Fraudulent booking detection system, fraudulent booking detection method, and program

Abstract

To provide a fraudulent reservation detection system and the like for presuming reservation of an accommodation facility by a reserving person for fraudulent purpose.SOLUTION: In a fraudulent reservation detection system 10, a fraudulence presuming apparatus has an acquiring unit for acquiring current reservation information of an accommodation facility, and a learning unit for carrying out learning by using, as teacher data, information of a reserving person included in past reservation information at a plurality of accommodation facilities and regarding any one of a reserving person of the accommodation facility or any one of an attribute of a visitor, the number of visitors, a scheduled check-in time registered by the reserving person, and a name of the reserving person, any of reservation status information of the same scheduled stay date or the same accommodation facility by the reserving person at a predetermined past point and per-period reservation number information relating to the number of reservation of the accommodation facility or the number of stays by the reserving person during a predetermined past period, and information indicating whether or not each of the plurality of accommodation facilities was used for fraudulent purpose, to output, in response to input of the current reservation information, information indicating a degree of possibility that the reserving person reserved a reservation target accommodation facility for the purpose of fraudulent act.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to an illegal reservation detection system, an illegal reservation detection method, and a program.

Background Art

[0002] An information processing apparatus is disclosed that estimates the risk of reservation cancellation based on reservation information regarding the reservation of a facility by a reservation user (for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Since the information processing apparatus described in Patent Document 1 can estimate the risk of reservation cancellation, for example, appropriate measures (such as invalidating the reservation) can be taken for the reservation by the time of the scheduled accommodation date.

[0005] However, in the travel industry in recent years, not only reservation cancellations but also damages caused by users called "UG (Undesirable Guest)" that cause various damages to accommodation facilities have been increasing. And one of the damages suffered by accommodation facilities is the use of accommodation facilities for the purpose of illegal acts including prostitution. Therefore, in order to prevent the use of accommodation facilities for the purpose of illegal acts, it is desired to realize a new mechanism for estimating whether a reservation made by a user is for the purpose of illegal acts.

[0006] Therefore, an object of the present invention is to provide an illegal reservation detection system capable of estimating whether a reservation of an accommodation facility is made by a user for the purpose of illegal acts.

Means for Solving the Problems

[0007] A fraudulent reservation detection system according to one aspect of the present invention includes: an acquisition unit that acquires reservation information, which is information relating to a current reservation of an accommodation facility by a predetermined person; a learning unit that learns as training data at least one of the following information included in past reservation information of multiple accommodation facilities: reservation information, which is information relating to at least one of the attributes of the person making the reservation or the person planning to stay at the accommodation facility, the number of persons planning to stay, the planned check-in time registered by the person, and the name of the person making the reservation; reservation status information, which is information relating to the reservation status of the same planned stay date or the same accommodation facility by the person making the reservation at a predetermined point in the past; and period reservation number information, which is information relating to the number of reservations or number of nights of stay by the person making the reservation at the accommodation facility during a predetermined period in the past; and information indicating whether each of the multiple accommodation facilities was used for fraudulent purposes; and outputs information indicating the degree to which the person making the reservation has made a reservation at the accommodation facility in question for fraudulent purposes, by inputting the current reservation information.

[0008] A method for detecting fraudulent reservations according to one aspect of the present invention involves a computer acquiring reservation information, which is information relating to a current reservation of an accommodation facility by a predetermined person; learning as training data at least one of the following from past reservation information of multiple accommodation facilities: reservation information, which is information relating to at least one of the attributes of the person making the reservation or the person planning to stay at the accommodation facility, the number of people planning to stay, the planned check-in time registered by the person, and the name of the person making the reservation; reservation status information, which is information relating to the reservation status of the same planned stay date or the same accommodation facility by the person making the reservation at a predetermined point in the past; and period reservation count information, which is information relating to the number of reservations or days of stay by the person making the reservation at the accommodation facility during a predetermined period in the past; and information indicating whether each of the multiple accommodation facilities was used for fraudulent purposes; and inputting the current reservation information to output information indicating the degree to which the person making the reservation has made a reservation at the accommodation facility in question for fraudulent purposes.

[0009] A program according to one aspect of the present invention causes a computer to acquire reservation information, which is information relating to a current reservation of an accommodation facility by a predetermined person; to learn as training data at least one of the following information included in past reservation information of multiple accommodation facilities: the attributes of the person who made the reservation or is scheduled to stay at the accommodation facility, the number of people scheduled to stay, the scheduled check-in time registered by the person, and the name of the person who made the reservation; the reservation status information, which is information relating to the reservation status of the same scheduled stay date or the same accommodation facility by the person at a predetermined point in the past; and the number of reservations or days of stay information, which is information relating to the number of reservations or days of stay by the person at the accommodation facility during a predetermined period in the past; and information indicating whether each of the multiple accommodation facilities was used for fraudulent purposes; and to output information indicating the degree to which the person made the reservation of the accommodation facility for fraudulent purposes by inputting the current reservation information. [Effects of the Invention]

[0010] According to the present invention, it is possible to provide a fraudulent reservation detection system that can estimate whether or not a reservation for accommodation was made by a person with the intention of committing fraud. [Brief explanation of the drawing]

[0011] [Figure 1] This figure shows an example of the configuration of a fraudulent reservation detection system. [Figure 2] This is a configuration diagram showing an example of the functional configuration of the various devices that make up the fraudulent reservation detection system. [Figure 3] This is a diagram showing reservation information D111. [Figure 4] This is a diagram showing reservation status information D112. [Figure 5] This is a diagram showing the reservation count information for the period, D113. [Figure 6] This is a diagram showing information D114 regarding the intent of fraudulent activity. [Figure 7] This is a flowchart showing the processing procedure for the fraudulent reservation detection system 10. [Figure 8] It is a diagram showing an example of the hardware configuration of a computer.

