JP6853526B2 - Reservation available time prediction system - Google Patents

Description

The present invention relates to a reservable time prediction system, and more particularly to a technique for predicting the reservable time in advance with respect to a limited number of goods and services such as accommodation rooms and airmail seats.

Recently, due to the growing awareness of health and the purpose of revitalizing local areas, large-scale citizen-participation-type sporting events are being held in various parts of Japan (non-patented). Document 1).
In addition, a support service that provides information on sports events held in various places and acts as an entry (application) for sports events that you wish to participate in has also appeared (Non-Patent Document 2).
By using such a support site, the user can easily enter various sporting events held all over Japan while staying at home.

Okinawa Marathon Internet URL: http://www.okinawa-marathon.com/ Search date: August 30, 2016 Your sport is here / SPORTS ENTRY Internet URL: https://www.sportsentry.ne.jp/ Search date: August 30, 2016

By the way, when participating in a distant sporting event, it is naturally necessary to arrange accommodations locally and secure seats for airplanes and railroads.
However, the reality is that many users feel relieved when they enter a sports competition and postpone the reservation of accommodation facilities and the securing of airline tickets to the last minute.
For this reason, although they made entries for sporting events, there were cases where they could not secure accommodation or transportation at the site.

The present invention has been devised to solve the above-mentioned conventional problems, and is a specific reservation target (for example, the day before the date of a large event such as a sports competition) on a future reservation target date (for example, in a specific area). The purpose is to provide technology that can present to the user in advance when reservations can be made for accommodation facilities and transportation to the area.

In order to achieve the above object, the reservable time prediction system according to claim 1 stores the past reservation record data including the reservation target date and the reservation acceptance date for a specific reservation target object. It is provided with a storage means, and when a prediction condition indicating a future reservation target date is input, a process for identifying a past reservation target date corresponding to this future reservation target date and storage in the reservation record data storage means. From the reservation record data, the process of extracting the reservation record data that matches the past reservation target date and the extracted reservation record data are aggregated based on the reservation reception date to reach the past reservation target date. The process of calculating the cumulative number of reservations for the above reserved objects for each day up to, and the process of dividing the cumulative number of reservations for each day by the quantity of the above reserved objects to calculate the daily reservation rate, and these past It is characterized in that a process of estimating a future reservable time and a process of outputting the reservable time are executed based on the reservation rate of each day up to the reservation target date.

The "past reservation target date corresponding to the future reservation target date" corresponds to, for example, the same month and day one year before the "future reservation target date". Alternatively, when a large-scale event is held on the "future reservation target date", the past month and day when the event was held corresponds to this.

"When a prediction condition indicating a future reservation target date is input" is not limited to the case where the "future reservation target date" is explicitly input by the user. For example, when application information for an event to be held at a specific date and time in the future is registered, the date or the day before the event is recognized as a "future reservation target date", and the system automatically automatically recognizes it. It is also possible to estimate the time when reservations can be made and output it in the form of sending an e-mail to the users participating in the event or presenting a WEB page.

The reservable time prediction system according to claim 2 is the system according to claim 1, and further, a plurality of the past reservation target dates are specified, and each reservation is made for each of the plurality of reservation target dates. The daily reservation rate up to the target date is calculated, the average value of each reservation rate is calculated for each of the corresponding multiple days, and future reservations are made based on the average value of these past daily reservation rates. The feature is that the possible time is estimated.

The reservable time prediction system according to claim 3 is the system according to claim 1, and further, a plurality of the past reservation target dates are specified, and each reservation is made for each of the plurality of reservation target dates. The adjusted reservation rate is calculated by applying a predetermined weight to the daily reservation rate up to the target date, the average value of the adjusted reservation rate is calculated for each of the corresponding multiple days, and these past daily adjustments are made. It is characterized in that the future reservation available time is estimated based on the average value of the reservation rate.

In claim 2 or 3, "multiple past reservation target dates are specified" means, for example, the same date one year before the "future reservation target date", the same month date two years ago, and three years ago. It means that the same date ... is specified. Alternatively, when a large-scale event is held on the "future reservation target date", the date one year before the event was held, the date two years ago, the date three years ago, and so on are ". It means that it is specified as "past reservation target date".

The reservable time prediction system according to claim 4 is the system according to claims 1 to 3, further, with respect to the reservation rate or the adjusted reservation rate for each day up to the past reservation target date. By applying the first judgment rule with a predetermined reservation rate or the adjusted reservation rate as the threshold value, the reliable reservation deadline is recognized, and the second judgment rule with the higher reservation rate or the adjusted reservation rate as the threshold value is recognized. By applying, the reservable deadline is recognized, and these reliable reservable deadlines and reservable deadlines are output as the above-mentioned future reservable deadlines.

The reservable time prediction system according to claim 5 is the system according to claims 1 to 4, and further, the holding period of the event held in a specific area and the number of participants on a daily basis during the holding period are determined. The event information storage means is stored, and the reservation target has a relationship with a specific area. As a result of referring to the event information storage means, in the related area of the reservation target, the past If it is detected that an event that was held on the reservation target date or the next day will not be held on the future reservation target date or the next day, at least a part of the number of participants for each day of the event will be accumulated. It is characterized by subtracting from the number to calculate the reservation rate for each day in the past.

The reservable time prediction system according to claim 6 is the system according to claims 1 to 4, and further, the holding period of the event held in a specific area and the number of participants on a daily basis during the holding period are determined. The event information storage means is stored, and the reservation target has a relationship with a specific area. As a result of referring to the event information storage means, in the related area of the reservation target, the past If it is detected that an event that was not held on the reserved day or the next day will be held on the future reserved day or the next day, at least a part of the number of participants for each day of the event will be added to the above. It is characterized by calculating the reservation rate for each day in the past by adding it to the cumulative number of reservations.

The reservable time prediction system according to claim 7 is the system according to claims 1 to 6, and further, the reservation target is a seat of an airline flight, and the flight number, departure airport, and the like of each airline flight. It is equipped with an airline flight storage means that stores the departure time, arrival airport, arrival time, and number of seats, and the reservation record data storage means stores the past reservation record data of each flight, as the above prediction condition. , When the future reservation target date and the departure airport and arrival airport are entered, the flights that match this prediction condition are extracted from the above flight storage means, and from the above reservation record data storage means to each flight. The feature is that the past reservation record data is extracted, the future reservation availability time of each flight is estimated based on these reservation record data, and the above reservation record time is output for each flight number. ..

