short-range communication module and a long-range communication module.

[0059] The short-range communication module can communicate with an external device (2) adjacent to the server device using a short-range communication method. Here, the short-range communication module can utilize one of the following communication methods: Bluetooth, Bluetooth low energy, infrared data association (IrDA), Zigbee, Wi-Fi, Wi-Fi direct, Ultra Wideband (UWB), or near field communication (NFC).

[0060] The remote communication module may include a communication module that performs various types of remote communication and may include a mobile communication unit. The mobile communication unit may transmit and receive a wireless signal with at least one of a base station, an external terminal, and an external device (2) on a mobile communication network. In addition, the remote communication module may communicate with an external device (2) or an external device (2) such as another electronic device through a surrounding access point (AP). The access point (AP) may connect a local area network (LAN) to which a server device is connected to a wide area network (WAN) to which a communication server is connected. Accordingly, the server device may be connected to the communication server through the wide area network (WAN) with the external device (2) and communicate with each other.

[0061] The wired communication unit (220) can connect to a wired communication network and communicate with an external device (2) through the wired communication network. For example, the wired communication unit (220) can connect to a wired communication network through Ethernet (IEEE 802.3 technology standard) or connect to a wired communication network through CAN communication, and transmit and receive data with the external devices (2) through the wired communication network.

[0062] A server device according to one embodiment may include an input / output interface (not shown). An interface may be provided that connects an input device (not shown) such as a keyboard, mouse, or touch panel, an output device (not shown) such as a display, and a processor (110) to transmit and receive data.

[0063] The memory (120) can store various types of information required for the operation of the server device. Specifically, the memory (120) can store an operating system and programs required for the operation of the server device, or store data required for the operation of the server device.

[0064] Specifically, the memory (120) can store various programs related to room data preprocessing, room sales price adjustment, and machine learning algorithms. Furthermore, the memory (120) can store feature data utilized in machine learning.

[0065] The memory (120) may include volatile memory (120) such as Static Random Access Memory (S-RAM) and Dynamic Random Access Memory (D-RAM) for temporarily storing data. In addition, the memory (120) may include nonvolatile memory (120) such as Read Only Memory (ROM), Erasable Programmable Read Only Memory (EPROM), and Electrically Erasable Programmable Read Only Memory (EEPROM) for long-term storage of data.

[0066] The processor (110) outputs control signals to control the overall server device. The processor (110) may include one or more central processing units (CPUs) and graphics processing units (GPUs). In this case, the processor (110) may be implemented as an array of a plurality of logic gates, or may be implemented as a combination of a general-purpose microprocessor (110) and a memory (120) storing a program that can be executed on the microprocessor (110).

[0067] The aforementioned memory (120) and processor (110) may be included in the control unit (100), and the control unit (100) may control the aforementioned components to determine whether to increase or decrease the room sales price.

[0068] Specifically, the control unit (100) can extract features for adjusting room sales prices from room data to generate preprocessed data, and generate a room reservation probability determination model and a room sales price adjustment model that use the preprocessed data as learning data. That is, the server device (1) according to one embodiment can determine the reservation probability of a room to dynamically adjust the room sales price, and separately input the reservation probability into the room sales price adjustment model to adjust the room sales price.

[0069] That is, the control unit (100) can input room data into a room reservation probability determination model to derive a room reservation probability as an output value, and input the derived room reservation probability into a room sales price adjustment model to derive room sales price adjustment data as an output value.

[0070] Thereafter, the control unit (100) can transmit the derived room sales price adjustment data to an external device (2), and the user can check the dynamically adjusted room sales price online.

[0071] The control unit (100) can determine a product demand index and a regional demand index based on the room reservation probability, and input the product demand index and the regional demand index into a room sales price adjustment model to derive room sales price adjustment data as an output value.

[0072] In addition, the control unit (100) can determine the average value of the business demand index and the room demand index including the room forecast reservation rate (sale rate) and the room pickup rate as the product demand index, and the average value can include the arithmetic mean, the geometric mean, and the harmonic mean.

[0073] At this time, the room reservation probability determination model may include a reservation occurrence probability prediction model that predicts the probability of a reservation occurring on a specific stay date for each room and a reservation occurrence count prediction model that predicts the number of reservations occurring on a specific stay date for each business, and the control unit (100) may input room data into the reservation occurrence count prediction model to output the predicted number of reservations for a room, and determine the predicted room reservation rate by dividing the predicted number of reservations for a room by the remaining number of rooms.