Embodiments for Carrying Out the Invention

[0012] ===Unauthorized Reservation Detection System 10=== <<Overview>> Referring to FIG. 1, the configuration of the unauthorized reservation detection system 10 according to the present embodiment will be described. FIG. 1 is a diagram showing an example of the configuration of the unauthorized reservation detection system 10.

[0013] As shown in FIG. 1, the unauthorized reservation detection system 10 includes, for example, an unauthorized estimation device 100 and a facility reservation management device 200. Although not shown, the unauthorized reservation detection system 10 may be configured by a plurality of facility reservation management devices 200.

[0014] Each device of the unauthorized reservation detection system 10 is communicably connected to each other by a communication network N. The communication network N may be either a wired network or a wireless network.

[0015] The unauthorized reservation detection system 10 is, for example, a system that estimates whether a reservation made by a reservation applicant for an accommodation facility as a reservation target is for an illegal purpose using an evaluation model learned based on accumulated past reservation information.

[0016] Specifically, the unauthorized reservation detection system 10 acquires, for example, information regarding the reservation of an accommodation facility (hereinafter referred to as "reservation information") from at least one facility reservation management device 200. The unauthorized reservation detection system 10 inputs the acquired reservation information into an evaluation model and outputs information indicating the degree (for example, probability) of the possibility that the accommodation reservation is a reservation for an illegal purpose. The unauthorized reservation detection system 10 provides a score (hereinafter referred to as "UG score") calculated based on the output information to the facility reservation management device 200.

[0017] Reservation information is, for example, information regarding the reservation of a person who uses an accommodation facility. The reservation information is transmitted to the facility reservation management device 200, for example, based on a user's operation input. The reservation information is information indicating, for example, a reservation number, a reservation date and time, the name and number of people of the person who reserves the accommodation facility, and the like. Specifically, the reservation information includes, for example, information stored in each item of the reservation person information D111 described later.

[0018] The purpose of illegal behavior is, for example, the purpose of an act that violates the usage regulations of an accommodation facility (including illegal acts), for example, the purpose of prostitution. Hereinafter, as an example, the purpose of illegal behavior will be described as the "purpose of prostitution".

[0019] The evaluation model is, for example, a regression model or a classification model. Specifically, the evaluation model may be a method using, for example, linear regression, logistic regression, random forest, boosting, support vector machine, or neural network. The evaluation model will be described later.

[0020] The illegal estimation device 100 is, for example, a server device that calculates a UG score for reservation information acquired from at least one facility reservation management device 200. The illegal estimation device 100 provides the calculated UG score to the facility reservation management device 200.

[0021] The facility reservation management device 200 is, for example, a server device that receives a reservation for an accommodation facility from a user. The facility reservation management device 200 is, for example, a device operated by an operator who operates a travel agency that receives reservations for accommodation facilities. When the facility reservation management device 200 acquires reservation information from a person making a reservation, for example, it transmits the information to the illegal estimation device 100. Then, it acquires the UG score for each piece of reservation information output by the illegal estimation device 100.

[0022] Thereby, for example, an operator who operates a travel agency can determine whether to take measures such as issuing a warning to a person making a reservation who may have made a reservation for an accommodation facility for the purpose of illegal behavior, based on the UG score received by the facility reservation management device 200.

[0023] The fraud detection device 100 and the facility reservation management device 200 may be, but are not limited to, a server computer, a personal computer (but are not limited to, a desktop, laptop, tablet, etc.), a media computer platform (e.g., cable, satellite set-top box, digital video recorder), a handheld computer device (e.g., PDA (Personal Digital Assistant), email client, etc.), or other types of computers or communication platforms.

[0024] <<Fraudulent Estimation Device 100>> The functional configuration of the fraud detection device 100 will be explained with reference to Figure 2. Figure 2 is a configuration diagram showing an example of the functional configuration of the various devices that make up the fraud reservation detection system 10.

[0025] The fraud detection device 100 is a device that obtains the reservation information of the person making the reservation from the facility reservation management device 200 and outputs a UG score based on the obtained reservation information.

[0026] As shown in Figure 2, the fraudulent estimation device 100 includes, as functional units, a storage unit 110, an acquisition unit 120, a learning unit 130, an estimation unit 140, and a transmission unit 150.

[0027] The memory unit 110 may include, for example, reservation information D111, reservation status information D112, number of reservations for a period information D113, and fraudulent activity purpose information D114.

[0028] The reservation information D111, reservation status information D112, number of reservations for a period information D113, and fraudulent activity purpose information D114 (hereinafter collectively referred to as the "reservation database") are, for example, databases for efficiently training the evaluation model of the learning unit 130, which will be described later, and databases for inputting the information of current reservations into the evaluation model. Specifically, the reservation database is a database compiled by users, who extract, edit, and combine the data that constitutes each of the reservation databases from a database (not shown) that stores current and past reservation information.