The reservable time prediction system according to claim 8 is the system according to claims 1 to 6, further, the object to be reserved is a guest room of an accommodation facility, and the name, location, and price range of each accommodation facility. An accommodation facility storage means in which another number of guest rooms is stored is provided, and the past reservation record data of each accommodation facility is stored in the reservation record data storage means. When the area is input, accommodation facilities that match this prediction condition are extracted from the above-mentioned accommodation facility storage means, and past reservation record data related to each accommodation facility are extracted from the above-mentioned reservation record data storage means. The feature is that the future reservable time of each price range is estimated based on the reservation record data, and the above reservable time is output for each price range.

According to the reservable time prediction system described in claim 1, the reservable time for the future reservation target date can be estimated based on the past reservation record data and presented to the user in advance, so that the reservation can be made carelessly. It is possible to avoid the occurrence of a situation where the time is missed.

In the case of the reservable time prediction system according to claim 2 or 3, the average value of the reservation rate or the adjusted reservation rate between a plurality of corresponding past reservation record data is calculated for each day, and the future is based on this. Since it is a method of estimating the reservable time, it is possible to even out the fluctuation of each year and obtain a more reliable prediction result.

According to the reservable time prediction system according to claim 4, the user is presented with both a more certain "reservable deadline" and a less certain "reservable deadline". Can be done.

In the case of the reservable time prediction system according to claim 5 or 6, since the influence of the presence or absence of the event can be reflected in the estimation of the reservable time, a more accurate reservable time can be derived.

According to the reservable time prediction system according to claim 7, since the user can recognize the reservable time in advance for the reservation of a specific air flight, it is possible to effectively prevent the reservation error of the air flight.

According to the reservable time prediction system described in claim 8, the user can recognize in advance the reservable time for each price range of accommodation facilities in a specific area, so that reservation mistakes of accommodation facilities can be effectively prevented. it can.

As shown in FIG. 1, the reservable time prediction system 10 according to the present invention includes a WEB server 12, a reservable time prediction server 14, and a DB server 16.
The WEB server 12, the reservable time prediction server 14, and the DB server 16 are connected to each other via a communication network.

The WEB server 12 is connected to a large number of client terminals 20 operated by the user via the Internet 18.
The client terminal 20 corresponds to a PC, tablet terminal, smartphone or the like equipped with a communication function and a WEB browser program.

The reservable time prediction server 14 includes an OS and a dedicated application program, and the CPU of the server computer executes necessary processing according to the application program to exhibit various functions described later.

The DB server 16 includes an OS and a database management program, and in its external storage device 22, an event DB 24, an event application DB 26, a user DB 28, a weather forecast DB 29, an accommodation facility DB 30, an accommodation reservation DB 32, and an accommodation facility prediction. Various databases such as DB34, airmail DB36, airmail reservation DB38, and airmail forecast DB40 are stored.

The DB server 16 is connected to a large number of external servers 42 and data registration terminals 44 via the Internet 18.
Examples of the external server 42 include an airline flight reservation management server, a reservation management server for each accommodation facility, a reservation management server for an event reservation agency, and the like.
The data registration terminal 44 corresponds to a PC or the like managed by the operator of this system 10.

The above event DB24 contains information on various events that are expected to attract a relatively large number of customers (mobilization) such as sports events, music events, exhibitions, expositions, motor shows, academic societies, summits, etc. held in various parts of Japan. It is stored.
For example, as shown in FIG. 2A, a record having data items such as event ID, event name, event category, start date / time, end date / time, venue, number of participants per day, organizer, registration date / time, etc. , Stored in event DB24.
These event information are transmitted from the data registration terminal 44 at any time, and are stored in the event DB 24 via the DB server 16.

Further, the event application DB 26 stores user application information for a specific event.
For example, as shown in FIG. 2B, a record including data items such as an event application ID, an event ID, a user ID, a participation date, and an application acceptance date / time is stored in the event application DB 26.
These event application information is transmitted from an external server 42 (reservation management server of an event reservation agency, etc.) at any time, and is stored in the event DB 24 via the DB server 16.

The attribute information of each user is stored in the user DB 28.
For example, as shown in FIG. 2C, a record including data items such as a user ID, a name, an address, a gender, a date of birth, an e-mail address, and a registration date is stored in the user DB 28.
These user information are transmitted from the client terminal 20 to the WEB server 12 at the time of user registration, and are stored in the user DB 28 via the DB server 16.

The weather forecast information is stored in the weather forecast DB 29.
For example, as shown in FIG. 2D, a record including data items such as a weather forecast ID, an applicable area, an applicable period, an increase / decrease rate, and a registration date is stored in the weather forecast DB 29.
These weather forecast information are transmitted from the data registration terminal 44 at any time and stored in the weather forecast DB 29 via the DB server 16.

The accommodation facility DB 30 stores attribute information of each accommodation facility.
For example, as shown in FIG. 3A, a record having data items such as an accommodation facility ID, a name, a category (hotel / inn / pension, etc.), a location, and the number of guest rooms by price range is stored in the accommodation facility DB30. Has been done.
These accommodation facility information is transmitted from the data registration terminal 44 at any time, and is stored in the accommodation facility DB 30 via the DB server 16.

The accommodation reservation DB 32 stores accommodation reservation information (reservation record data) at accommodation facilities such as major hotels and inns in each region.
For example, as shown in FIG. 3B, a record including data items such as an accommodation reservation ID, an accommodation facility ID, an accommodation date, a price range, and a reservation acceptance date / time is stored in the accommodation reservation DB 32.
These accommodation reservation information is transmitted from an external server 42 (reservation management server of each accommodation facility, etc.) at any time, and is stored in the accommodation reservation DB 32 via the DB server 16.