[0074] In addition, the control unit (100) can determine the room pickup rate by dividing the number of rooms actually sold from the room data by the total number of rooms, input the room data into a reservation occurrence probability prediction model to output the reservation probability of the room, and determine the reservation probability of the room as a room demand index. That is, the control unit (100) can determine the product demand index based on the business demand index and the room demand index in order to input the product demand index as an input value into a room sales price adjustment model including an AEEF (Asymmetric Exponent-Exponential Form) model.

[0075] In addition, the control unit (100) may modify the room sales price adjustment data based on receiving at least one of the lowest room price line and the price priority between rooms through the communication unit (200), or may modify the room sales price adjustment data based on an amount obtained by excluding the user's coupon usage amount from the room sales price adjustment data.

[0076] Accordingly, the server device (1) can dynamically suggest a room sales price by utilizing a machine learning model, thereby having the effect of maximizing the sales profit of the room seller.

[0077] FIG. 2 illustrates input and output values ​​of a machine learning model utilized in a server device (1) according to one embodiment.

[0078] The control unit (100) can train an artificial neural network using potential demand data, reservation information data, and seasonal data of the room as training data (a).

[0079] When the control unit (100) includes an artificial intelligence processor (110) (e.g., NPU) for training an artificial neural network, the processor (110) can train the artificial neural network by utilizing the weight data stored in the memory (120) as training data for the artificial neural network.

[0080] Examples of learning algorithms include, but are not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning.

[0081] An artificial neural network can be composed of multiple neural network layers. Each of the multiple neural network layers has multiple weight values ​​and performs neural network operations through operations between the calculation results of the previous layer and the multiple weights. The multiple weights of the multiple neural network layers can be optimized based on the learning results of an artificial intelligence model. For example, during the learning process, the multiple weights can be updated so that the loss or cost values ​​obtained from the artificial intelligence model are reduced or minimized.

[0082] The artificial neural network may include a deep neural network (DNN), for example, a convolutional neural network (CNN), a deep neural network (DNN), a recurrent neural network (RNN), a restricted boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), or a deep Q-network (900) (Deep Q-Networks), but is not limited to the examples described above.

[0083] The control unit (100) can learn the correlation between potential demand data of the room, reservation information data, seasonality data, and sales price increase and decrease data based on the selected artificial intelligence model.

[0084] Specifically, the control unit (100) can output room sales price adjustment data (b) including sales price increase data and sales price decrease data based on a classification and regression learning model that uses potential demand data, reservation information data, and seasonality data of the room as input values.

[0085] Figures 3 and 4 illustrate a schematic flowchart of a server device (1) adjusting a room sales price according to one embodiment.

[0086] The configurations 101 to 108 in FIGS. 3 and 4 are implemented in the form of software blocks, and can be stored in memory (120) and executed by the processor (110).

[0087] First, referring to FIG. 3, the control unit (100) can receive room data via the communication unit (200) and store the received room data in the raw data DB (121). At this time, the room data can include basic information, price information, reservation information, room location information, guest feedback information, service information, time-based reservation information, and availability information.

[0088] Specifically, basic information may include room number or name, room type (e.g., standard, suite, deluxe), room size and capacity (e.g., single bed, queen bed, twin bed), price information may include daily or hourly price information, seasonal price change information, discount information, and promotional information, and reservation information may include reservation status information, reservation date information, reservation time information, reservation customer information (e.g., name, contact information, email, etc.), and previous reservation history (preference) information.

[0089] Additionally, room location information may include the name of the accommodation, location information (city, region, etc.), nearby tourist attractions and facility information, and guest feedback information may include previous customer review information, evaluation information, and information on improvements made through feedback and reviews.

[0090] Additionally, service information may include information on in-room facilities (wireless internet, TV, refrigerator, bathroom, etc.), information on accommodation facilities (restaurant, swimming pool, fitness center, etc.), and information on room services and auxiliary facilities, and time-based reservation information and availability information may include information on availability of rooms at a specific time period, and information on the reserved time period and date.