[0029] Furthermore, the reservation database does not necessarily have to be stored in the memory unit 110, for example. In this case, the fraud estimation device 100 only needs to generate information corresponding to the items included in the reservation database when training the evaluation model of the learning unit 130 based on a database (not shown) that stores current and past reservation information.

[0030] Please refer to Figure 3 to explain the reservation information D111. Figure 3 is a diagram showing the reservation information D111.

[0031] Reservation information D111 is a database compiled by extracting and editing information such as the attributes of the reservation holder and information about the accommodation facility they booked, from current and past reservation information.

[0032] As shown in Figure 3, the reservation information D111 includes items such as [Reservation ID], [Reservationee ID], [Reservationee Name], [Name of Guest], [Number of Guests], [Gender], [Scheduled Check-in Time], [Maximum Occupancy of Reserved Room], [Payment Method], [Age], [Number of Rooms], [Accommodation Type], and [Difference in Name].

[0033] The [Reservation ID] field stores, for example, an identification code that can identify each individual reservation.

[0034] The [Reservation ID] field stores, for example, an identification code that can identify the reservation holder.

[0035] [Reservation Name] will store, for example, the name of the person making the reservation.

[0036] [Name of prospective guest] will store, for example, the name of the prospective guest.

[0037] The [Number of people expected to stay] field stores information such as the number of people who are expected to use the accommodation.

[0038] The [Gender] field stores information indicating, for example, the gender of the person planning to stay at the accommodation.

[0039] The [Scheduled Check-in Time] field stores information such as the scheduled check-in time for guests staying at the accommodation.

[0040] The [Maximum Capacity of Reserved Room] field stores information such as the maximum capacity of the room being reserved at an accommodation facility.

[0041] The [Payment Method] field stores information indicating the payment method for, for example, accommodation fees. Specifically, the payment methods for accommodation users include "prepayment," where the reservation holder pays in advance using a credit card, and "on-site payment," where the reservation holder or guest pays at the accommodation.

[0042] The [Age] field stores information indicating the age of the person planning to stay at the accommodation. Alternatively, the [Age] field may store information indicating the age group of the person planning to stay. Specifically, this age group information might be expressed in 5-year or 10-year increments.

[0043] The [Number of Rooms] field stores information such as the total number of rooms in the accommodation facility being booked.

[0044] The [Accommodation Category] field stores information indicating the type of accommodation (e.g., business hotel, inn, etc.).

[0045] The [Name Difference] field stores information indicating whether, for example, the name of the person making the reservation and the name of the person scheduled to stay are different. Specifically, for example, if in past reservation information the name of the person making the reservation is "A" and the name of the person scheduled to stay is "B", then there is a name difference, and "1" is entered in [Name Difference]. On the other hand, for example, if the name of the person making the reservation is "C" and the name of the person scheduled to stay is "C", then there is no name difference, and "0" is entered in [Name Difference].

[0046] Next, we will explain the reservation status information D112 with reference to Figure 4. Figure 4 is a diagram showing the reservation status information D112.

[0047] Reservation status information D112 is a database containing information about the reservation status of the same planned stay date or the same accommodation, compiled based on information included in current and past reservation information, for example.

[0048] As shown in Figure 4, the reservation status information D112 may include items such as [Reservation ID], [Number of reservations for the same stay date], [Number of facility reservations for the same stay date], [Number of consecutive nights at the same facility], and [Vendor usage flag].

[0049] The [Reservation ID] field stores, for example, an identification code that can identify the reservation holder.

[0050] The [Number of bookings for the same stay date] field stores, for example, the number of bookings made by a user for the planned stay date.

[0051] Specifically, for example, if guest P has booked "Room a at Accommodation A" on July 5, 2022, and has also booked "Room b at Accommodation A", "Room c at Accommodation B", "Room d at Accommodation D", and "Room e at Accommodation D" on the same day, then "5" will be stored in [Number of bookings for the same stay date].

[0052] The [Number of facility reservations per stay date] column stores, for example, the number of reservations for facilities other than the designated facility booked by the user on the same stay date.

[0053] Specifically, for example, if guest P has booked "Room a at Accommodation A" on July 5, 2022, and has also booked "Room c at Accommodation B" and "Room d at Accommodation D", then "2" will be stored in [Number of facility bookings for the same stay date].

[0054] [Number of consecutive nights at the same facility] stores information about the number of consecutive nights, for example, if you have booked the same accommodation facility consecutively.

[0055] Specifically, for example, if user B makes a reservation for "Room a at Accommodation A" on July 12, 2022, and had already made a reservation for "Room a at Accommodation A" on the previous day, July 11, 2022, then "2" will be stored in [Number of consecutive nights at the same facility].

[0056] The [Vendor Usage Flag] stores information, for example, from past reservation data, indicating whether or not the reservation was made through a vendor that makes reservations for fraudulent purposes.

[0057] Specifically, for example, if past reservation information matches the conditions for being a fraudulent booking broker, "1" will be stored in the [Broker Use Flag]. Conditions for being a fraudulent booking broker include, for example, the reservation holder's name is not a personal name, and the number of reservations for the same accommodation date and the number of reservations for the same accommodation date at different facilities exceed a predetermined number of reservations (for example, "2" for the same accommodation date and "4" for the same accommodation date at different facilities). Alternatively, conditions for being a fraudulent booking broker may also include, for example, the reservation holder's name is not a personal name, and the number of consecutive nights stayed at the same facility at different facilities exceeding a predetermined number of nights (for example, 30 days).