In the accommodation facility prediction DB 34, the prediction data of the reservation possible time calculated by the reservation possible time prediction server 14 is sequentially stored via the DB server 16.
For example, as shown in FIG. 3 (c), data items such as accommodation facility prediction ID, user ID, prediction target date, prediction target area, price range, reliable reservation possible deadline, reservation possible deadline, and prediction execution date and time are provided. The record is stored in the accommodation facility prediction DB 34.

The airmail DB36 stores the attribute information of each airmail.
For example, as shown in FIG. 4A, a record including data items such as flight ID, flight number, departure airport, arrival airport, departure time, arrival time, number of seats, etc. is stored in the flight DB36. ing.
These airmail information is transmitted from the data registration terminal 44 at any time, and is stored in the airmail DB 36 via the DB server 16.

The flight reservation DB 38 stores flight reservation information (reservation record data) for each airline's flight.
For example, as shown in FIG. 4B, a record including data items such as an airmail reservation ID, an airmail ID, a boarding date, and a reservation acceptance date / time is stored in the airmail reservation DB 38.
These flight reservation information is transmitted from an external server 42 (airline reservation management server of each airline company, etc.) at any time, and is stored in the flight reservation DB 38 via the DB server 16.

In the flight forecast DB 40, the forecast data of the reservable time calculated by the reservable time prediction server 14 is sequentially stored via the DB server 16.
For example, as shown in FIG. 4 (c), a record including data items such as an airmail forecast ID, a user ID, a forecast target date, a departure airport, an arrival airport, a reliable reservation deadline, a reservation deadline, and a forecast execution date and time. Is stored in the airmail forecast DB40.

Next, according to the flowchart of FIG. 5, the processing procedure of predicting the reservation available time of the accommodation facility by this system 10 will be described.
First, when a request for predicting the availability of accommodation facilities is sent from the client terminal 20 to the WEB server 12 (S10), the prediction condition specification screen (not shown) is displayed from the WEB server 12 to the client terminal 20. It is transmitted (S12).

On the other hand, when the user specifies the scheduled accommodation date (for example, "December 1, 2016") and the area (for example, "Naha City, Okinawa Prefecture") on the above screen, these conditions are set from the client terminal 20 to the WEB server. It is transmitted to 12 (S14).

The reservable time prediction server 14 that has received these prediction conditions via the WEB server 12 calculates the "certain reservable deadline" and the "reservable deadline" through the processes of S16 to S20 described later, and passes them to the WEB server 12.
Next, the reservationable time prediction server 14 generates accommodation facility prediction data including the “certain reservation possible deadline” and the “reservable deadline”, and stores the accommodation facility prediction data in the accommodation facility prediction DB 34 via the DB server 16 (S22).

On the other hand, the WEB server 12 generates a prediction result presentation screen in which the "certain reservation possible deadline" and the "reservable possible deadline" are described, and transmits the prediction result presentation screen to the client terminal 20 (S24).
FIG. 6 shows the prediction result presentation screen 50 regarding the accommodation facility displayed on the WEB browser of the client terminal 20, and as shown in the figure, the price ranges (1) to (5) are displayed on this screen. "Reservable deadline" and "Reservable deadline" are displayed.
Here, the "reliable reservation deadline" means a date on which reservations for the scheduled accommodation date can be made with a margin.
In addition, the "reservation deadline" indicates the limit point at which a reservation can be made with respect to the scheduled accommodation date, and means that it will be extremely difficult to make a reservation from the next day onward.

For example, in the case of a low-priced accommodation facility of less than 3,000 yen per night, it is necessary to move by the guaranteed reservation deadline of "August 21, 2016" in order to secure the reservation on December 1, 2016. It will be.
Of course, it is possible to make a reservation even after this certain reservation deadline, but even so, it will be very difficult to make a reservation after the reservation deadline of "September 21, 2016".

Therefore, it is safe for users to make a reservation by "August 21, 2016" in order to secure a relatively cheap accommodation for less than 3,000 yen, and make a reservation by "September 21, 2016" at the latest. You can recognize in advance that you must put in.
On the other hand, if the user wants to secure an expensive accommodation facility of 12,000 yen or more per night, it can be seen that the reservation can be reserved until the reservation deadline of "November 27, 2016" at the maximum.

The user who decides to make a reservation at the present time (June 1, 2016) clicks the reservation button 52 related to the desired price range.
As a result, the client terminal 20 is connected to the WEB server of the travel agency, and a reservation application screen (not shown) listing accommodation facilities in the desired price range is displayed.

On the other hand, if the reservation deadline in the desired price range has passed (for example, when making a reservation in the price range of less than 3,000 yen from the beginning of November), the user can enter "Forecast condition change column 54". After reselecting "Forecast Date" or "Forecast Region", click the Re-Forecast button 56.
As a result, a screen listing the prediction results that match the changed prediction conditions is displayed on the client terminal 20.

Even if you cannot make a reservation for accommodations for less than 3,000 yen on "December 1, 2016" the day before the event in "Naha City" where the event is held, if you expand the area to the neighborhood, it will be November. Since there is a possibility that reservations can be made even at a certain point in time, it is meaningful to change the prediction conditions and re-predict as described above.

Next, the calculation method of the above-mentioned "reliable reservation deadline" and "reservable deadline" will be described.
First, the reservable time prediction server 14 refers to the accommodation facility DB30 and extracts the accommodation facilities located in the designated area (S16).
Next, the reservation available time prediction server 14 refers to the accommodation reservation DB 32, and calculates the reservation rate for the past predetermined day within the past predetermined period for each of the extracted accommodation facilities on a daily basis (S18).

For example, if "December 1, 2016" is the target date for this reservation, the reservable time prediction server 14 corresponds to December 1, 2015, which is one year before that, as the target date for this reservation. For each month and day belonging to December 2, 2014, which is recognized as a "past reservation target date" and goes back one year from this, the reservation rate for December 1, 2015 is set to the designated area. Calculated according to the price range of the accommodation facility where you are located.

Assuming that the total number of guest rooms in all accommodations in Naha City, Okinawa Prefecture, whose price range is "less than 3,000 yen" is 1,000, and the number of reservations for "December 1, 2015" as of July 1, 2015. If there are 300 rooms, the reservation rate on 2015/07/01 = 30%.
On the other hand, if the number of reservations for "December 1, 2015" of the same accommodation facility as of September 1, 2015 was 750, "Reservation rate for 2015/09/01 = 75%". Become.