[0091] The control unit (100) may perform feature engineering in the feature engineering execution unit (101) to process the room data received via the communication unit (200) into learning data. Feature engineering may refer to the task of appropriately modifying, selecting, or generating data characteristics or features in machine learning and data analysis to improve model performance and enhance prediction accuracy.

[0092] The control unit (100) can store features generated by feature engineering in a feature data DB (122) and perform feature selection in the feature data DB (122).

[0093] In a server device (1) according to one embodiment, the model-based variable selection (SelectFromModel) algorithm of Scikit-learn can be used to perform feature selection.

[0094] Specifically, the feature selection unit (102) of the control unit (100) can select suitable features to propose dynamic room sales prices. Feature selection is one of the data preprocessing processes used in machine learning and data analysis, and can refer to the task of selecting the most important features (characteristics) for model training and removing unnecessary features.

[0095] The control unit (100) can check the entire feature list extracted from the raw data DB (121) and evaluate feature importance to select suitable features for suggesting dynamic room sales prices. That is, the control unit (100) can evaluate feature importance to determine which features among various features related to room sales price adjustment play the most important role in reservation prediction and sales price adjustment.

[0096] The control unit (100) may use random forest, XGBoost, or other tree-based algorithms to evaluate feature importance.

[0097] Thereafter, the control unit (100) can remove features with a low impact on reservation rate prediction and analyze the correlation between features to identify and remove features with high interdependence. After performing feature selection, the control unit (100) can train and evaluate the model to assess its performance, thereby determining the impact of feature selection on model performance.

[0098] The control unit (100) can perform feature selection and store the selected features in a machine learning metadata DB (123), and the machine learning metadata DB (123) can mean a database that stores and manages metadata related to a machine learning model and project.

[0099] Thereafter, the hyper parameter tuning unit (103) of the control unit (100) can tune the hyper parameters, and the hyper parameters can mean setting values ​​that control the learning process and structure of the machine learning model, and can include, for example, a learning rate, a tree depth, and a regularization strength.

[0100] The control unit (100) can find optimal hyperparameter values ​​to optimize the performance of the model through hyperparameter tuning, prevent overfitting, and enable the model to better generalize to data.

[0101] In a server device (1) according to one embodiment, the GridSearchCV algorithm of Scikit-learn can be used as an algorithm for tuning hyperparameters. GridSearchCV is one of the methods for finding the optimal combination by trying all possible hyperparameter combinations. GridSearchCV uses cross-validation to evaluate the performance of each combination, thereby reliably measuring the generalization performance of a model.

[0102] Likewise, the control unit (100) can store the hyper parameters for which tuning has been performed in the machine learning metadata DB (123).

[0103] Thereafter, the machine learning model learning unit (104) of the control unit (100) can learn the model based on the determined hyperparameters and selected features. Specifically, the control unit (100) can learn the model parameters using the selected features as input data to adjust the room sales price.

[0104] According to one embodiment, a server device (1) can use a random forest model to predict room sales prices, and can be applied to a regression problem of predicting room sales prices using given features. Specifically, the control unit (100) can learn the relationship between given features and room sales prices.

[0105] Here, the random forest model can refer to an ensemble learning model that combines multiple decision trees to perform predictions. Specifically, the random forest model can learn the relationship between features and room sales prices through steps such as bootstrap sampling, random feature selection, ensemble prediction, and hyperparameter tuning.

[0106] Afterwards, the control unit (100) can store the learned machine learning model in the machine learning learning model DB.

[0107] Next, referring to FIG. 4, the prebuilt data generation unit (105) of the control unit (100) can generate prebuilt data for model learning and testing, and the prebuilt data can be used in model learning and evaluation, feature engineering, hyperparameter tuning, data visualization, and exploration. Thereafter, the control unit (100) can store the generated prebuilt data in the prebuilt data DB (125).

[0108] The parameter fitting unit (106) of the control unit (100) can fit parameters to optimize the performance of the learning model and obtain the best prediction results.

[0109] A server device (1) according to one embodiment can utilize an AEEF (Asymmetric Exponent-Exponential Form) model as a room sales price adjustment model, and the control unit (100) can adjust hyperparameters so that the AEEF model can appropriately suggest room sales prices. For example, the control unit (100) can adjust the learning rate of the regression model, the number of trees, the depth, etc.