[0058] Next, we will explain the period reservation information D113 with reference to Figure 5. Figure 5 is a diagram showing the period reservation information D113.

[0059] The period reservation count information D113 is a database that includes information about the number of accommodation reservations or the number of nights booked by a reservation holder during a predetermined period in the past relative to a predetermined point in time (the present or a past point in time), compiled based on reservation information. As shown in Figure 5, the period reservation count information D113 may include items such as [Reservation Holder ID], [Aggregation Period], [Maximum Number of Reservations for Same Night Stay], [Maximum Number of Facility Reservations per Same Night Stay], [Total Number of Nights Stayed at Same Facility], and [Total Number of Facilities Used].

[0060] The [Reservation ID] field stores, for example, an identification code that can identify the reservation holder.

[0061] The [Aggregation Period] field stores information indicating a specific period in the past.

[0062] [Maximum number of reservations for the same stay date] stores, for example, the maximum number of reservations for the same scheduled stay date within a specified period in the past.

[0063] Specifically, for example, if, in the 30 days prior to the planned stay date of June 1st as indicated in the reservation information, reservation holder P made reservations for "Room a at Accommodation A" and "Room b at Accommodation B" on May 10th (reservation count "2"), and further reservations for "Room c at Accommodation B", "Room d at Accommodation C", and "Room e at Accommodation D" on May 20th (reservation count "3"), then "3" will be stored in [Maximum number of reservations for the same stay date].

[0064] [Maximum number of facility reservations per same stay date] stores, for example, the maximum number of reservations for the same planned stay date at a different facility from the one booked, over a specified period in the past.

[0065] Specifically, for example, if, in the 30 days prior to the planned stay date of June 4th as indicated in the reservation information, reservation holder P made a reservation for "Accommodation A" on May 10th, and also made reservations for "Accommodation B" and "Accommodation C," which are different accommodations from "Accommodation A," then "2" will be stored in [Maximum number of accommodation reservations per same stay date].

[0066] [Total number of nights stayed at the same facility] stores, for example, the total number of days spent at the same accommodation facility during a specified period in the past.

[0067] Specifically, for example, if, within the 30 days prior to the scheduled stay date of June 4th as indicated in the reservation information, the reservation holder P used "Accommodation Facility A" on May 10th, May 13th, and May 16th, then "3" will be stored in [Total number of nights stayed at the same facility].

[0068] The [Total Number of Facilities Used] column stores, for example, the total number of accommodations used during a specified period in the past.

[0069] Specifically, for example, if the person who made the reservation, P, used "Accommodation A," "Accommodation B," and "Accommodation C" during the 30 days prior to the scheduled stay date of June 4th as indicated in the reservation information, then "3" will be stored in [Total number of facilities used].

[0070] Next, we will explain the fraudulent activity intent information D114 with reference to Figure 6. Figure 6 is a diagram showing the fraudulent activity intent information D114.

[0071] The fraudulent activity purpose information D114 is a database that compiles information indicating, for example, whether a past reservation for an accommodation facility was made by a user with the intention of engaging in fraudulent activity. As shown in Figure 6, the fraudulent activity purpose information D114 may include items such as [Reservation ID], [Reservation User ID], and [Fraudulent Activity Flag].

[0072] The [Reservation ID] field stores, for example, an identification code that can identify each individual reservation.

[0073] The [Reservation ID] field stores, for example, an identification code that can identify the reservation holder.

[0074] The [Fraudulent Activity Flag] stores information indicating, for example, whether a past reservation was for the purpose of fraudulent activity (in this case, prostitution).

[0075] Here, the fraudulent activity flag is obtained, for example, from the operators of travel agencies that accept accommodation reservations, to determine whether or not past reservation information was actually made with the intention of fraudulent activity. A "1" is stored for reservation information that was made with the intention of fraudulent activity, and a "0" is stored for reservation information that was not made with the intention of fraudulent activity.

[0076] The acquisition unit 120 acquires current and past reservation information from, for example, the facility reservation management device 200. The current and past reservation information is stored in the storage unit 110.

[0077] The learning unit 130 has, for example, an evaluation model that performs machine learning using the information stored in the memory unit 110. The evaluation model evaluates, for example, the likelihood that a user's reservation for accommodation is for fraudulent purposes.

[0078] Here, the evaluation model is a model that has been trained using, for example, information contained in past reservation information for multiple accommodations (in this case, information contained in a reservation database) and information indicating whether or not each of those accommodations was used for fraudulent purposes (in this case, a fraud flag) as training data. When the evaluation model receives the current reservation information of a person making a reservation, it outputs information indicating the degree to which that person is likely to have made a reservation for the target accommodation for fraudulent purposes.

[0079] Specifically, to give one example, the learning unit 130 can obtain a regression model using the "fraudulent activity flag" of the fraudulent activity purpose information D114, which indicates whether or not a hotel reservation for an accommodation facility was a reservation made for the purpose of fraudulent activity, as the dependent variable (ground truth data), and at least one piece of information extracted from at least one of the reservation database (reservation attribute information D111, reservation status information D112, and number of reservations for a period information D113) as the independent variable (feature).

[0080] For example, the learning unit 130 obtains a regression model as shown in equation (1) below as the evaluation model. y = W1·X1 + W2·X2 + ... + W n ·X n ...(1)

[0081] In equation (1), "X" is an explanatory variable corresponding to the information extracted from the reservation database. "y" is the dependent variable corresponding to the "fraudulent activity flag," which is "1" if the reservation was for fraudulent purposes, and "-1" or "0" if the reservation was not for fraudulent purposes.