Next, the reservable time prediction server 14 analyzes the transition of the daily reservation rate within the above period, and specifies the "reliable reservation deadline" and "reservable deadline" for December 1, 2015 for each price range. (S20).

For example, as shown in Fig. 7 (a), for accommodations with a price range of "less than 3,000 yen", the reservation rate on August 21, 2015 was "46%", while the reservation rate was "46%" on August 22, 2015. If this has risen to "52%", the reservation available time prediction server 14 will be on August 21 according to the first judgment rule that "the day before the day when the reservation rate exceeds 50% is set as the reliable reservation deadline". Recognize the day as the "reliable reservation deadline".

Also, if the reservation rate on September 21, 2015 was "87%", but it increased to "91%" on September 22, the following day, the "reservation rate exceeded 90%". According to the second determination rule that "the day before is the reservable deadline", the reservable time prediction server 14 recognizes September 21 as the "reservable deadline".
It goes without saying that the above judgment rule is just an example, and other rules can be applied (the same applies hereinafter).

Alternatively, as shown in Fig. 7 (b), for accommodations with a price range of "8,000 yen or more and less than 12,000 yen", the reservation rate on November 1, 2015 was "48%", while the following 11 If this rises to "51%" on the 2nd of the month, the reservation available time prediction server 14 will follow the first judgment rule that "the day before the day when the reservation rate exceeds 50% will be the sure reservation deadline". November 1st is recognized as the "reservable deadline".
Also, if the reservation rate on November 20, 2015 was "88%", but it increased to "90%" on November 21, 2015, the "reservation rate exceeded 90%". According to the second judgment rule that "the day before is the reservable deadline", the reservable time prediction server 14 recognizes November 20 as the "reservable deadline".

Next, the reservable time prediction server 14 simply adds one year to the above-mentioned "reliable reservable deadline" and "reservable deadline" for December 1, 2015, so that December 1, 2016 Estimate the "reservable deadline" and "reservable deadline" for the day.

As a result, for accommodations with a price range of "less than 3,000 yen", the reservation deadline for December 1, 2016 will be "August 21, 2016", and the reservation deadline will be "September 21, 2016". ".
Similarly, for accommodations with a price range of "8,000 yen or more and less than 12,000 yen", the reliable reservation deadline for December 1, 2016 is "November 1, 2016", and the reservation deadline is "November 2016". It will be "20th of the month".

The above calculation method is particularly effective when the same large-scale event is held in the same area in the same period every year, but the accuracy can be further improved by reflecting the change in the environment every year.
Hereinafter, a specific description will be given.

[Correction when the event date is determined by the day of the week]
In other words, in the case where the reservation status of accommodation facilities in a specific area fluctuates greatly depending on whether or not a specific event is held, if the date of the event is determined not by the month and day but by the day of the week, the reservation target date of this time. It is desirable to specify the event date of last year as the "past reservation target date" corresponding to.

For example, if the date of the NAHA Marathon is specified as "the first Sunday of December every year", a forecast request with "December 3, 2016 (the day before the 2016 date)" as the reservation target date If entered, the reservation possibility prediction server 14 recognizes it as an "event-related reservation" and identifies "December 5, 2015", which is the day before the same event date last year, as the "past reservation target date". ..

As a result, if "August 10, 2015" is approved as a reliable reservation deadline for "December 5, 2015" based on the results of last year, the difference in the event date will be reflected and "2016" two days ago. "August 8, 2014" is estimated to be the reliable reservation date for this year.
However, in large cities such as Tokyo and Osaka, which have a large population, international super-large events such as the Olympic Games and the World's Fair are just a corner, and holding general events causes such a big change in accommodation reservation trends. Since this is not the case, it is not necessary to perform the above correction process.

[Reflection of reservation record up to now]
Comparing the reservation rate on the specified date of this year or the average reservation rate in the specified period of this year with the reservation rate on the specified day of last year or the average reservation rate in the corresponding period of last year, there is a certain difference between the two. If so, it is applicable to reflect this in the calculation result.

For example, the reservation rate for accommodations with a price range of "less than 3,000 yen" for December 1, 2016 has reached 32% at the present time (June 1, 2016), but at the same time last year ( If it was 20% on June 1, 2015), as shown in Fig. 8 (a), it corresponds to uniformly adding 12% to the daily reservation rate calculated based on the results of last year. ..

As a result, the secure reservation deadline for December 1, 2016 will move to a time before August 21, 2016, and the reservation deadline will also move to a time before September 21, 2016. It becomes.
Of course, instead of adding the increased 12% as it is, it is possible to stop adding only a part (for example, 6% of half).

On the contrary, the reservation rate of accommodations with a price range of "less than 3,000 yen" for December 1, 2016 is only 8% at the present time (June 1, 2016), but at the same time last year. If the point (June 1, 2015) is 20%, as shown in Fig. 8 (b), the daily reservation rate calculated based on the results of last year is uniformly reduced by 12%. Will be done.

As a result, the secure reservation deadline for December 1, 2016 will move to a time after August 21, 2016, and the reservation deadline will also move to a time after September 21, 2016. It becomes.
In this case as well, instead of subtracting the reduced 12% as it is, it is possible to stop the subtraction of a part (for example, 6% of half).

[Reflecting the impact of holding an event]
If a large event that was held in the forecast area at the forecast time last year is not held this year, if the date is changed, or if a large event that did not exist last year is held for the first time this year , It is desirable to reflect the planned number of participants and the number of participants in the forecast results.

For example, if the date of the event (number of participants: 3,000) held in the same area on December 1st of last year (2015) is changed to December 8th of this year (2016), December 2016 When predicting the available reservation time for the 1st of the month, divide 1,500, which is 50% of the number of participants, by 5, which is the number of price ranges, to obtain the number of reductions of "-300".
Next, when calculating the reservation rate by price range for each day belonging to the period from December 2, 2014 to December 1, 2015, subtract 300 from each reservation number.