[0110] Afterwards, the control unit (100) can store the fitted hyper parameters in the machine learning learning model DB (124).

[0111] The API (Application Programming Interface) deployment unit (107) of the control unit (100) can perform a process of deploying (installing and executing) the developed API to a server or hosting environment so that it can be used in an actual operating environment. In other words, the control unit (100) can integrate the developed room price adjustment service with other applications or provide it for use by external developers or users.

[0112] The API providing unit (108) of the control unit (100) can provide a room price adjustment service to the serving layer so that the developed room price adjustment service can interact with the user.

[0113] Accordingly, the user can send a command to the model API to request a room price through the wrapper API in the serving layer, and the model API can send the optimal price to the user's external device (2) through the wrapper API.

[0114] Although each step described above in FIGS. 3 and 4 is depicted sequentially, each step may be performed in parallel by the CPU and GPU.

[0115] FIG. 5 illustrates features utilized in a server device (1) according to one embodiment.

[0116] Referring to FIG. 5, the feature engineering execution unit (101) of the control unit (100) can perform feature engineering based on a feature engineering technique, and specifically, the control unit (100) can use a log transformation and one hot encoding algorithm.

[0117] The log transformation algorithm is used when data does not follow a normal distribution or when the scale of the data is greatly different, and the control unit (100) can generate preprocessing data by transforming the room data to be closer to a normal distribution when the room data has a one-sided distribution.

[0118] The One Hot Encoding algorithm is used to input categorical data from room data into a machine learning model, and the control unit (100) can convert categorical variables into binary vectors consisting of 0 and 1 so that the learning model can understand and utilize the categorical data.

[0119] According to the algorithm described above, the control unit (100) can select features as shown in FIG. 5, and the attributes of the features can be distinguished by variable name (a), time dimension (b), space dimension (c), description (d), and variable type (e).

[0120] Specifically, the variable name (a) can include a date concept representing a category, such as day_of_week for variable 1, week_of_year for variable 2, and month for variable 3.

[0121] Additionally, the time dimension (b) indicates whether each variable comes from features associated with the stay date, features associated with the reservation date, or features associated with both the stay date and the reservation date, which can determine the influence of the stay date or reservation date on the prediction of room sales.

[0122] Additionally, the spatial dimension (c) can determine whether each variable originates from a feature related to a room or a business. Variables 8 (roomprice) through 14 (revenue_weekday_365day) are related to room characteristics, and thus can be determined as variables with a room dimension. Accordingly, the control unit (100) can determine the impact of each feature on the prediction of room sales based on its relevance to a room or business.

[0123] Additionally, the description (d) includes a specific description of each variable, and the variable type (e) determines whether each variable is a categorical variable, a binary variable, or a continuous variable, so that the influence of each feature on the prediction of room sales can be determined according to the variable type derived from each feature.

[0124] According to one embodiment, a server device (1) can determine features in the feature engineering section and feature selection section of the control section (100) and utilize them as shown in FIG. 5.

[0125] FIG. 6 illustrates factors utilized by the control unit (100) of the server device (1) according to one embodiment to adjust the room sales price.

[0126] Referring to FIG. 6, the control unit (100) can determine a business demand index and a room demand index to determine a product demand index used as an input value of a room sales adjustment model, and can determine a room predicted reservation rate (sale rate) and a room pickup rate to determine the business demand index.

[0127] Specifically, the business demand indicator includes a predicted room reservation rate (sale rate) and a room pickup rate, and the control unit (100) can calculate the predicted room reservation rate (sale rate) by dividing the predicted number of reservations (c) by the remaining number of rooms (b). In addition, the control unit (100) can calculate the room pickup rate by dividing the number of actually sold rooms (d) by the total number of rooms (a).

[0128] In addition, the control unit (100) can input room data into a reservation occurrence probability prediction model to output the reservation probability of the room, and determine the reservation probability of the room as a room demand index.

[0129] Thereafter, the control unit (100) can calculate the average value of the determined business demand index and the room demand index as a product demand index, and input the calculated product demand index into the AEEF model to output room sales price adjustment data.

[0130] At this time, the average values ​​of the business demand index and the room demand index may include the arithmetic mean, the geometric mean, and the harmonic mean, and the geometric mean and the harmonic mean may be utilized when a relatively conservative room sales price proposal is required.