[0082] Furthermore, in equation (1), "W" is the coefficient of "X" and represents the weight value. Specifically, "W1" is the weight value of "X1", "W2" is the weight value of "X2", and "W n " is "X n This is the weight value of ". Thus, equation (1) is created by combining an explanatory variable "X" which corresponds to information extracted from at least one of the reservation attribute information D111, reservation status information D112, and period reservation number information D113, with a variable (for example, "W1·X1") which includes the weight value "W".

[0083] First, we will explain the rationale for using the information contained in the reservation database (reservation attribute information D111, reservation status information D112, and number of reservations for a given period information D113) as training data to train the evaluation model, which allows for highly accurate estimation of whether or not a reservation is made for fraudulent purposes.

[0084] In other words, the fraudulent reservation detection system 10 can solve the problem that simply training an evaluation model with general reservation information (e.g., name, age, accommodation, number of nights, etc.) was not sufficient to adequately estimate whether a reservation was made for fraudulent purposes.

[0085] The following describes the results of a comparison of trends between bookings made for legitimate use and bookings made for fraudulent purposes, based on a vast amount of past accommodation reservations. Specifically, the following describes the prominent trends shown in the information regarding bookings made for fraudulent purposes. It can be seen that by training an evaluation model with information showing such trends as training data, it is possible to generate an evaluation model that can make accurate estimations.

[0086] Regarding the "gender" of those making reservations, while the ratio tends to be 2 men to 1 woman for normal use, it was found that reservations made by women overwhelmingly outnumbered those made by those engaging in fraudulent activity.

[0087] Regarding the "number of people planning to stay" for reservations, it was found that while regular use tends to show a tendency towards 1-2 people, fraudulent use overwhelmingly shows a tendency towards reservations for 1 person.

[0088] Regarding the "planned check-in time" of reservations, it was found that for normal use, 3:00 PM was the most common time, and this gradually decreased towards midnight as time progressed. In contrast, reservations made for fraudulent purposes tended to have a concentrated check-in time around 3:00 PM.

[0089] Regarding the "maximum occupancy of reserved rooms," while reservations for rooms accommodating 3 or more people are common for normal use, it was found that reservations for 1-2 people overwhelmingly tend to be made for fraudulent purposes.

[0090] Regarding the "payment methods" of those making reservations, while normal use tends to show a ratio of 1:2 (on-site payment to advance payment), it was found that in cases of fraudulent use, almost all reservations were paid for on-site.

[0091] Regarding the discrepancy between the name of the person making the reservation and the name of the person scheduled to stay (name discrepancy), it was found that in normal use, the names tend to match in most reservations, while in cases of fraudulent use, the names tend to be mismatched in almost all reservations.

[0092] Regarding the "accommodation type" of booked accommodations, it was found that while bookings for regular use tend to include inns and resorts, bookings for fraudulent use tend to consist almost entirely of business hotels and city hotels.

[0093] Regarding the number of "accommodation rooms" booked, it was found that while bookings for normal use tend to be for small to medium-sized facilities, bookings for fraudulent purposes tend to be for very few small facilities.

[0094] Regarding the "age range" of those making reservations, it was found that compared to regular users, those making reservations for fraudulent purposes tend to be younger and concentrated around the age of 30.

[0095] Regarding the number of reservations made for the same day, the number of reservations made for the same facility on the same day, and the number of consecutive nights at the same facility, it was found that those making reservations for fraudulent purposes tend to make multiple reservations on the same day compared to those making reservations for regular use.

[0096] Regarding the fact that the person making the reservation is a "business," it was found that if the person meets the criteria for being a broker of fraudulent activities compared to a regular user, they are more likely to be a highly malicious person making a reservation with the intent of fraudulent activity, and they tend to make further reservations for fraudulent purposes.

[0097] Regarding the "maximum number of reservations for the same stay" and the "maximum number of reservations for different facilities for the same stay," it was found that those making reservations for fraudulent purposes tend to make multiple reservations on the same day compared to those making reservations for regular use.

[0098] Regarding the "total number of nights stayed at the same accommodation facility" for each reservation, it was found that those who made reservations using fraudulent means tended to stay at the same facility less frequently compared to those who made reservations using the facility normally.

[0099] Regarding the "total number of facilities used" by those who made reservations, it was found that those who made reservations for fraudulent use tended to use the same facility less frequently compared to those who made reservations for regular use.

[0100] Next, we will explain a specific combination of information from the reservation database that can more accurately estimate whether or not a reservation is made for fraudulent purposes.

[0101] Specifically, the learning unit 130 preferably includes, for example, a combination of information from the reservation information indicating the "gender" of the person scheduled to stay, the "number of people scheduled to stay," and the "scheduled check-in time" of the person scheduled to stay, as explanatory variables.

[0102] This is because reservation information intended for fraudulent purposes often exhibits characteristics such as "the prospective guest is female, the number of guests is 1, and the planned check-in time is around 3:00 PM."

[0103] This allows the fraudulent booking detection system 10 to accurately assess whether or not booking information for accommodation facilities is intended for fraudulent purposes.

[0104] Furthermore, the learning unit 130 may, for example, add information indicating "name differences" among the reservation information, specifically whether or not there is a difference between the name of the person making the reservation and the name of the person scheduled to stay, as an explanatory variable to the above combination.