As a result, for example, the reliable reservation deadline for December 1, 2016 in the price range of "less than 3,000 yen" will be postponed after August 21, 2016, and the reservation deadline for December 1, 2016 will be delayed. However, it will be postponed after September 21, 2016.
Even if the scale of the existing event (scheduled number of participants and capacity) is significantly reduced, the same correction process as above is performed.

On the contrary, if a new event (number of participants: 3,000) that did not exist on December 1st of last year (2015) will be held on December 1st of this year (2016), 2016 When predicting the available reservation time for December 1, 2014, divide 1,500, which is 50% of the number of participants, by 5, which is the number of price ranges, to obtain the increase of "+300".
Next, when calculating the reservation rate by price range for each day belonging to the period from December 2, 2014 to December 1, 2015, add 300 to each reservation number and calculate.

As a result, for example, the deadline for secure reservations for December 1, 2016 in the price range of "less than 3,000 yen" has been advanced before August 21, 2016, and reservations for December 1, 2016 are possible. The deadline will also be advanced before September 21, 2016.
Even if the scale of the existing event (planned participation number and capacity) is significantly expanded, the same correction process as above will be performed.

In the above, an example is shown in which 50% of the number of participants increasing or decreasing is reflected in the prediction result, but this ratio is only an example, and it is naturally possible to apply other ratios.
Also, instead of simply dividing the increase or decrease number calculated based on the above ratio by the number of price ranges to obtain the number to be added or subtracted from the number of reservations in each price range, each is according to a predetermined rule. It is also possible to calculate the number of increase or decrease to be allocated to the price range (for example, the price range of "less than 3,000 yen": 50%, the price range of "3,000 yen or more and less than 5,000 yen": 20%, etc.).

[Reflection of weather forecast information]
It is also possible to reflect the weather forecast information in the prediction result.
For example, if a long-range forecast is issued this winter (2016) that the Nagano prefecture region will be a warm winter, the demand for accommodation facilities around the ski resort is expected to decrease, so regarding accommodation facilities in that region. It is desirable to reflect this in the prediction process.

For this purpose, a record defining the applicable area, the applicable period, and the rate of increase / decrease is stored in the weather forecast DB 29 of the DB server 16 from the data registration terminal 44.
On the other hand, the reservationable time prediction server 14 refers to the weather forecast DB 29, and when the weather forecast data matching the prediction target area and the prediction target date requested by the user is registered, the increase / decrease rate is reserved in the past. Apply to the number of reservations calculated based on the data.

For example, for accommodations for "less than 3,000 yen" in Tateshina-cho, Kitasaku-gun, Nagano Prefecture, the reservation rate for December 1, 2015 reached 50% on September 1, 2015, based on the results of last year. Even so, since it is a warm winter this year, the rate of increase / decrease of "-10%" is uniformly applied, and the reservation rate on September 1, 2016 is predicted to be 40%.

[Utilization of accommodation facility forecast data]
As described above, the accommodation facility prediction DB 34 accumulates the prediction target date, the prediction target area, the price range, the reliable reservation deadline, and the reservation possible deadline each time the prediction process is executed in response to the user's request. By analyzing this prediction data later, it is possible to improve the prediction accuracy in this system 10.

That is, as follows, the current prediction accuracy is verified by comparing the prediction data stored in the accommodation facility prediction DB 34 with the actual data calculated based on the accommodation reservation data stored in the accommodation reservation DB 32. Can be done.

[Forecast data]
Forecast Date: June 1, 2016 Forecast Date: December 1, 2016 Forecast Area: Naha City, Okinawa Prefecture Price Range: Less than 3,000 Yen Reservation Available Deadline: August 21, 2016 (Reservation Rate 50%) Deadline not to exceed)
Reservation deadline: September 21, 2016 (reservation rate does not exceed 90%)

[Actual data (1)]
Analysis target date: August 21, 2016 Forecast target date: December 1, 2016 Forecast target area: Naha City, Okinawa Prefecture Price range: Less than 3,000 yen Reservation rate: 30%
Calculation date: August 22, 2016

[Actual data (2)]
Analysis target date: September 21, 2016 Forecast target date: December 1, 2016 Forecast target area: Naha City, Okinawa Prefecture Price range: Less than 3,000 yen Reservation rate: 85%
Calculation date: September 22, 2016

First, the actual data (1) shows the actual reservation rate on "August 21, 2016", which is the deadline for certain reservations in the forecast data, and it can be seen that this is only "30%". ..
As mentioned above, the reliable reservation deadline indicates the "deadline that does not exceed the reservation rate of 50%", so it means that there was actually a deviation of nearly 20%.

In addition, the actual data (2) shows the actual reservation rate on "September 21, 2016", which is the reservation deadline in the forecast data, and it can be seen that this is also only "85%". ..
As described above, the reservation deadline indicates the "deadline that does not exceed the reservation rate of 90%", so that there is actually a deviation of nearly 5%.

As described above, by repeating the comparison between each prediction data and the actual data, it is possible to grasp the current state of prediction accuracy, and based on this, it is also possible to optimize various parameters and correction values used for prediction. It becomes.

Next, according to the flowchart of FIG. 9, the processing procedure of predicting the reservable time of an air flight by this system 10 will be described.
First, when a request for predicting the available flight reservation time is sent from the client terminal 20 to the WEB server 12 (S30), the prediction condition specification screen (not shown) is displayed from the WEB server 12 to the client terminal 20. It is transmitted (S32).

On the other hand, on the above screen, the combination of the scheduled boarding date (for example, "December 1, 2016"), the departure airport (for example, "Haneda Airport") and the arrival airport (for example, "Naha Airport"), that is, When the route is specified, these conditions are transmitted from the client terminal 20 to the WEB server 12 (S34).

The reservable time prediction server 14 that has received these prediction conditions via the WEB server 12 calculates the "certain reservable deadline" and the "reservable deadline" through the processes of S36 to S40 described later, and passes them to the WEB server 12.
Next, the reservationable time prediction server 14 generates the flight prediction data including the “certain reservation possible deadline” and the “reservable deadline”, and stores the flight prediction data in the flight prediction DB 40 via the DB server 16 (S42).