[0131] FIG. 7 illustrates the performance of a room sales adjustment model of a server device (1) according to one embodiment.

[0132] Referring to FIG. 7, the control unit (100) can obtain room sales price adjustment data by inputting a product demand index and a regional demand index into a room sales price adjustment model, and the room sales price adjustment model can include an AEEF (Asymmetric Exponent-Exponential Form) model having the product demand index and the regional demand index as variables.

[0133] The control unit (100) can change the price based on the AEEF (Asymmetric Exponent-Exponential Form) model, and can specifically determine the price based on the predicted reservation probability.

[0134] That is, the control unit (100) can increase the room sales price as the predicted reservation probability increases, as in (a) of Fig. 7. The AEEF (Asymmetric Exponent-Exponential Form) model is composed of the following mathematical equations 1 and 2, and can calculate the suggested price (Psug) using the product demand index derived from the reservation probability (q) and the regional demand index (D) as variables.

[0135] [Mathematical Formula 1]

[0136]

[0137] [Equation 2]

[0138]

[0139] For each parameter, determines the minimum and maximum ranges of the price fluctuation ratio, It moves the graph with a price reduction bias and a price increase bias depending on its size, class can determine the reservation probability that serves as the basis for price increase, and specifically As the number of reservations increases, the standard value, which is the basis for price increases, may decrease. As the number of reservations increases, the baseline, which is the probability of a reservation that becomes the basis for a price increase, may increase.

[0140] A server device (1) according to one embodiment can obtain room sales price adjustment data based on an AEEF model, calculate a product demand index and a regional demand index using the aforementioned algorithm, input them into the AEEF model, and suggest an appropriate sales price.

[0141] Fig. 8 illustrates an example in which the sales price of a room is adjusted in a server device (1) according to one embodiment.

[0142] Fig. 8 shows an example of a dataset utilized in the learning model of the server device (1), and the server device (1) can determine a recommended price (e) based on the aforementioned algorithm and suggest it to the user.

[0143] Specifically, the control unit (100) can determine the reservation rate up to the target time point as the predicted reservation rate (c) based on the actual price (d) one day prior to the partition time point (a), which is the reference time point for separating the room data. The control unit (100) can calculate the predicted room reservation rate (c) by dividing the number of predicted reservations by the number of remaining rooms, and the control unit (100) can calculate the actual reservation rate (b) by dividing the number of rooms actually sold by the total number of rooms.

[0144] The control unit (100) can calculate a recommended price (e) based on the calculated predicted reservation rate (c) and actual reservation rate (b), and determine a harmonic average value of a business demand index and a room demand index including the predicted reservation rate (c) and actual reservation rate (b) as a product demand index, so that a room sales price adjustment model can input it.

[0145] In addition, the control unit (100) can raise or lower the room price according to a reference value for raising or lowering the room price, and assuming that the reference value is a predicted reservation rate (c) of 45% as in FIG. 8, when the partition time points (a) are 20230312 and 20230313, the predicted reservation rate (c) is less than 45%, so the actual price is lowered to determine the recommended price, and for the remaining partition time points (a), the predicted reservation rate (c) is 45% or more, so the actual price is raised to determine the recommended price.

[0146] FIG. 9 illustrates a method for verifying a room sales adjustment result in a server device (1) according to one embodiment, and FIG. 10 illustrates a graph comparing a room sales adjustment result in a server device (1) according to one embodiment with an existing model.

[0147] First, referring to FIG. 9, the control unit (100) can verify and evaluate the model by modifying the confusion matrix used in the classification problem to verify the results of adjusting the room sales.

[0148] Specifically, the control unit (100) can compare the model suggested price and the actual selling price, and distinguish between Good / Bad cases based on whether or not the product is sold, and use this for verification.

[0149] In the case of (A) of Fig. 9, if a price increase is proposed for the rooms sold, a higher unit price can be received if the price is raised, so it can be classified as a Good case, and in the case of (B) of Fig. 9, if a price reduction is proposed for the rooms sold, a loss equivalent to the difference can be incurred if the price is lowered, so it can be classified as a Bad case.

[0150] In addition, in the case of (C) of Fig. 9, if a price increase was proposed for unsold rooms, reservations would not have occurred if the price had been raised, so it can be classified as a Bad case, and in the case of (D) of Fig. 9, if a price reduction was proposed for unsold rooms, a loss equivalent to the difference could occur if the price was lowered, so it can be classified as a Bad case.