[0105] This is because reservation information intended for fraudulent purposes often has a characteristic where the name of the person making the reservation and the name of the person staying at the hotel are different.

[0106] As a result, the fraudulent booking detection system 10 can accurately evaluate whether or not booking information for accommodation facilities is intended for fraudulent activity by adding information indicating "differences in names" as an explanatory variable.

[0107] The estimation unit 140 estimates the UG score for the current reservation information obtained by the acquisition unit 120. For example, the estimation unit 140 extracts past reservation information of reservation holder P associated with the reservation holder ID included in the current reservation information obtained by the acquisition unit 120 from the storage unit 110.

[0108] The estimation unit 140 can then calculate the UG score Sc using the evaluation model of equation (1), which uses multiple pieces of information extracted from the extracted past reservation information and the current reservation information acquired by the acquisition unit 120 as explanatory variables.

[0109] The estimation unit 140 can calculate the UG score Sc using formula (1). The fraud estimation device 100 transmits the result calculated by the estimation unit 140 to the facility reservation management device 200.

[0110] The transmission unit 150 transmits the UG score calculated by the estimation unit 140 to the facility reservation management device 200.

[0111] <<Facility Reservation Management Device 200>> Returning to Figure 2, the functional configuration of the facility reservation management device 200 will be described. As shown in Figure 2, the facility reservation management device 200 includes, for example, a storage unit 210, an acquisition unit 220, and a transmission unit 230 as functional units.

[0112] The memory unit 210 stores, for example, reservation information of current and past customers. The reservation information stored in the memory unit 210 may be, for example, information stored in the items included in the customer information D111 shown in Figure 3.

[0113] The acquisition unit 220 acquires, for example, current reservation information from the person making the reservation. The acquisition unit 220 also acquires, for example, the UG score from the fraud detection device 100.

[0114] The transmitting unit 230 transmits, for example, the current reservation information obtained from the reservation holder to the fraudulent reservation estimation device 100. The transmitting unit 230 may also transmit, for example, past reservation information to the fraudulent reservation estimation device 100 in response to a request from the fraudulent reservation estimation device 100.

[0115] ===Processing Procedure=== The processing procedure of the fraudulent reservation detection system 10 will be explained with reference to Figure 7. Figure 7 is a flowchart showing the processing procedure of the fraudulent reservation detection system 10. As an example, Figure 7 describes the processing procedure in which the fraud estimation device 100 sends a UG score, indicating the possibility that the reservation is intended for fraudulent activity, to the facility reservation management device 200, based on the current reservation information regarding the reservation made by the reservation holder.

[0116] In step S101, the facility reservation management device 200 receives, for example, user input and transmits reservation information regarding the accommodation facility to the fraud detection device 100.

[0117] In step S102, the fraud detection device 100 stores the information contained in the acquired current reservation information in the storage unit 110. At this time, the fraud detection device 100 may, for example, generate information corresponding to the items shown in the reservation status information D112 and the period reservation count information D113 (hereinafter referred to as "reservation-related information") based on the information contained in the current reservation information and the past reservation information of the person making the reservation.

[0118] In step S103, the fraud estimation device 100 inputs, for example, the current reservation information and reservation-related information into the evaluation model.

[0119] In step S104, the fraud estimation device 100 calculates a UG score from the evaluation model, for example.

[0120] In step S105, the fraud detection device 100 transmits the UG score to the facility reservation management device 200.

[0121] In step S106, the facility reservation management device 200 displays the received UG score.

[0122] As a result, the fraudulent booking detection system 10 can accurately estimate, based on the booking information for the accommodation facility, whether or not the booking was made by a person with the intention of engaging in fraudulent activity.

[0123] ===Hardware Configuration=== Referring to Figure 8, an example of a hardware configuration when the fraud detection device 100 and the facility reservation management device 200 are implemented using a computer 1000 will be described. Note that the various functions of the fraud detection device 100 and the facility reservation management device 200 can be implemented by dividing them among multiple devices.

[0124] Figure 8 shows an example of a computer hardware configuration. As shown in Figure 7, the computer 1000 includes, for example, a processor 1001, memory 1002, storage device 1003, input I / F unit 1004, data I / F unit 1005, communication I / F unit 1006, and display unit 1007.

[0125] The processor 1001 is a control unit that controls various processes in the computer 1000 by executing programs stored in the memory 1002.

[0126] Memory 1002 is a storage medium such as RAM (Random Access Memory). Memory 1002 temporarily stores the program code of the program executed by the processor 1001, as well as data required during program execution.

[0127] The storage device 1003 is a non-volatile storage medium such as a hard disk drive (HDD) or flash memory. The storage device 1003 stores the operating system and various programs for realizing the above configurations.

[0128] The input interface unit 1004 is a device for receiving input from the user. Specific examples of the input interface unit 1004 include keyboards, mice, touch panels, various sensors, and wearable devices. The input interface unit 1004 may be connected to the computer 1000 via an interface such as USB (Universal Serial Bus).

[0129] The data I / F unit 1005 is a device for inputting data from outside the computer 1000. Specific examples of the data I / F unit 1005 include drive devices for reading data stored on various storage media. The data I / F unit 1005 may also be located outside the computer 1000. In that case, the data I / F unit 1005 would be connected to the computer 1000 via an interface such as USB.