On the other hand, the WEB server 12 generates a prediction result presentation screen in which the "certain reservation possible deadline" and the "reservable possible deadline" are described, and transmits the prediction result presentation screen to the client terminal 20 (S44).
FIG. 10 shows a prediction result presentation screen 70 regarding airmail displayed on the WEB browser of the client terminal 20, and as shown in the figure, each of the airmails (1) to (6) ... Is displayed on this screen. , "Reservable deadline" and "Reservable deadline" are displayed.
Here, the "reservable deadline" means the day when the reservation for the flight can be made with a margin, and the "reservation deadline" means the reservation for the flight. Means the possible limit.

For example, in order to make a reservation for flight (6) JAL 903 departing from Haneda at 7:55 on December 1, 2016, it is necessary to move by the certain reservation deadline of "September 15, 2016". It will be.
Of course, it is possible to make a reservation even after this certain reservation deadline, but even so, it will be very difficult to make a reservation after the reservation deadline of "October 30, 2016".

Therefore, it is safe for users to make a reservation by "September 15, 2016" in order to secure a ticket for JAL 903 flight on December 1, 2016, at the latest "October 30, 2016". Recognize that you must make a reservation by "day".
On the other hand, for users who wish to secure a ticket for ANA 461 flight on December 1, 2016, the reservation can be reserved until the reservation deadline of "November 7, 2016" at the maximum. Become.

The user who decides to make a reservation at this time (June 1, 2016) clicks the reservation button 72 for the desired flight.
As a result, the client terminal 20 is connected to the WEB server of the airline company or the travel agency, and the reservation application screen (not shown) of the desired airline flight is displayed.

On the other hand, if the reservation period of the desired flight has expired, the user reselects "Predicted date" or "Predicted route" in the prediction condition change field 74, and then clicks the re-prediction button 76. To do.
As a result, a screen listing the prediction results of the reservable time that matches the changed prediction conditions is displayed on the client terminal 20.

Next, the calculation method of the above-mentioned "reliable reservation deadline" and "reservable deadline" will be described.
First, the reservable time prediction server 14 refers to the airmail DB36 and extracts the airmail related to the designated route (S36).
Next, the reservation possible time prediction server 14 refers to the flight reservation DB 38, and calculates the reservation rate for the past predetermined day within the past predetermined period for each of the extracted flights on a daily basis (S38).

For example, if "December 1, 2016" is the desired reservation date for this time, the reservation available time prediction server 14 corresponds to December 1, 2015, which is one year before that, as the reservation target date for this time. For each month and day belonging to December 2, 2014, which is recognized as a "past reservation target date" and goes back one year from this, the reservation rate for December 1, 2015 is calculated for each extracted flight flight. calculate.

If ANA 461 has 200 seats and the number of reservations for "December 1, 2015" as of June 1, 2015 is 20, "Reservation for 2015/06/01" Rate = 10% ".
On the other hand, if the number of reservations for "December 1, 2015" on the same flight as of July 1, 2015 was 50, "Reservation rate for 2015/07/01 = 25%". Become.

Next, the reservable time prediction server 14 analyzes the transition of the daily reservation rate within the above period, and specifies the "reliable reservation deadline" and "reservable deadline" for December 1, 2015 for each flight. To do.

For example, as shown in Fig. 11, for (1) ANA 461 flight, the reservation rate on September 30, 2015 was "47%", but on October 1, 2015, it increased to "51%". If so, according to the first judgment rule that "the day before the day when the reservation rate exceeds 50% is set as the reliable reservation deadline", the reservation possible time prediction server 14 sets September 30 as the "certain reservation deadline". To certify.

Also, if the reservation rate on November 7, 2015 was "89%", but it increased to "92%" on November 8, 2015, the "reservation rate exceeded 90%". According to the second determination rule that "the day before is the reservable deadline", the reservable time prediction server 14 recognizes November 7 as the "reservable deadline".

Next, the reservable time prediction server 14 simply adds one year to the above-mentioned "reliable reservable deadline" and "reservable deadline" for December 1, 2015, so that December 1, 2016 Estimate the "reservable deadline" and "reservable deadline" for the day.
As a result, for ANA 461 flight, the reliable reservation deadline for December 1, 2016 will be "September 30, 2016", and the reservation deadline will be "November 7, 2016".

Regarding the forecast of the time when reservations can be made for airmail, the above-mentioned "correction when the event date is determined by the day of the week" and "reflection of the reservation record up to the present time", and "presence or absence of event holding" held near the arrival airport It is possible to improve the prediction accuracy by performing the same operations as "Reflecting the influence of" and "Reflecting the weather forecast information".
In addition, by comparing the flight prediction data stored in the flight prediction DB 40 with the actual data calculated based on the flight reservation data stored in the flight reservation DB 38, the current prediction accuracy can be analyzed and added to this. It is also possible to improve the prediction accuracy based on this.

[Reflect user information]
Regarding the prediction of the flight reservation time, it is possible to further improve the prediction accuracy by analyzing the address of the user who has entered the event and reflecting the result.
For example, the number of participants in large-scale events held near the arrival airport on the day after the forecast target date was relatively large in Kansai last year, while the proportion of applicants this year is increasing in Kanto. If so, it is desirable to add the prescribed ratio to the reservation rate of last year's results for the route "Haneda Airport → Naha Airport".
On the contrary, for the route "Osaka Airport → Naha Airport", it is desirable to subtract the prescribed ratio from the reservation rate of last year's results.

Specifically, the reservationable time prediction server 14 refers to the venue, start date / time, end date / time, and planned number of participation of each event information stored in the event application DB 26, and the prediction target date and prediction target area (arrival airport). Check if there are events of a certain size or larger in the vicinity).

Then, when the corresponding event exists, the reservable time prediction server 14 refers to the event application DB 26 and the user DB 28, and determines the address distribution of each user who applied for the event last year.
Next, the reservable time prediction server 14 refers to the event application DB 26 and the user DB 28, and determines the address distribution of each user who has applied for the event this year.

Next, the reservation available time prediction server 14 sets a preset value for the reservation rate calculated based on the actual results of last year when there is a certain deviation in the user's address distribution between last year and this year. Make a positive or negative correction with.

The process of predicting the reservable time by this system 10 can naturally be applied to other reservable objects.
For example, reservation of reserved seats of a train that stops in a specific area on a specific day in the future, reservation of a table of a restaurant located in a specific area, and the like are applicable.