[0151] Based on the above indicators, the control unit (100) can determine the price decrease recall (PDR), the price decrease precision (PDP), the price decrease precision (PIR), the price increase recall (PIR), and the price increase precision (PIP), as shown in the following mathematical formulas.

[0152] [Equation 3]

[0153] PDR = D / (C + D)

[0154] [Equation 4]

[0155] PDP = D / (B + D)

[0156] [Equation 5]

[0157] PIR = A / (A + B)

[0158] [Equation 6]

[0159] PIP = A / (A + C)

[0160] In addition, the control unit (100) can determine the reservation price error rate BR (Booking Regret) as the median of the error rates between the proposed price and the actual reservation price among cases A and B.

[0161] According to one embodiment, a server device (1) can evaluate a model by applying a high weight to a PDR corresponding to mathematical expression 3 and a PIP corresponding to mathematical expression 6 in the process of verifying a learning model.

[0162] That is, the control unit (100) can evaluate the model by placing a high weight on price reduction recall and price increase precision, thereby reducing Bad Price Increase cases where sales are not made despite a proposed price increase. This minimizes the risk of not making a sale despite a proposed price increase, thereby maximizing user benefits.

[0163] Referring to FIG. 10, the upper graph of FIG. 10 may represent the result of plotting the PDR (Price Decrease Recall), which is the price decrease recall rate, by lead time, and the lower graph may represent the result of plotting the PIP (Price Increase Precision), which is the price increase precision, by lead time.

[0164] First, referring to the upper graph of FIG. 10, (a) may refer to the price reduction reproducibility when a server device (1) according to one embodiment dynamically suggests a room sales price, and (b) may refer to the price reduction reproducibility when a competitive room model, which is a conventional technology, dynamically suggests a room sales price.

[0165] That is, in the case of the price reduction recurrence rate for suggesting a price reduction for unsold rooms, it can be confirmed that the performance of the dynamic sales proposal by the server device (1) according to one embodiment is improved by about two times compared to the existing competitive room model.

[0166] Next, referring to the lower graph of FIG. 10, (c) may refer to the price increase precision when the server device (1) according to one embodiment dynamically suggests the sales price of a room, and (d) may refer to the price increase precision when the competitive room model of the prior art dynamically suggests the sales price of a room.

[0167] That is, in the case of price increase precision, which indicates the ratio of rooms that were actually sold among rooms for which a price increase was proposed, it can be confirmed that the performance of dynamic sales proposals by the server device (1) according to one embodiment is improved by about two times or more compared to existing competitive room models.

[0168] In this way, the server device (1) according to one embodiment can verify and evaluate a learning model that proposes a room sales price by transforming a confusion matrix, so that the overall performance of the learning model can be easily confirmed and improved based on verification data.

[0169] FIG. 11 is a flowchart regarding a control method of a server device (1) according to one embodiment, and FIG. 12 is a drawing that continues the flowchart regarding a control method of a server device (1) according to one embodiment from FIG. 11.

[0170] Referring to FIG. 11, the control unit (100) can receive room data from the communication unit (200) (1100). Thereafter, the control unit (100) can extract features for adjusting room sales prices from the room data and generate preprocessed data (1110). In this case, the preprocessed data may refer to the resulting data after feature engineering and feature selection are performed on the room data.

[0171] The control unit (100) can learn and generate a room reservation probability determination model and a room sales price adjustment model using preprocessed data as learning data (1120). The control unit (100) can input preprocessed data into the room reservation probability determination model and derive a room reservation probability as an output value (1130).

[0172] Thereafter, the control unit (100) can input the room reservation probability into the room sales price adjustment model to derive the room sales price adjustment data as an output value (1140), and specifically, determine the product demand index based on the room reservation probability, and input the product demand index and the regional demand index into the room sales price adjustment model.

[0173] Referring to FIG. 12, the control unit (100) can determine whether at least one of the lowest room price line and the price priority between rooms has been received through the communication unit (200) (1200).

[0174] Specifically, a user may input a minimum room price line into the user terminal, and the minimum room price line may mean a minimum price at which the room sales price can be adjusted based on the user's operating profit.