[0130] The communication interface unit 1006 is a device for performing data communication via the Internet N with external devices of the computer 1000, either via wired or wireless connection. The communication interface unit 1006 may also be located outside the computer 1000. In that case, the communication interface unit 1006 is connected to the computer 1000 via an interface such as USB.

[0131] The display unit 1007 is a device for displaying various types of information. Specific examples of the display unit 1007 include liquid crystal displays, organic EL (Electro-Luminescence) displays, and displays for wearable devices. The display unit 1007 may be located outside the computer 1000. In that case, the display unit 1007 is connected to the computer 1000 via, for example, a display cable. Furthermore, if a touch panel is used as the input I / F unit 1004, the display unit 1007 can be integrated with the input I / F unit 1004.

[0132] ===Summary=== The fraudulent reservation detection system 10 includes: an acquisition unit 120 that acquires reservation information, which is information about a current reservation of an accommodation facility by a predetermined reservation holder; a learning unit 130 that learns from information included in past reservation information of multiple accommodation facilities, which is reservation attribute information, which is information about at least one of the following: the attributes of the reservation holder or prospective guest of the accommodation facility, the number of prospective guests, the planned check-in time registered by the reservation holder, and the name of the reservation holder; reservation status information, which is information about the reservation status of the same planned stay date or the same accommodation facility by the reservation holder at a predetermined point in the past; and period reservation number information, which is information about the number of reservations or number of nights of stay by the reservation holder during a predetermined period in the past, along with information indicating whether each of the multiple accommodation facilities was used for fraudulent purposes, as training data and outputs information indicating the risk that the reservation holder will use the accommodation facility in question for fraudulent purposes; and an estimation unit 140 that estimates the possibility that the reservation holder will use the accommodation facility in question for fraudulent purposes based on the output result output from the learning unit 130 when the current reservation information is input to the learning unit 130. This allows the fraudulent booking detection system 10 to notify travel agencies and accommodation facilities about potential fraudulent bookings made by individuals. As a result, travel agencies and accommodation facilities can take appropriate action against the individuals making the bookings (e.g., cancel the bookings).

[0133] Furthermore, the learning unit 130 of the fraudulent reservation detection system 10 learns from the information included in the reservation holder information, specifically gender information indicating the gender of the person scheduled to stay at the accommodation facility, the number of people scheduled to stay at the accommodation facility, and the check-in time information indicating the check-in time of the person scheduled to stay at the accommodation facility, as training data. As a result, the fraudulent reservation detection system 10 learns an evaluation model by combining information that indicates the characteristics of a person who makes a reservation with the intention of fraudulent activity, and can therefore estimate with greater accuracy whether or not the person made the reservation with the intention of fraudulent activity.

[0134] Furthermore, the learning unit 130 of the fraudulent reservation detection system 10 can also further learn name difference information, which indicates the difference between the name of the person making the reservation and the name of the person scheduled to stay, as training data from the information included in the reservation holder information. As a result, the fraudulent reservation detection system 10 has further trained its evaluation model by adding prominent characteristics of people who make reservations with the intention of fraudulent activity, so it can estimate with greater accuracy whether or not a person made a reservation with the intention of fraudulent activity.

[0135] Furthermore, the learning unit 130 of the fraudulent booking detection system 10 can also be further trained using at least one of the following as training data: capacity information indicating the maximum number of people that can be accommodated in a room at the accommodation facility to be booked; payment method information indicating the payment method that the prospective guest at the accommodation facility will use; age group information indicating the age group of the prospective guest at the accommodation facility to be booked; number of rooms information indicating the total number of rooms at the accommodation facility to be booked; and accommodation category information indicating the hotel category of the accommodation facility to be booked. This allows the fraudulent booking detection system 10 to estimate with greater accuracy who has made a reservation with the intention of fraudulent activity.

[0136] Furthermore, the learning unit 130 of the fraudulent reservation detection system 10 can also learn as training data at least one of the following from the reservation status information: information on the number of reservations for the same accommodation date, which shows the number of reservations made by the reservation holder on the same day as the planned accommodation date of the accommodation in question at a predetermined point in the past; information on the number of reservations for different accommodations, which shows the number of reservations made by the reservation holder on the planned accommodation date of the accommodation in question; and information on the number of consecutive nights of stay at the same accommodation, which shows the number of reservations made by the reservation holder for consecutive days at the same accommodation. This allows the fraudulent reservation detection system 10 to evaluate with even greater accuracy whether or not a reservation is made with the intention of fraudulent activity.

[0137] Furthermore, the learning unit 130 of the fraudulent reservation detection system 10 can also learn at least one of the following as training data: information included in the period reservation count information, which shows the maximum number of reservations for the same accommodation facility on the same day by the same reservation holder during a predetermined period in the past; information showing the maximum number of reservations for accommodation facilities other than the accommodation facility in question, made by the same reservation holder on the same scheduled accommodation date; information showing the total number of nights the same accommodation facility was used by the same reservation holder; and information showing the total number of facilities used by the same reservation holder. This allows the fraudulent reservation detection system 10 to evaluate with even greater accuracy whether or not a reservation is made with the intention of fraudulent activity.

[0138] The embodiments described above are provided to facilitate understanding of the present invention and are not intended to limit its interpretation. The elements, arrangement, materials, conditions, shapes, and sizes of the embodiments are not limited to those exemplified and can be modified as appropriate. Furthermore, configurations shown in different embodiments can be partially substituted or combined. [Explanation of symbols]

[0139] 10...Fraudulent reservation detection system, 100...Fraudulent reservation estimation device, 110...Storage unit, 120...Acquisition unit, 130...Learning unit, 140...Estimation unit, 150...Transmission unit, 200...Facility reservation management device.