In the above, the method of determining the future reservationable time based on the actual data of the past one year has been described, but the present invention is not limited to this.
For example, it is possible to estimate the future reservation available time based on the average value of the actual data for the past three years.

That is, when "December 1, 2016" is presented as a future reservation target date, the reservation possibility prediction server 14 first has (1) December 1, 2015, and (2) December 1, 2014. , (3) The three dates of December 1, 2013 are recognized as past reservation target dates, and the daily reservation rate up to each reservation target date is calculated for each year.
Next, the reservation possibility prediction server 14 calculates the average value of the reservation rate in the corresponding months and days of each year (for example, October 1, 2015, October 1, 2014, October 1, 2013).

Next, the reservation possibility prediction server 14 obtains the reliable reservation possible deadline and the reservation possible deadline based on the past reservation record data by applying the same determination rule as above to the average value of the reservation rate for each day.
Finally, the reservability prediction server 14 transfers the above-mentioned reliable reservation deadline and reservable deadline to this year (2016), so that the reliable reservable deadline and reservable deadline for December 1, 2016 Presumed to be.

When specifying a plurality of reservation target dates in the past, it is naturally possible to apply the above-mentioned "correction when the event date is determined by the day of the week".
In other words, "December 3, 2016" was presented as a future reservation target date under the situation that the date of the large-scale event held every year in the target area is defined as "the first Sunday of December". In this case, the reservation possibility prediction server 14 recognizes three of (1) December 5, 2015, (2) December 7, 2014, and (3) December 1, 2013 as past reservation target dates. Then, the daily reservation rate up to each reservation target date is calculated for each year.

Next, the reservation possibility prediction server 14 performs "adjustment" to reflect the deviation of the past reservation target days when calculating the average value of the reservation rate in the corresponding months and days of each year. As a result, for example, on October 1, 2015, October 3, 2014, and September 27, 2013, the average value of each reservation rate is calculated.

Also, when calculating the average value of reservation rates for the corresponding months and days of each year, instead of simply adding each reservation rate and dividing by the number of years, a predetermined weighting is applied to the reservation rate for each year. You can also make "adjustments".
That is, "increase / decrease in the number of guest rooms in accommodation facilities", "status of competing events", "increase / decrease in the number of event participants requiring accommodation (inferred from the address of each event applicant)", and "climate conditions (inferred from the address of each event applicant)". Since there is a possibility that the reservation status of the year may change due to "warm winter / cold winter, etc.)", it is possible to make a reservation in advance by associating a weighted value for each year that reflects these conditions with the event, region, reservation category, etc. Set in the prediction server 14.

For example, if the following weighting is set, the following adjustments have been made when calculating the average value between October 1, 2015, October 3, 2014, and September 27, 2013. The reservation rate is calculated, and the average value between them is calculated.
(1) [Weighted value]
2015 NAHA Marathon related ・ ・ ・ 0.9
2014 NAHA Marathon related ・ ・ ・ 0.7
2013 NAHA Marathon related ・ ・ ・ 1.0
(2) [Actual reservation rate]
October 1, 2015 ・ ・ ・ 50%
October 3, 2014 ・ ・ ・ 45%
September 27, 2013 ・ ・ ・ 60%
(3) [Adjusted reservation rate]
October 1, 2015 ・ ・ ・ 50 × 0.9 = 45%
October 3, 2014 ・ ・ ・ 45% × 0.7 ≒ 32%
September 27, 2013 ・ ・ ・ 60% x 1.0 = 60%

In addition to setting the above weighting value in the system by human hands in advance, the reservation possibility prediction server 14 automatically calculates the weighting value by referring to the predetermined data according to the predetermined rule. Of course, it is also possible to configure.

In the above, an example is shown in which the reservability prediction server 14 predicts the reliable reservation deadline and the reservable deadline for a specific reservation target date after receiving an explicit prediction request from the user. It is not limited to this.
For example, each time new event application data that implies a future reservation target date is registered (input) in the event application DB26, the time when the accommodation facility or airmail required for the user to participate in the event can be reserved. It can also be configured to have a mechanism for the reservation possibility prediction server 14 to automatically calculate and notify (output) to the user via a WEB page (My Page) or e-mail.
Incidentally, the necessity of booking accommodation facilities and airmail is determined by the address of the user registered in the user DB 28.

It is an overall block diagram of the reservation possible time prediction system which concerns on this invention. It is a figure which shows the structure of the event DB, the event application DB, the user DB, and the weather forecast DB. It is a figure which shows the structure of the accommodation facility DB, the accommodation reservation DB, and the accommodation facility prediction DB. It is a figure which shows the structure of the airmail DB, the airmail reservation DB, and the airmail prediction DB. It is a flowchart which shows the processing procedure of the reservation possible time prediction about the accommodation facility by this system. It is a figure which shows the prediction result presentation screen of the reservation possible time about accommodation facilities. It is a figure which shows the process of predicting the reliable reservation possible deadline and the reservation possible deadline about accommodation facilities. It is a figure which shows the process which reflects the reservation record up to now in the prediction result. It is a flowchart which shows the processing procedure of the reservation possible time prediction about the flight by this system. It is a figure which shows the prediction result presentation screen of the reservation possible time about the airmail. It is a figure which shows the process of predicting the reliable reservation possible deadline and the reservation possible deadline for an airmail.