[0175] When the lowest room price line is received (e.g., 1200), the control unit (100) can modify the room sales price adjustment data (1220) by considering the lowest room price line entered by the user. For example, if the room price adjusted by the room sales price adjustment data is 80,000 won, and the entered lowest room price line is 81,000 won, the room price adjusted by the room sales price adjustment data can be modified to 81,000 won.

[0176] Additionally, the control unit (100) may receive a price priority between rooms from the user, and the price priority between rooms may refer to a priority for which a plurality of rooms meeting similar conditions are desired to be sold first. The control unit (100) may modify the room sales price adjustment data based on the price priority between rooms input by the user. For example, among two rooms offered with the same room sales price, the control unit (100) may modify the room sales price adjustment data to lower the sales price of the room with the higher sales priority compared to the remaining rooms.

[0177] In addition, the control unit (100) can determine whether an amount excluding the user's coupon usage amount has occurred in the room sales adjustment data due to coupon usage (1210).

[0178] If the control unit (100) determines that the amount of the room sales price adjustment has been deducted from the user's coupon usage amount due to coupon use (example 1210), the control unit (100) may modify the room sales price adjustment data based on the coupon usage amount (1220). For example, if the maximum coupon amount for a room has been exhausted, the control unit (100) may modify the room sales price adjustment data to lower the room sales price to maintain the coupon-applied sales price.

[0179] Thereafter, the control unit (100) can transmit room sales price adjustment data to an external device (2) including a user terminal and an external server device (1) (1230), and the user can maximize sales profits by reflecting the room sales price that maximizes sales profits.

[0180] In this way, according to the server device (1) according to one embodiment, the room sales price can be dynamically adjusted to maximize sales profits by considering changes in various commercial demands, thereby reducing the effort and cost for adjusting the room sales price and having the effect of maximizing operating profits.

[0181] Meanwhile, the disclosed embodiments may be implemented in the form of a recording medium storing computer-executable instructions. The instructions may be stored in the form of program code, and when executed by a processor, may generate program modules to perform the operations of the disclosed embodiments. The recording medium may be implemented as a computer-readable recording medium.

[0182] Computer-readable storage media include all types of storage media that store instructions that can be deciphered by a computer. Examples include read-only memory (ROM), random access memory (RAM), magnetic tape, magnetic disks, flash memory, and optical data storage devices.

[0183] Additionally, a computer-readable storage medium may be provided in the form of a non-transitory storage medium. Here, the term "non-transitory storage medium" simply means a tangible device that does not contain signals (e.g., electromagnetic waves). This term does not distinguish between cases where data is permanently stored in the storage medium and cases where data is temporarily stored. For example, a "non-transitory storage medium" may include a buffer in which data is temporarily stored.

[0184] According to one embodiment, the method according to the various embodiments disclosed in the present document may be provided as included in a computer program product. The computer program product may be traded as a product between a seller and a buyer. The computer program product may be distributed in the form of a machine-readable recording medium (e.g., compact disc read only memory (CD-ROM)), or may be distributed online (e.g., downloaded or uploaded) via an application store (e.g., Play Store™) or directly between two user devices (e.g., smartphones). In the case of online distribution, at least a portion of the computer program product (e.g., a downloadable app) may be at least temporarily stored or temporarily generated in a machine-readable recording medium, such as a memory of a manufacturer's server, an application store's server, or an intermediary server.

[0185] The above has illustrated and described specific embodiments. However, the invention is not limited to the above-described embodiments, and those skilled in the art will appreciate that various modifications and implementations can be made without departing from the spirit and scope of the invention as set forth in the claims below.

Claims

1. A communication unit that receives room data for training a machine learning model related to room sales price adjustment; A memory for storing the above room data; and From the above room data, features for adjusting room sales are extracted to create preprocessed data, A room reservation probability determination model and a room sales price adjustment model are created using the above preprocessed data as learning data, A server device including a control unit that inputs the room data into the room reservation probability determination model to derive a room reservation probability as an output value, and inputs the derived room reservation probability into the room sales price adjustment model to derive room sales price adjustment data as an output value.

2. In claim 1, The above control unit A server device that determines a product demand index and a regional demand index based on the above-mentioned room reservation probability, inputs the product demand index and the regional demand index into the above-mentioned room sales price adjustment model, and derives the above-mentioned room sales price adjustment data as an output value.