Claims

1. An acquisition unit that acquires reservation information, which is information regarding the current reservation of accommodation by a designated person, Among the information contained in past booking records for multiple accommodations, Reservation information, which includes information relating to at least one of the following: the attributes of the person who made the reservation or is scheduled to stay at the accommodation facility, the number of people scheduled to stay, the scheduled check-in time registered by the person who made the reservation, and the name of the person who made the reservation. Reservation status information, which is information regarding the reservation status of the aforementioned reservation holder for the same planned stay date or the same accommodation at a predetermined point in the past, and Period reservation information, which is information regarding the number of reservations or nights of stay at accommodations made by the aforementioned reservation holder during a specified period in the past. At least one of the following, Information indicating whether each of the aforementioned accommodations was used for fraudulent purposes, A learning unit that learns using the above as training data and outputs information indicating the degree to which the person who made the reservation made the reservation for fraudulent purposes by inputting the above-mentioned current reservation information, A fraudulent booking detection system, including one.

2. The learning unit, of the information included in the reservation information, Gender information indicating the gender of the person scheduled to stay at the accommodation facility being booked, The number of people who are scheduled to stay at the accommodation facility being booked, The fraudulent reservation detection system according to claim 1, which learns check-in time information indicating the check-in time of a person scheduled to stay at the accommodation facility to be reserved, as training data.

3. The learning unit further learns, using as training data, name difference information, which shows the difference between the name of the person making the reservation and the name of the person scheduled to stay, from the information included in the reservation information. The fraudulent reservation detection system according to claim 2.

4. The learning unit, of the information included in the reservation information, The information includes the maximum occupancy of the room at the accommodation facility being booked, and Payment method information indicating the payment method that the person planning to stay at the accommodation facility in question intends to use, Age group information indicating the age range of guests scheduled to stay at the accommodation facility being booked, Room count information showing the total number of rooms in the accommodation facility eligible for booking, Accommodation classification information that indicates the hotel classification of the accommodation facility to be booked, The fraudulent reservation detection system according to claim 3, wherein at least one of the above is learned as training data.

5. The learning unit, of the information included in the reservation status information, At a specific point in the past, The information includes the number of reservations made by the same person for the same day as the planned stay date at the accommodation facility in question, and Information on the number of reservations for the same accommodation on the scheduled stay date of the accommodation in question, which shows the number of reservations for accommodations other than the accommodation in question that the aforementioned reservation holder has also booked, and Information on consecutive nights of stay at the same accommodation, which shows the number of reservations made by the same person at the same accommodation, The fraudulent reservation detection system according to any one of claims 1 to 4, wherein at least one of the above is learned as training data.

6. The learning unit, among the information included in the information on the number of reservations for the period, In a specified period in the past, The maximum number of reservations for the same accommodation on the same day by the same reservation holder is shown in the maximum number of reservations for the same accommodation on the same day, and The maximum number of reservations for the same accommodation on the same scheduled stay date, which is different from the accommodation booked by the same person, is shown in the maximum accommodation reservation information for the same scheduled stay date. The total number of nights stayed at the same accommodation facility, which shows the number of nights the same accommodation facility was used by the same person who made the reservation, Total information on the number of facilities used, which shows the number of facilities reserved by the same person who made the reservation, The fraudulent reservation detection system according to any one of claims 1 to 4, wherein at least one of the above is learned as training data.

7. Computers To obtain reservation information, which is information regarding the current reservation of accommodation by a designated person, Among the information contained in past booking records for multiple accommodations, Reservation information, which includes information relating to at least one of the following: the attributes of the person who made the reservation or is scheduled to stay at the accommodation facility, the number of people scheduled to stay, the scheduled check-in time registered by the person who made the reservation, and the name of the person who made the reservation. Reservation status information, which is information regarding the reservation status of the aforementioned reservation holder for the same planned stay date or the same accommodation at a predetermined point in the past, and Period reservation information, which is information regarding the number of reservations or nights of stay at accommodations made by the aforementioned reservation holder during a specified period in the past. At least one of the following, Information indicating whether each of the aforementioned accommodations was used for fraudulent purposes, The system learns from this data and outputs information indicating the degree to which the person who made the reservation likely booked the accommodation facility with fraudulent intent, A method for detecting fraudulent reservations.

8. On the computer, To obtain reservation information, which is information regarding the current reservation of accommodation by a designated person, Among the information contained in past booking records for multiple accommodations, Reservation information, which includes information relating to at least one of the following: the attributes of the person who made the reservation or is scheduled to stay at the accommodation facility, the number of people scheduled to stay, the scheduled check-in time registered by the person who made the reservation, and the name of the person who made the reservation. Reservation status information, which is information regarding the reservation status of the aforementioned reservation holder for the same planned stay date or the same accommodation at a predetermined point in the past, and Period reservation information, which is information regarding the number of reservations or nights of stay at accommodations made by the aforementioned reservation holder during a specified period in the past. At least one of the following, Information indicating whether each of the aforementioned accommodations was used for fraudulent purposes, The system learns from this data and outputs information indicating the degree to which the person who made the reservation likely booked the accommodation facility with fraudulent intent, A program that executes the command.

Images