10 Reservation available time prediction system
12 WEB server
14 Reservation available time prediction server
16 DB server
18 internet
20 Client terminal
22 External storage
24 Event DB
26 Event application DB
28 User DB
29 Weather forecast DB
30 Accommodation DB
32 Accommodation reservation DB
34 Accommodation Forecast DB
36 Airmail DB
38 Airmail Reservation DB
40 Airmail Forecast DB
42 External server
44 Data registration terminal
50 Prediction result presentation screen
52 Reservation button
54 Forecast condition change column
56 Reprediction button
70 Prediction result presentation screen
72 Reservation button
74 Forecast condition change column
76 Reprediction button

Claims (7)

It is equipped with a reservation record data storage means for storing past reservation record data having a reservation target date and a reservation reception date for a specific reservation target object.
When a forecast condition indicating a future reservation target date is entered, a process to identify the past reservation target date corresponding to this future reservation target date, and
The process of extracting the reservation record data that matches the past reservation target date from the reservation record data stored in the reservation record data storage means, and
A process of totaling each extracted reservation record data based on the reservation reception date and calculating the cumulative number of reservations of the reservation target object for each day up to the reservation target date in the past.
The process of calculating the daily reservation rate by dividing the cumulative number of reservations for each day by the quantity of the above reservation objects, and
Based on the reservation rate of each day up to these past reservation target dates, the process of estimating the future reservation available time and
The process to output this reservable time and
Is a reservable time prediction system that is executed
By applying the first determination rule with a predetermined reservation rate as a threshold value to the reservation rate of each day up to the past reservation target date, the reliable reservation possible deadline is recognized and a higher reservation rate is set as a threshold value. By applying the second judgment rule, the reservation deadline is recognized, and
A reservable time prediction system characterized by outputting these reliable reservable deadlines and reservable deadlines as the above-mentioned future reservable time. It is equipped with a reservation record data storage means for storing past reservation record data having a reservation target date and a reservation reception date for a specific reservation target object.
When a forecast condition indicating a future reservation target date is entered, a process to identify the past reservation target date corresponding to this future reservation target date, and
The process of extracting the reservation record data that matches the past reservation target date from the reservation record data stored in the reservation record data storage means, and
A process of totaling each extracted reservation record data based on the reservation reception date and calculating the cumulative number of reservations of the reservation target object for each day up to the reservation target date in the past.
The process of calculating the daily reservation rate by dividing the cumulative number of reservations for each day by the quantity of the above reservation objects, and
Based on the reservation rate of each day up to these past reservation target dates, the process of estimating the future reservation available time and
The process to output this reservable time and
Is a reservable time prediction system that is executed
Along with the identification of multiple past reservation target dates,
For each of these multiple reservation target days, the daily reservation rate up to each reservation target date is calculated.
The average reservation rate, which is the average value of each reservation rate, is calculated for each of the corresponding multiple days.
By applying the first judgment rule with a predetermined average reservation rate as a threshold value to the average reservation rate of each day up to the above-mentioned past reservation target date, the reliable reservation possible deadline is recognized and a higher average reservation is made. By applying the second judgment rule with the rate as the threshold value, the reservable deadline is recognized,
A reservable time prediction system characterized by outputting these reliable reservable deadlines and reservable deadlines as the above-mentioned future reservable time. It is equipped with a reservation record data storage means for storing past reservation record data having a reservation target date and a reservation reception date for a specific reservation target object.
When a forecast condition indicating a future reservation target date is entered, a process to identify the past reservation target date corresponding to this future reservation target date, and
The process of extracting the reservation record data that matches the past reservation target date from the reservation record data stored in the reservation record data storage means, and
A process of totaling each extracted reservation record data based on the reservation reception date and calculating the cumulative number of reservations of the reservation target object for each day up to the reservation target date in the past.
The process of calculating the daily reservation rate by dividing the cumulative number of reservations for each day by the quantity of the above reservation objects, and
Based on the reservation rate of each day up to these past reservation target dates, the process of estimating the future reservation available time and
The process to output this reservable time and
Is a reservable time prediction system that is executed
Along with the identification of multiple past reservation target dates,
For each of these multiple reservation target days, an adjusted reservation rate is calculated by applying a predetermined weight to the daily reservation rate up to each reservation target date.
The adjusted average reservation rate, which is the average value of each adjusted reservation rate, is calculated for each of the corresponding multiple days.
By applying the first judgment rule with the predetermined adjusted average reservation rate as the threshold value to the adjusted average reservation rate for each day up to the past reservation target date, the reliable reservation deadline is recognized and the reservation possible deadline is recognized. Reservation deadlines are recognized by applying a second decision rule with a higher adjusted average reservation rate as the threshold.
A reservable time prediction system characterized by outputting these reliable reservable deadlines and reservable deadlines as the above-mentioned future reservable time. Equipped with an event information storage means that stores the period of the event held in a specific area and the number of participants per day during the period.
The above reserved items are related to a specific area and
As a result of referring to the event information storage means, the event held on the past reservation target date or the next day in the related area of the reservation target object may not be held on the future reservation target date or the next day. If detected, any of claims 1 to 3 , wherein at least a part of the daily number of participants of the event is subtracted from the cumulative number of reservations to calculate the reservation rate for each day in the past. The listed reservable time prediction system. Equipped with an event information storage means that stores the period of the event held in a specific area and the number of participants per day during the period.
The above reserved items are related to a specific area and
As a result of referring to the event information storage means, an event that was not held on the past reservation target date or the next day is held on the future reservation target date or the next day in the related area of the reservation target object. Any of claims 1 to 3 , wherein when it is detected, at least a part of the daily number of participants of the event is added to the cumulative number of reservations to calculate the reservation rate for each day in the past. Reservable time prediction system described in. The above reservation target is an airline seat,
Equipped with an airmail storage means that stores the flight number, departure airport, departure time, arrival airport, arrival time, and number of seats of each flight.
In the above reservation record data storage means, the past reservation record data of each airline flight is stored.
When a future reservation target date and a departure airport and an arrival airport are input as the above prediction conditions, flights that match these prediction conditions are extracted from the above flight storage means.
Past reservation record data related to each flight is extracted from the above reservation record data storage means, and
Based on these booking record data, the future booking availability of each flight is estimated.
The reservable time prediction system according to any one of claims 1 to 5 , wherein the reservable time is output in units of airmail. The above reservation target is the guest room of the accommodation facility,
Equipped with accommodation facility storage means that stores the name, location, and number of guest rooms by price range of each accommodation facility.
The past reservation record data of each accommodation facility is stored in the above reservation record data storage means.
When a future reservation target date and region are input as the above prediction conditions, accommodation facilities that match the prediction conditions are extracted from the above accommodation facility storage means.
Past reservation record data related to each accommodation facility is extracted from the above reservation record data storage means, and
Based on these reservation record data, the future reservation availability time of each price range is estimated,
The reservable time prediction system according to any one of claims 1 to 5 , wherein the reservable time is output for each price range.

Images