3. In claim 2, The above control unit, A server device that determines the average value of the business demand index and the room demand index, including the room forecast reservation rate (sale rate) and the room pickup rate (pickup rate), as the product demand index.

4. In claim 3, The above room reservation probability determination model is, A server device including a reservation occurrence probability prediction model that predicts the probability of a reservation occurring on a specific stay date for each room and a reservation occurrence count prediction model that predicts the number of reservations occurring on a specific stay date for each business.

5. In claim 4, The above control unit, A server device that inputs the above room data into the above reservation occurrence number prediction model to output the predicted number of reservations for the room, and determines the predicted reservation rate for the room by dividing the predicted number of reservations for the room by the remaining number of rooms.

6. In claim 4, The above control unit, A server device that determines the room pickup rate by dividing the number of rooms actually sold by the total number of rooms in the above room data.

7. In claim 4, The above control unit, A server device that inputs the above room data into the above reservation occurrence probability prediction model to output the reservation probability of the room, and determines the reservation probability of the room as the room demand index.

8. In claim 2, The above room sales price adjustment model is, A server device including an AEEF (Asymmetric Exponent-Exponential Form) model having the above product demand index and the above regional demand index as variables.

9. In claim 1, The above control unit, A server device that modifies the room sales price adjustment data based on receiving at least one of the lowest room price line and the price priority between rooms through the communication unit.

10. In claim 1, The above control unit, A server device that modifies the room sales price adjustment data based on the amount excluding the user's coupon usage amount from the room sales price adjustment data.

11. In claim 1, The above control unit, A server device that transmits the above room sales adjustment data to an external device.

12. Receive room data for training a machine learning model related to room sales price adjustment; Extract features for adjusting room sales from the above room data to create preprocessed data; A room reservation probability determination model and a room sales price adjustment model are created using the above preprocessed data as learning data; By inputting the above room data into the above room reservation probability determination model, the room reservation probability is derived as an output value; A control method of a server device, comprising: inputting the above-mentioned room reservation probability into the above-mentioned room sales price adjustment model, and deriving room sales price adjustment data as an output value.

13. In claim 12, The above room sales adjustment data is derived as output values. A control method of a server device that determines a product demand index and a regional demand index based on the above-mentioned room reservation probability, inputs the product demand index and the regional demand index into the above-mentioned room sales price adjustment model, and derives the above-mentioned room sales price adjustment data as an output value.

14. In claim 13, Determining the above product demand indicators is: A control method for a server device that determines the average value of the business demand index and the room demand index including the room forecast reservation rate (sale rate) and the room pickup rate as the product demand index.

15. In claim 14, The above room reservation probability determination model is, A control method for a server device, comprising a reservation occurrence probability prediction model that predicts the probability of a reservation occurring on a specific stay date for each room and a reservation occurrence count prediction model that predicts the number of reservations occurring on a specific stay date for each business.

16. In claim 15, Predicting the above room reservation rate is: A control method of a server device that inputs the above-mentioned room data into the above-mentioned reservation occurrence number prediction model to output the predicted number of reservations for the room, and determines the predicted reservation rate for the room by dividing the predicted number of reservations for the room by the remaining number of rooms.

17. In claim 15, What determines the above room pick-up rate is: A control method for a server device that determines the room pickup rate by dividing the number of rooms actually sold by the total number of rooms in the above room data.

18. In claim 15, Determining the above room demand indicators is: A control method of a server device that inputs the above room data into the above reservation occurrence probability prediction model to output the reservation probability of the room, and determines the reservation probability of the room as the room demand index.

19. In claim 13, The above room sales price adjustment model is, A control method for a server device including an AEEF (Asymmetric Exponent-Exponential Form) model having the above product demand index and the above regional demand index as variables.

20. In claim 12, A control method of a server device further comprising: modifying the room sales adjustment data based on receiving at least one of a room lowest price line and a room-to-room price priority through a communication unit; 21. In claim 12, A control method of a server device further comprising: reducing the room sales price adjustment data based on an amount excluding the user's coupon usage amount from the room sales price adjustment data; 22. In claim 12, A control method of a server device further comprising transmitting the above-mentioned room sales adjustment data to an external device.

Images